HHS Startup Day pt 7 Pitches Through Close

Just another WordPress site

HHS Startup Day pt 7 Pitches Through Close

[music playing] >> Erin Hawley: It seems to be the mystery thing for everybody All right We — try it on this? Got it? There you go Hi everyone I’m Erin Howley, I’m the Vice President for DataRobot We are a machine learning company and as Becky was trying to explain earlier, it’s a little hard in five minutes to kind of move through this But I’ll share with you a little bit about what we’re doing as a company if I can get this thing to work Okay We’re going to talk about automated machine learning That’s specifically what our company focuses in on And what that means is Health and Human Services as we know, you guys have — this is not working You already have massive amounts of data I don’t think it’s a question about whether or not, you know, you need to gather more data What we focus in on is really helping agencies and corporations on taking the data they already have and being able to build out the most predictive model from that data and then make, be, take that model and then put it out into production so that you can — as the data changes, your models will change, and your information and your business insights will change So, if I could share with you and say to you, if you’d like to know who’s going to win March Madness, that’s something that DataRobot has been able to do We help the Major League Baseball teams predict the best players that they need to bring into their teams So, if you’ve seen the movie Money Ball, that’s similar to what we’ve done And how that all got started was by the idea that these founders got together We have some of the top data scientists in the world, who got together We have four of the number ones, and these folks pulled together, and they said, you know what, each of us is a specialist in one particular modeling algorithm Whether or not that’s Python or R or TensorFlow, let’s bring it all and put it into a single platform so that rather than data scientist taking weeks and months to build out one single model, let’s put all of the models into one platform so as our data comes in, the models will quickly go through and figure out what’s the best model? It could be a combination of these you see on the bottom It could be TensorFlow from Google, it could be H2O, it could be R. It’s going to actually show up for you within minutes what the best model is for your particular data set These data scientists got together — Amanda [spelled phonetically] is the one here in the center She’s the number one data scientist in the world right now and her — she’s the one who actually won March Madness Never knew anything about basketball, which is kind of ironic She lives in the U.K. and she’s featured on ESPN for her work there But the idea behind it is, let’s help people more quickly and efficiently build out these models using our software platform and being able to eliminate the complexity that happens right now I go into agencies all the time and the number one thing they say is we don’t even have any data scientists We can hire a lot of integrators and a lot of consultants, who of course as a software company we like to partner with, but at the same time we don’t have this ability to spend weeks and months waiting for this information to share — to come back forward for us So, the platform is developed, and the service was provided so that agencies could really help eliminate some of the problems they have right now Now that these top three adoption barriers that we see both in our commercial business and in our federal business, I think the biggest highlight is that 73 percent of the organizations just strictly don’t have data scientist or data analysts inside their workspace that understand how to build these models as quickly as we do So that’s part of the challenge that we’re trying to overcome and help people go with Kind of addressed this, there’s this unmet need for this — this unmet demand for data scientists Lots and lots of data, but we’re taking way too long, especially the federal government, with being able to produce more accurate and predictive models and results for the questions that they are trying to solve I was at the CIA in an unclassified data set they provided me It had taken them seven months and six of their top data scientists to produce a model that we outbid or out beat by seven, at least seven models better We did it in 19 minutes, total So And their model was a little weak We helped them enrich the data by using companies like Thresher to help them build out a better data set So, we’re going to help turn analysts and data scientists into being able to do more data science practices that they were not able to do prior to that I’ve got one-minute left I’m going to quickly just share with you a little bit more about the company because I think this is more interesting, specifically end-use cases If I didn’t spend so much time trying to get this — okay Here’s our federal use cases I think in the case of Health and Human Services, what we’re doing in Medicare and Medicaid specifically we have not actually — this is not a customer of ours, but it’s something that we’re able to share right now We show it actually in the demonstration We can actually help identify fraudulent claims and where, specifically down to the city and

town where doctors are submitting false claims We spend a lot of time in the intelligence community working on loan wolf, looking at social unrest, cybersecurity, being able to help predict and prevent some of the issues that they’re having in their networks, helping them understand what networks are going to go down and when they’re going to go down based on that prior data that we collected and then in workforce analytics Interestingly, in the government, that’s so far number use case Everybody wants to know more about the employees they have in their environment, what employees are likely to quit, what employees are likely to be at risk for certain things I have no time left So, this is a little bit more about the company We have 400 million models at this point on our commercial cloud We are a software provider on-prem, on the Cloud, whether it’s AWS or Google, it doesn’t really matter Over 300 employees, 180 of which are in the data science field Based on Boston, Massachusetts where they are quite said today that the Patriots lost [laughter] That’s it So, thank you so much [applause] >> Female Speaker: So, you mentioned that it can augment data scientists, so two-part question The first is, if I actually, the agency doesn’t have a team of data scientists, like you mentioned, but I have data analysts, is the tool sufficient enough for just a data analyst to start to learn and understand the outcome of what models are selected? >> Erin Hawley: Yeah >> Female Speaker: And that if your data scientists, would they be able to import customer models and run against your large data set of models? >> Erin Hawley: Right Both great questions So yes, so inside this model pool, if you will, we have hundreds of open source models, but we of course want to include those models that were built by your data scientists, we’ll put them into the platform And what happens is when the software runs — you can go to our website, it’s like a two-minute demo, and you’ll see how when it runs, a leader board creates so that all of the models are running at the same time and the number one models will show up and it’s a survival of the fittest It’s like Hunger Games for data scientists And the tops models will perform And if you’re data scientists had chosen — they might choose number three, because it might not have been the fastest, but they liked the way it was built better because it was a combination perhaps of what’s called an ensemble model It could have been an H2O model and an R model As far as data analysts, inside the federal government, that’s our sweet spot really because those are the folks who are most interested and most willing They know the data, they have to always know the data We’ll never know your data They have to know the data but then we give them these tools that allow them to very quickly build out these predictive models I’d always suggest that they, of course, go and work with a data scientist or at least make sure that they’re validating it against somebody who understands that information most clearly But data analysts are using Tableau Other technologies don’t have this predictive capability within their tool We spent three-and-a-half years of our five years building out the tool and that’s why there’s really nobody that we actually compete with I mean, IBM, sure, they have what they have, and bigger companies have models, but nobody has literally every single open source model available >> Female Speaker: Do you have any current work in the healthcare space? >> Erin Hawley: In healthcare yes, not federal healthcare But yeah, we do a lot in hospitals, we do a lot around remittance rates We can tell you down to what patients, with reason codes, what patients are most likely for instance to come back within a 30 days window based on all of the pre, the data you had previously collected about your patients That’s one of the demos I do and I’m not a data scientist I can tell you out of these 10,000 patients, 6,000 have not come back into the hospital within 30 days, but these 4,000 have, specifically why, so that then the hospitals have much better care treatments in place and they don’t risk not getting paid by Medicare/Medicaid >> Female Speaker: And how accurate are your models? >> Erin Hawley: Very accurate I mean, it just really depends on your data set, but when you see the results, you can see down to .99 percent, whatever it was based on the data Now, if we don’t — if the data, it has to be prepped and it has to be massaged, if you will, like an ETL tool would do We work with Becky’s company quite a bit so that when the data comes in to DataRobot, we’re taking off the data that’s been prepared so that when it comes into our system it then produces results So, you can see down to the accurately, the number one most predicted models down to the models that are the worst and why we kicked them out >> Male Speaker: Can you talk a little bit about how you decide to partner? You talked about working with some other companies in the sector right now Talk about your business model and how you choose partners to bring forward >> Erin Hawley: Right So, we love to work with partners, we are a software company I’ve been in the federal government my whole life in that contractor space and of course most of the town is filled with consultants and integrators, so we look to partner in opportunities like that where they’re going after these different proposals and they’re

looking for data science-type work, artificial intelligence, machine learning Those are the areas that we’re going to target We do best actually when we work directly with the agency and the agency has said to the integrator, “Listen, I know you have 5,000 people being trained to be data scientists, but we’d like you to actually partner with this technology or this company who will help, you know, bring to bear faster results I ‘ve had — I started the division a year ago and some of our best use cases are those where the government directly got involved and said, “We can’t wait five months for you to produce a single model, we want to have something much more quickly.” So, we’re open to partnering We really enjoy working with Becky’s company and anybody who is actually trying to deal with, you know, data So, it doesn’t matter if it’s Tableau, Cloudera, Amazon, Hortonworks Agencies have massive amounts of it and now they want to actually do something with it and have better insights and what better way to say, you know, I have this data, I want to understand, can I do this with it? So that’s why we’re very use case agnostic It’s kind of hard for me to say we just do healthcare, we just do financial services, but, by the way, financial services are our largest market Eight of the top ten banks are making phone calls to you when they catch a fraudulent activity We just busted up a $350 million money laundering scheme about six weeks ago with one of the banks that they never saw coming >> Male Speaker: So just wondering, have you targeted Fin Syn and that ginormous data set — >> Erin Hawley: Right >> Male Speaker: — that they have? I mean obviously The other piece is like how are you, just describe the way that you are actually awarded a federal contract or a federal OTA and was it cost based, was it fixed price, was it sole source directly to you? >> Erin Hawley: [unintelligible] >> Male Speaker: Because you talked about partners, but you started a year ago so I’m assuming you’re pivoting to more of a direct sale approach >> Erin Hawley: Right >> Male Speaker: If you could just talk about that a little bit >> Erin Hawley: Yeah, absolutely Conveniently or coincidently I should say, I guess Megan [spelled phonetically] from Decode is one of the organizations that we worked with who helped us identify two of our contracts and that one was with DHS We’re helping them on global travel assessment system Helping them identify high risk passengers >> Male Speaker: Yup >> Erin Howley: And we just finished up that program That was an OTA A subcontractor to two, on two of our contracts through a larger systems integrator, who I will say kind of went kicking and screaming into it The government had to say, “You will sit here, and you will leverage this technology because I just don’t have time for you to waste time.” Inside the agencies, it’s pretty much we want to identify the next potential Benghazi Can you come in and help us look at our data sets and help us predict the models? Did I answer your question? >> Male Speaker: Yeah So, and do you have pipeline? I mean, do you have like a three to five-year pipeline of opportunity or are you guys still at that point of development yet? >> Erin Howley: We’re — actually, you know what, I am, I started like a year ago and I always say, “Why do I keep doing this to myself?” I started Cloudera five years before their federal group So, it’s very, very hard and we have so much opportunity, but at the same time, you really have to — we’re spending a lot of time honestly on data science for executives We’re going a lot of training around what machine learning is, the differences between automated, which is what we do, and supervised, versus just the random artificial intelligence So, we do a lot of training That’s helped us build these really nice piolets I like to do it this way I’d rather, instead of saying, “Here, give me this ELA.” Because I’m not a big believer in that, “let me come in and do a piolet and if it works out, great” So, we have several six-month, yearlong piolets, and some of them, they give us six months, we knocked out their results within three weeks >> Male Speaker: So, just a piece or few back, that’s a good approach on pilots I would the roundabout way I would identify every single settee support small business that exists in DHS, Fin Syn in particular, maybe that’s — >> Erin Hawley: Yeah, we just hired somebody for that >> Male Speaker: — partner with them, educate them on your capability, because they’re going to actually structure the solicitations that you’ll want to compete for You’re going to want to go after high risk, highly political because a lot of people won’t compete for that, so it won’t eat up a ton of your — >> Erin Hawley: Thank you >> Male Speaker: — minimal proposal costs Educate them– >> Erin Hawley: Yeah >> Male Speaker: — and I think you’ll have a really good opportunity to grow >> Erin Hawley: Yeah I just hired somebody to cover financial services We’re just so overwhelmed by trying to understand like what — everybody thinks, “Oh you do machine learning, we need some of that.” And they don’t really understand what exactly we do But one that’s come up really interestingly is how much help we’ve been able to provide in just some piolets on grants management I mean there is a tremendous amount in grants management Anything else? >> Male Speaker: Two things One is, around the fraud abuse piece >> Erin Hawley: Yeah >> Male Speaker: You’ve been working with the inspector general here at HHS I suppose, if not, a place you ought to start >> Erin Hawley: Yeah >> Male Speaker: The other piece is the fraud and abuse work is distributed around the country as well Every state runs a Medicaid program

There’s an enormous amount of fraud and abuse The amount of technology that’s applied against the, either identifying or stopping fraud and abuse is really at a low level today So that’s a fertile place to look >> Erin Hawley: Okay >> Male Speaker: The next question is can you talk a little bit about in your investors? You have quite a bit capitalized there >> Erin Hawley: Yeah >> Male Speaker: And so, it’s interesting >> Erin Hawley: Yeah Excuse me, I’m losing my voice Our CEO, Jerry Eichen [spelled phonetically], he was one of the top data scientists you saw on the screen and in the first three years he said, “I can build a product and have it out the door in six months.” And then he started the product The company was like, I don’t want to do that I want to build something — and he kind of kept it secret for three years I was at Cloudera and knew Jeff Hammerbacher [spelled phonetically] was the founder of Cloud Era having started data science at Facebook So, when I knew, he was an investor to DataRobot, I was like, they must have something going on So, Jeremy’s the master at being able to take this platform and literally, he’s asked by VCs all the time Let us give you money and he doesn’t want it We feel like he doesn’t want to just be the labor to the VCs We have spent, I think the first $28 million that he got was on building out the product So, our VCs, the number one is NEA here in Chevy Chase and then In Q Tel is a small investor on the IC side and then, I should know these by heart, there’s a couple that are — but yeah, he’s done very well on raising money And now there’s a new approach I think now he’s doing something where it’s more — so many financial institutions want to get engaged, he’s like, I want you to be a customer because that’s more important to us because your use cases are so phenomenally good for our business >> Male Speaker: Really impressive >> Erin Hawley: Thanks >> Male Speaker: Have you had a chance to talk to anybody at Medicare? >> Erin Hawley: No So, here’s the problem, I started — and honestly Megan is amazing, and she really has helped me in some ways I’ve done federal for a long time, but this is a whole new ball of wax This machine learning, AI, and so I’ve just hired a couple of sales people and we have honestly not even our, some of our use cases are all around healthcare and fraud on Medicare, but we haven’t had anybody focus in on this space because I’m trying to build the company >> Male Speaker: I see One that’s just not for you, in each state we want grant out Medicaid, but the Medicaid dollars are hard to track because they’re at the state level And my sense is the states would be very interested because through it you could track many things including potentially the opioid crisis in a predictive way >> Erin Hawley: We actually are doing a demo on that We’re building out — getting the data is always the hardest thing but we’re building out a demo, I’d love to share it with one of you guys, it’s on opioid crisis The DEA said, “Can you help us understand where the people are moving? Where the drug dollar is going?” So, we tried to do that, but we need the data and then we have one right now, we’re going to do a webinar on February 27th and my data scientist is like, I can build anything, but right now the data is okay for an opioid crisis type of webinar and he’s hoping to get some more information to make it a richer data set Because we just don’t have data We’re never — if we do it on-prem, it’s your data, we’re not taking it back with us or anything like that so that’s why piolets are good for us too I’d love the opportunity We haven’t done the states just because I’m barely keeping up with federal [applause] Thank you [dramatic music] Drew Clark: Good afternoon My name is Drew Clark and I’m the CEO of Aperio Health and for the record, I’m not a data scientist, I’m not a technologist, and I didn’t sleep at a Holiday Inn last night, so I can’t qualify through that default But John, thanks for the segway to Medicaid and the opioid crisis We’re going to talk a little bit about our product which is really not, we’re not a vendor to HHS, we’re actually a solution provider to the providers and the patients that are actually the beneficiaries or the recipients of funding through those programs at CMS I think the clicker’s working Real quick, the team Lots of gray hair here The founder, John Knight [spelled phonetically] and myself and our CTO, Kelly Dire [spelled phonetically], all happen to be alum of Washington and Lee University, none of us were there at the same time Deep backgrounds in both the, let’s call it the public sector, in terms of publicly traded companies, and also in the entrepreneurial and venture capital world Healthcare and technology experiences that we’ve come together around our passion and vision to help bend the cost curve in healthcare So, opioid crisis We all know about, you can’t, you know, pick up your phone, you can’t pick up the Journal, you can’t listen to the radio or watch anything on CNN or MSNBC or any other program and not hear about this It’s a huge epidemic It’s actually a subset of behavioral health that’s even a bigger problem for this country which also drives the physical healthcare costs that John was mentioning earlier today 3.2 to 3.6 trillion Going through the roof We can’t sustain this growth pattern

We’ve got to change the way that we deliver healthcare, the way that we manage healthcare, and the way that the patient and her or his family members are engaged in healthcare So, real quick, this is a stat, a little actually dated right now, but the National Counsel for Behavioral Health is a big proponent, and I agree, that if we don’t fix the behavioral side of the equation, we’re never going to fix the physical health side, we’re never going to be able to bend the cost curve Value and devalue revolution Folks at CMS know about this, are involved with lots of different piolet programs to help integrate behavioral and primary care We started in the old world of what I’ll call sick care, health care 1.0 Fee for service, volume driven, provide the procedure, bill for the procedure, collect for the procedure, yaddy, yaddy, yadda Fast forward we get into an electronic health record mandated by the federal government and insurers in the commercial space and we’re now just beginning this transition into value care Ultimately, it’s going to be population care Providers are going to take full risk for subsets of the population and deliver that care Today, the electronic health records and the supporting healthcare information technology tools don’t enable anybody to do that And we’re not trying to compete with the Epics or Cernas of the world, we are community-based focused and I’ll talk a little bit more about that In terms of this revolution that’s occurring There’s a couple of great uses, two examples in the private sector of companies that are actually merging this world of behavioral health and primary care Also, in the CMS world you can think about PACE The programs that are all inclusive care for the elderly Well, Iora and Wellbe are doing that in the private sector Iora raised about 123 million in their rounds A lot of that went to build the technology platform that didn’t exist in the marketplace for them today So, if Aperio had been around two-and-a-half years ago, when Iora was coming out of the ground, they would not have had to spend as much of their capital on the technology, they could’ve used our platform This is an example of what we’re talking about in terms of community-based healthcare This is an example, Mosaic is an outpatient provider in Baltimore, they’re part of the Sheppard-Pratt health system which is an inpatient psychiatric facility and they are dealing with this married of community activities around a patient or around a population What’s interesting, if you look at this slide and think about all those different touch points in the community, none of those communicate with one another Data’s not portable, they’re not interoperable Sheppard-Pratt as an example has three different EMR solutions that they utilize, and other systems within the hospital to manage care They don’t talk to their outpatient providers like Way Station or Mosaic So, this is just trying to give you a visual of what’s going on in every community in America and the lack of integration and portability of data What do we need in community-based care? And this is not inside the four walls of the hospitals, but the hospitals are part of that ecosystem at the end of the day Right, we’ve got to have patient engagement, we have to make sure that there’s connectivity, that there’s data and there’s outcomes that are being driven from the care that’s being provided at the end of the day So, this is just a quick snapshot of what we think integrated community health delivery looks like Just a quick slide for those folks that have any interest We’ve built our solution from the ground up We’ve tried to use the more current technologies Tried to create a user interface and a user experience for the providers and their patients that�s much more in tune with today’s social media and the tool sets that people are using So, this is not your father’s or your grandfather’s EHR And just some data points about the structure that we use with metadata, we created some widgets and some tools that are proprietary around workflows and forms for the provider so you don’t need a data scientist, you don’t need a technologist, to create a clinical pathway, like a drug trial, to be able to manage that healthcare, create the outcomes, report back to your payers, report back to your constituents Just a quick snapshot of some of the products that are in the market place around behavioral health and primary care Obviously self-serving but we like to try to rank ourselves against the competition You can see that we obviously give ourselves high rankings across most of these data points but the user interface, user experience is critical, having a tool set and a platform that really does enable the provider to move forward And I’m going to get the hook here, in traction, we’ve got 2 million in capital that’s been delivered, we�ve got four customers in the queue, we’ve got some interesting strategic partnerships going on And I’ll turn it over for questions [applause] >> Female Speaker: Can you talk more about what you actually do? >> Drew Clark: So, we are an EHR on steroids, if you will We integrate — so think about a patient health record, the EMR component, think about social interaction and engagement with the patient We have APIs that are built in our library So not only is it healthcare on the physical and behavioral health side and substance abuse, all coordinated in one patient-centered view Because today, if you go into the marketplace, for the most part, you’ve got a separate EHR

and a separate patient view on the physical health side, behavioral health and substance abuse is over here, the two never cross Also, then building in the social determent So, tying in meals on wheels, tying in housing programs, tying in employment, food shortages, et cetera So, trying to build that network and incorporate tools like Healthify which is a partner of ours as well >> Female Speaker: How many of those data sources have you actually successfully integrated? >> Drew Clark: I’m sorry >> Female Speaker: How many of those data sources that you just mentioned, specifically the social determinants and behavioral health side? >> Drew Clark: We’ve got over a dozen APIs written today and anybody that allows us to write one, we’ll create that connectivity >> Female Speaker: And how are you dealing with issues related to standards? >> Drew Clark: So, at the end of the day, you know, our tool set’s been built around HIPAA compliance, our hosting is done at AWS in the Cloud, in their healthcare vertical, we’ve got our technology team that’s a nearshore and offshore combination, that works with the, you know, the particular provider on the other side to make sure that those data sets and that that standardization is in place So, again, I’m not a technologist, but I can turn you over to the engineers >> Male Speaker: Who is the customer? How big is this market? >> Drew Clark: So, the market today in the U.S. if you look at federally qualified health centers, community clinics, behavioral health providers, it’s about 15 thousand in aggregate We think its $2.5 billion market place just in the behavioral health side If you look at the clinic side and the integration, that’s going to be probably a seven-and-a-have to $10 billion market And again, we’re not trying to compete inside the four walls of the hospital with Cerna or Epic >> Male Speaker: So, the FQHCs generally have EHRs deployed around the country >> Drew Clark: They do >> Male Speaker: There’s an enormous amount of behavioral health providers that are still working off of paper or have another system >> Drew Clark: Right >> Male Speaker: So, how do you win share away from those that are already committed and how do you convince those without an EHR to finally get one? >> Drew Clark: So, the value proposition and the pitch at the end of the day is, if you’re not ready for 2.0 and beyond, you’re not going to survive Whether you’re a mission-driven, passionate, community-based provider that’s trying to do good in the community, you’re not going to survive There’s a number of aggregators already starting to occur Most of the competitors today that we would compete against, whether it be a NextGen or Creditable, et cetera, they’re customers are actually coming to us saying we can’t get the data, we can’t do the care planning, we can’t do clinical pathways, we can’t report outcomes without a lot of extra work on our behalf And oh, by the way, we don’t have any technology team, we don’t have a lot of funding, so if you can provide a SA solution that delivers a lot of that, we�re all in We’re going to rip and replace So, we’ve got 12 customers already that are committed to rip out and replace with us >> Male Speaker: And this could just be my lack of understanding, but something does concern me about the approach, is the APIs I mean the APIs that is a very complex, architecture So what level of scale have you actually delivered that service at? I mean, when you’re talking about multiple APIs like that, that’s a weak point There are companies that won’t do business with us because if we force them to integrate with one our APIs and it fails, we put them at risk Your entire model, if I understand it correctly, is built around APIs and you’re taking in tons of that data >> Drew Clark: Let me correct that So, our model is not built around APIs, we want to have the APIs in place in a library so treatment plans or, let’s say mood app If you’ve got a mood app on your phone and your therapist wants to check in with you right now that clinician would have to go online, pull up the portal for that app and see that information So that ancillary data we’re importing into the patient health record, but we’re freestanding, fully functioning EHR, CNR capabilities that are all built in our platform So, we’re not dependent on those tools out there in the marketplace, but if somebody wants to integrate that, that’s where our API library comes in >> Male Speaker: And what level of scale are you doing this at? Like from a dollar perspective, how many clients do you have? >> Drew Clark: We’ve got, in theory just relaunched back this year We’ve got four customers in the, that are ready to be installed, another 10 in the queue, so we’ve got about 1.8 million in contract revenue and we’ve got other strategic opportunity that could create another half a million to a million dollars in revenue So, we’re still early >> Female Speaker: So, I have a very basic question and maybe it’s actually a quest-estion [spelled phonetically] if you will >> Drew Clark: Yeah >> Female Speaker: You have some amazing people, heavy hitters from HHS here, but what’s your specific ask for HHS as it relates to your tool? >> Drew Clark: To be honest, there’s not a specific ask for HHS other than looking at some of the pilot potentials that are going on in either Medicare or Medicaid and around The PACE program is an example, the Impact Act, ER diversion programs, where we’re partnering with Hopkins, right not we’re talking to them about an ER diversion program for these call it frequent fliers, the high risk folks that are most vulnerable in the community that don’t need to go into the ER but that is their primary care doc at the end of the day

So I think, you know, for me trying to create some opportunities where there are pilot programs or thoughts about doing things different in community-based healthcare, that’s where we could hopefully partner and learn more from HHS >> Female Speaker: I’m going to follow that up with the reason why I was asking about standards and the APIs is that I genuinely believe that the model you’re approaching healthcare with is 100 percent the future I think there are a lot of contingencies right now that are dependent on choices that both the Office of National Coordinator and CMS makes in the next few years regarding interoperability and standards So, my suggestion to you would be, as new requirements come out from ONC, whether that be from Twenty-First Century Cures, or CMS, please respond with comments >> Drew Clark: Right >> Female Speaker: I think that there’s tremendous opportunity from your experiences in trying to integrate social determinants information for our agencies to understand ways that we need to adjust our requirements as far as health IT is concerned >> Drew Clark: We’ll take that Thank you [applause] [dramatic music] >> Ed Connors: Good afternoon, I’m Ed Connors, CEO of Heudia Health Heudia connects people in need with people who care Let me give you a little bit about of our experience and where we’re heading with the company So, my background’s in engineering and technology I was fortunate enough to win a phase one, phase two SPIR project to address the problem of childhood obesity My fundamental discovery was that unless you can get a lot of community-based organizations and engaged and working in the same direction to solve this problem, you would never be effective My phase one porotype was adopted into Charlotte, North Carolina, not to address the problem of childhood obesity, but to improve primary care access for a group of vulnerable people that were twice as likely to use the emergency room for primary care The takeaway from that experience was the ability to use our technology stack to address a common set of barriers that impinge on a large number of population health problems such as prenatal care or the opioid crisis and be able to address them in a very cost-effective manner Essentially the next step in the evolution of the company was to pair that technology with a methodology that came out of an NIHR 24 study that allowed that group to increase the number of folks using a community healthcare clinic by 180 percent over three-year period of time That correlated to a 20 percent reduction in avoidable emergency room costs for the healthcare provider So, when you think about what Access We Care does, it does for millions of people what any one of us would do in this room for a friend, family member or loved one It allows somebody to figure out what the needs are for the individual, identify the right set of providers, empower them and encourage them to take action to see those providers, getting them transportation if they need that, and then monitor that system of care to make sure it remains consistent with their changing healthcare needs What I did for my father who had cancer, is the same thing I now can do for a million of people using a cloud-based piece of technology called AccessMeCare Essentially the problem we’re solving in the market place is the vulnerable individuals need resources that are inconvenient and accessible and don’t address the barriers of care or the community-based care providers that exist in those communities need better networking intelligence tools to get the job done And this inefficiency and effectiveness of this problem bares on every payer, provider, and healthcare program in our country So, fundamentally, what’s the technology doing? It’s a community-based, cloud-based, focused collaboration platform designed to leverage a shared process so that anybody, anywhere can screen, navigate, and plan care for a Medicaid beneficiary, a veteran, somebody who needs ancillary community-based services using a de-identified, anonymous process because we don’t capture PHI We’re just matching the needs of the individual with the providers that exist in their community The technology stack is broken into two parts, one is directed against community engagement and contact creation and the other part is about user engagement and tactical outreach processes Ultimately, I’m supported by a great team of folks that are very passionate about improving health outcomes for vulnerable individuals We’ve done four projects so far, most recently with the Office for the National Coordinator of Health IT on a community health HIE program in Peddie, South Carolina, where we increased the number of vulnerable women going to a prenatal care program by 90 percent in 90 days The fundamental lesson learned is our lead measure, which allows us to get a large number of community-based organizations engaged and involved and using our technology within a

very short period of time Over a five-year period of time we expect that we will be able to get 300 thousand community-based organizations on board and working together using a turnkey software driven process called AccessMeCare Our immediate strategic objective is to drive AccessMeCare into a hundred counties throughout the Southeastern part of the United States where health disparities are the greatest and life expectancies are the shortest, to really demonstrate innovation impact the scale and really address this problem using a private-public partnership model Thank you very much [applause] >> Female Speaker: I just want to give you some props for doing a use case-based approach to this I think one of the things we struggle with in looking at these data aggregation platforms for delivery of better population-based care is that people try to solve all of the problems at once So, having done that is really cool that you picked a single problem Who is your primary customer though? Who is actually paying for this? >> Ed Conner: So, we have got two sales to healthcare providers and we’ve redirected our marketing efforts into Medicaid and managed care organizations where we believe that the financial value prob is four times greater with this population It’s also lines up with a number of federal initiatives to reduce ED avoidance and address a lot of other very important public problems facing the U.S. Department of Health and Human Services and the VA >> Male Speaker: So, two things, one is, can you talk about the source of the data that you use and then the other is, you gave prenatal care visits as one of the metrics in a project that you did Talk about some others where you have some experience and what you’ve moved the needle on >> Ed Conner: Yeah, so, let’s talk about cancer One of our greatest successes came with a project that we sold into a hospital as part of Chipper Demonstration Grant going on up in Northeast Pennsylvania We were — this will address your question about the Hewes [spelled phonetically] case We were out at a community health wellness fair at a Head Start program, and we were navigating folks, “Hey, you know, do you need health? Social service? What do you need to live a healthier, more productive life?” And a young grandmother, let’s say, with a six-year old child came up and said, “Do you have — we need vision services for this six-year old boy Can you help us?” And then as we’re going through that, she says, “Oh yeah, can you help me find a cancer doctor? I’ve had a cancer on my nose for over a year, I don’t know where to go to what to do.” The woman had four barriers to care She had a income barrier because she was laid off She had an insurance barrier because she was laid off living in a state that didn’t expand Medicaid When asked, “Do you know about the new cancer center that’s two miles from where we’re standing,” she says, “No.” And, of course, they had the navigators on board and the social workers to sort of address the first two barriers And then she had fear And that was very key, you know, we got her over that hurdle The other part of your question, I’m afraid I might not have answered it >> Male Speaker: The key sources of data >> Ed Connor: Yeah, so we actually crowdsource this data with the community Part of that NIH process is a series of focus groups, community forums, key informant interviews so that we get directly embedded with the community itself We do that over a very limited amount of labor to get the project started, and then we can do a lot of that over virtual services as well So, over time as we evolve the technology, more of that will come online and remove that brute force component giving us more better margins But we know that this area, the fundamental data and the content curation that needs to go on at a hyper-local level, does not exist And being able to do that at very low, incremental costs is one of the core components of our company >> Female Speaker: So, how will you translate that into collecting the data, let’s say, at the V.A. for example, where it’s a, you know, the population’s going to be a little bit different not when it comes to collecting community data >> Ed Connor: So So, we’re looking at the portion of the V.A patients that are living beyond a reasonable distance of the medical center themselves and need a lot of support services in their community It is more — it’s a broader regional effort that we would have to implement But it still could be done Fundamentally, the same barriers are the same >> Male Speaker: So, how would I do business with you if I wanted to do business? It sounds like there’s a service component and there’s a — so, how would you actually — have you done any work with the federal government, and how have they actually engaged you? And then, how would you like to be engaged if I wanted to do business with you >> Ed Connor: I will like to be engaged through a private-public partnership where we got some — maybe an IDIQ contract to address Medicaid beneficiaries in a particular region of the country Or through a demonstration project that would allow us, like, the outreach and engagement

component This is really where we’re trying to figure out today is how do we line up We’ve got a very strong federal stakeholder that would be very interested in helping us come into the Southeast part of the United States doing what we’re proposing So, we could do it on a PMPM model or outreach engagement Yes >> Male Speaker: Yeah, I was going to say, so, the largest managed Medicaid companies are Centene, Molina, United Healthcare, Aetna Remember, we block grant the states and then the states distribute those dollars So, I make the point that it might be useful to, on the Medicaid side, to partner with a firm that’s already engaged with all of them or most of them — but particularly Centene and Molina because they’re the ones who are then managing the managed Medicaid environment The other team here that might be useful for you to think about is George Segunas’s [spelled phonetically] team that manages Community Health Because that, you know, we have federal community health centers all over the United States And those, I think, really fit with what you’re talking about >> Ed Connor: Yes To your point, we are looking at engaging conversations with a larger MCO, managed care MCO’s, as well as some of the state level folks because we could line up with some of the state Medicaid plans as well But thank you very much for that insight [applause] >> Male Speaker: Second to last Second to last team >> Yali Friedman: Thank you very much for this opportunity This is not a sales pitch, I don’t want your money I want your data and possibly your support And I’ll start as soon as the slides come up Great So, my name is Yali Friedman I’m here to tell you about DrugPriceTracker It’s a project that started — the seeds were planted in November And I had a lot of fun tweeting about it, actually, right after the State of the Union So, it’s rather timely And the problem as it exists is that there’s a lot of players who need to know about drug prices, they need drug price intelligence Generic manufacturers need to know which drugs they should be launching, where there’s a good delta between manufacturing costs, and how much they can sell it for And when they do enter, they need to know how much they should charge for their drugs And there’s just too many companies who don’t have these data I was just talking to a company in Korea last week who don’t know — through IMS they can’t get the data they need And you would think IMS would be able to provide it to them Patients want to know about the drug prices Where can you get the best price for your drug? And when you do see an increase is it because of the supply chain, or the manufacturer, or what have you And media, of course, needs to keep industry accountable And so, the problem is that the existing solutions leave a gap You’ve got Medi-Span, and First Databank, and others They’re expensive, they have dated, cumbersome interfaces, and they’re focused on enterprise sales And so, that leaves a blue ocean underneath There’s 1,500 American pharmaceutical companies with fewer than 50 employees There’s a lot more abroad And all of these companies have really easy acquisition channels They can pay for things on a credit card, which is what I enjoy I’ve been publishing in the pharmaceutical intelligence space for 16 years now, and DrugPatentWatch is a new project which I want to bring to bear on this problem So, the solution is to combine a couple of federal datasets HHS has the national average drug acquisition cost, or a good surrogate for the average wholesale price of a drug V.A. has the best price available for a drug by law, and our Rx-list has the consumer-facing price And so, the methodology then is to combine these datasets There’s a published methodology showing just how the national average drug acquisition cost does align with the average wholesale price Like I said, V.A. has the best available price, and our Rx-list shows you what consumers ought to pay And so, here’s some screenshots I just threw together I was tweeting the day after the State of the Union the Xarelto price increase and got attention from The New York Times And so, that’s initial proof of concept All this comes from also an initial customer who’s paying for the development of this because they want to know where the increases are, so they know which markets to enter We can also tweet daily or weekly price increases to hold industry accountable, and this is a great way to market the product So, while serving the need of industry — of consumers, rather, and of media, the product markets itself It’s the same philosophy behind DrugPatentWatch, which is a company I’ve been running for quite some time now And as you see on the bottom left is the retail price for consumers and then historic prices

for industry in the two panels on the right So, the users within pharmaceutical companies, I won’t go into this in much detail You know the space, you see that it’s rather broad, we can hit a lot of different people And so, lot of potential customers out there The business model is very simple, subscription sales to commercial clients At $2,000 a subscription, it’s rather small It’s $3 million total address while on market Which you’d have to discount back, but you can also hit international customers The objective is to get initial traction with this very, very simple model using free data, and then expand from there That’s how DrugPatentWatch grew over the last 16 years, and that’s the model which I want to propose for DrugPriceTracker as well So, as I said our existing customers at DrugPatentWatch have asked for this product One of the customers has agreed to pay for the development of this product, and so traction is there It’s a brand-new startup The seed was planted, like I said, in November We have tons of clients across the pharmaceutical space, inside government, outside government It goes on a PCard, so it’s really easy to acquire Twenty-four thousand newsletter readers, 8,000 LinkedIn followers, 13,000 twitter followers — so the marketing is taken care of as well Team The Director of Business Development is up in New Jersey Marketing, there’s three individuals leading that Advisors, a former Executive Director of Licensing and Business Development with Wolters Kluwer Health which owned Medi-Span — or still owns Medi-Span And channel partner is Springer Nature It’s a $2 billion company with a global salesforce with employees in 50 countries And then there’s me, publisher of DrugPatentWatch since 2002, publisher of the Journal of Commercial Biotechnology, and I also do data analytics for Scientific American Custom Media So, finally the use case for HHS is that you want these data out there It’s not always so easy for you to get them out there, but middle-industry players like myself it’s what we do best I want to democratize your data and save customers, both in industry and on the consumer level, a lot of money Thank you for your time [applause] >> Female Speaker: I have a quick question Do you have a plan — in your business plan do you account for government affairs and the drug industry that probably does not like what you’re doing taking you to the hill with a bunch of lobbyists? >> Yali Friedman: That’s never happened before with DrugPatentWatch A number of cease and desist letters have come in But if they’re sending the data to the feds or, in this case the data’s coming from the pharmacies, the industry doesn’t have a role in it And, you know, it’s up to CMS The reason why CMS collects the data is congressionally ordained And so, it’s — >> Female Speaker: You know, I’m saying someone will probably lobby to help keep CMS from providing that data to a tool like yourself >> Yali Friedman: Right You know, it’s not something I�m going to worry about at this time >> Female Speaker: I would go ahead and worry about it [laughs] >> Male Speaker: So, having said that, what data exactly do you want? You open the pitch with, “I don’t want your money, I want your data.” So, be very specific >> Yali Friedman: I can’t be at this time because what I want to do is, as I’ve done before, follow the lead of my clients It’s a mistake to posit what you think that your clients want before they ask for it And there’s various ways to ask them, you don’t necessarily want to ask directly But I know that there’s more information, which CMS has around drug pricing Some of it you cannot release, but it’s worthwhile exploring what you can release to accomplish your goals I’m sorry if that’s a bit roundabout >> Male Speaker: Are you looking at any utilization data so that you can do some predictives according to, you know, use of certain drugs over time and then predict what the prices might be in the future? >> Yali Friedman: Yeah, there’s a whole area you can go into there I know that there’s more value on the industrial utilization side, which you guys are blind to, obviously, because you’re government But within Medicare and V.A., you’ve got a ton of utilization data which can probably be interpreted as a matter of looking at the literature and seeing how it aligns and how it’s useful But it’s very exciting >> Male Speaker: Yeah, you raised a great point I think that both on Medicare Part D and our Medicaid drug price negotiations, I believe that data is probably well protected statutorily >> Male Speaker: [affirmative] >> Male Speaker: But maybe we could talk afterward I know the name of a company in Kansas City that’s doing some similar work, and being very successful in the private sector, that you may be a nice augment with And they’re very good >> Yali Friedman: Cool, that’s really exciting I was scared of you at first [laughs] >> Male Speaker: Sorry >> Female Speaker: I would just make a recommendation, which is be very specific about the datasets that you want If you need to ask somebody else, ask them — >> Yali Friedman: Right >> Female Speaker: — and then come back I think that there are a number of champions currently who are interested in releasing

more data But in order to make that effective for the industry and for government, you need to be very specific about what it is that you’re looking for >> Yali Friedman: Yeah, we’ve had a lot of success with the Food and Drug Administration, getting data which they want out — out And everyone wins Thank you [applause] >> Male Speaker: Okay last one, but certainly not least >> Hua Wang: Raise your hand if you have been impacted by cancer, whether it is you, a family, a friend Statistically speaking, even more shocking than the statistic, up to 44 percent of cancer diagnosis and treatment plans are incorrect If there’s any moment in your life when you want to be absolutely certain, it is when you have cancer I have lost family members to cancer, and I know firsthand the fear, worry, and stress involved with trying to find the right information and the right team of doctors SmartBridge Health is a solution that I wish had existed when I was going through those difficult times What is our solution? We give patients, survivors, and caregivers remote access to top cancer doctors from leading academic research institutions We have 75 cancer doctors on our platform, and they have incredible expertise and have significantly lowered their rates to be part of our platform and help us cure cancer at scale Our business model is not brain surgery We give 70 percent of our revenue to our doctors and keep 30 percent We offer three products One is a next day call where you can talk about anything cancer related And the pinpoint in that is that it usually takes a few weeks to get an appointment with a doctor And during that time, people are Googling online getting even more stressed So, we take that fear away and match you with a specialist the next day, just to prepare you for that doctors meeting and answer any questions you might have We also have — the next product is the expert second opinion And we have found that our customers love this service in terms of peace of mind Even if we agree with everything the doctor said — the treatment, diagnosis — they’re just so grateful that they got another pair of eyes on this And then we just launched a new service, the clinical trial navigation, and that is based on customer demand We help people find the right clinical trials and help them get enrolled in these trials And we have improved health outcomes based on our three services We have a 95 percent customer satisfaction rate We remove a lot of unnecessary procedures We had, for example, one customer in a remote area And her doctor was a general oncologist and prescribed all these treatments that didn’t hurt, but just was unnecessary from our perspective Our competitors We are different from other telemedicine competitors out there We focus only on cancer, and we are working to democratize access to top cancer doctors The traction We have initial paying customers who love us The woman right there, her doctor actually give her a list of 85 clinical trials to research in getting into And she was on the verge of a nervous breakdown With our help, we connected her with somebody at Duke, a breast cancer specialist — she had stage four breast cancer — and narrowed the list down to five trials and helped her get into a trial And today she has improved health outcomes We are partnered with leading cancer organizations such as ZERO Prostate Cancer that raises $50 million plus for the cancer community And they’re kind of our channel partners, they’re helping us spread our message to the cancer community And we’re very excited to announce today that we are launching a pilot with the Inova Schar Cancer Institute to really prove that we improve health outcomes for a group of 75 breast cancer survivors The team I am a former lawyer and healthcare consultant We offer physician and faculty at Johns Hopkins, and we have an Emmy-nominated producer to work on our marketing efforts And also, our strategic advisor was part of a leadership of a team — of a startup — that went from an idea to IPO And our roadmap So, we started off B2C and now we are migrating to the B2B market with Inova, and so kind of our first B2B partner And we plan on figure out how to get employers to pay for our services as a wellness benefit, and also how to get insurance coverage So, but this year we’re focusing on a Inova pilot, getting that data published in a scientific oncology journal, and we really think that will provide the roadmap for other B2B customers

And we’re working with a few AI partners right now to figure out how the AI piece works to help us scale to millions of customers And our big vision is around big data We want to become the global source for real-time patient-reported outcomes where every data point, a patient’s story, can be used and reused to accelerate cancer research And so, I ask today for HHS, we are looking for advisors from HHS to help us navigate the federal government space We need warm introductions We need kind of help, kind of more guidance on the Inova pilot Kind of what kind of data do we need in order to get federal government to give us another pilot, for example So, thank you And join us in making optimal cancer care accessible and affordable [applause] >> Male Speaker: So, in one of your slides you talked about 40 percent reduction, I think, in unnecessary services — or services in 66 percent change in the care that was delivered So, this seems like you eliminate unnecessary care >> Hua Wang: Yes, that’s what we have found [affirmative] >> Male Speaker: So, can you put a price tag on what your forecasts are? That on a per episode basis what you’re able to reduce the overall cost of care? >> Hua Wang: We don’t know [laughs] because we only — we launched last year, and we have initial paying customers And it’s hard for us to quantify, you know, based on a few customers what the projection is for other customers But that’s why we’re trying to become more institutionalized, and that’s why the pilot with Inova will really help us get that data and figure it out >> Male Speaker: Is that part of what the research plan is? >> Hua Wang: Yes And that’s why we also need advisors, to kind of help us figure that piece out, yeah >> Female Speaker: How many customers do you currently have? >> Hua Wang: Yeah, so we have initial paying customers — 15 customers But the pilot with Inova is 75 with expected 75k in revenues We’re early-stage startup, so >> Female Speaker: And how is your clinical trial finder different than the Flatiron model? >> Hua Wang: The Flatiron — so, I’m not sure about — I know Flatiron, they’re amazing They’re really focused on the big data piece as well I feel they’re a little more enterprise heavy, they’re not really focused on the consumer itself We really pride ourselves in taking a very personalized and hands-on approach for every customer So, one woman with the 85 clinical trials We spent four hours with her to find the right trial And doctors in other companies are not taking that level of service for them >> Female Speaker: So, that’s fantastic — sorry But how — so you talked about adding in the big data component to scale Is every — of those three business models that you have, is every interaction that labor intensive from your perspective? Meaning, if I’m calling someone obviously I need to talk to somebody And then you have the other two How are you going to scale when you are doing hundreds of thousands of people? You can’t spend four hours with everybody I would imagine >> Hua Wang: Yeah So, right now we are gathering templates for expert second opinions And we are talking to a few technology partners possibly to figure out how to — where the AI piece comes in So, our vision is we’ll have an algorithm that will match the patient to the right doctor Right now, it’s manual We do that matching We will have — right now when we get the patient’s medical records it’s usually 100 plus pages, and then we have a doctor kind of make a summary to a two-page and then give it to our specialist And we think AI could provide, perhaps, provide that summary or also help us translate the clinical trial data So, all the clinical trials are publicly available, but a lot of people found them really confusing to understand and read without a doctor’s advice And we think — right now we’re translating that complicated language into something a patient understands and knows what the next steps are And we think AI could play role in that as well >> Male Speaker: So, you mentioned the medical record I assume that any second expert’s [unintelligible] opinion is only based on the existing data in the patient’s record There’s no additional test can be ordered or procedures for — if there are additional questions asked, is there any way to communicate that to their care team, or how do you share data back to the care team? Is just another document they walk in and show their oncologist, or? >> Hua Wang: Sure So, right now we base it on existing medical records and tests, but there are cases where we suggest, “Well, we suggest you get a biopsy You need a CAT scan.” We’ll tell them what of this new information we think is helpful and can need We also found we work with a lot of international customers, and the United States is number one in cancer So, everybody wants to know what United States doctors think about cancer And we have found that they collaborate They put our doctors with their doctors on the phone They share our information with them, so it’s a very collaborative environment >> Male Speaker: Related to that, how do you recruit your doctors? You said it’s pretty small trials now, so it’s a small pool How do you source and find qualified people to make those [inaudible].” >> Hua Wang: Sure So, we have a doctor who is a co-founder at Johns Hopkins And I feel like, based on our network, we have really found people with the right motivation and kind of desire for global impact to be part of our effort

They are getting paid a small amount, but that’s kind of secondary And we — yeah >> Male Speaker: So, I think you’re onto something interesting It’s not well known outside of the clinical circles, but there are no standard protocols for cancer treatment anywhere in the United — there’s not a national protocol And so, what you’re dealing with and identifying for consumers is how to simplify the development of their understanding of what “what” is Because everywhere you go, the protocols are actually widely different It’s fascinating, but they all get reimbursed They all get paid for it, whether it’s by the federal government or by private insurance companies And I think what you’re doing is building a consumer market that creates an intelligent consumer as to what their options are So, I really applaud you for that because it’s hard to do that But what you’ve fallen right on top of is the reality that there is no national standard protocol for any cancer treatment >> Hua Wang: Yeah [affirmative] And only about 10 percent of cancer doctors have the expertise to really match you with the right clinical trial And there are no platforms [laughs] [applause] >> Male Speaker: Okay That concludes today’s “Shark Tank” experiment — experience and experiment here at HHS If we could have a round of applause for our sharks and our teams, and my boss Bruce Greenstein will help close us out [applause] >> Bruce Greenstein: Okay So, it is a few minutes past 3:00 p.m So, before everyone leaves, and if everybody just kind of stays still for a second, this is the first time we’ve done this I’ve learned a ton And as a matter of fact, when this idea was getting kicked around in the office somebody brought up that episode of “West Wing”, the “big hunk of cheese” one where, if you remember, people come in from the outside And then everyone takes meetings with people from the citizenry And it’s an opportunity for people that, you know, are so busy and important to — that keep their head down and working within their office and within their structure And they miss the opportunity to talk to and with the people that they’re either regulating, or that provide the solutions to their problems And so, this was a really interesting day It’s a sense for us of humility Where we have to open our eyes and we know that there are all these parts of the healthcare world that we navigate, or control, or pay for — but we don’t have our hands around We don’t have all the solutions And we need to go out to others and say, “Hey We need some help What do you guys think? What can you solve of our problems?” And this was a great example We had eight different teams — eight different companies This was not, like, from the garage where people are, you know, just got a laptop and they have some kind of idea and they’re not really sure These were rock-solid companies that are solving some major problems that we have, many of which we don’t even know how to procure or even how to ask for help on So, it’s kind of subversive the way that you all can show problems and solutions, and many of which we might not even know exactly that we have these problems And it helps us start to think about them, and then take action on it So, thank you very, very much for being here But let me ask one more thing from you guys So, we’re going to do these again, and we want to fill the house, and we want to make it non-stop action And so, we want some feedback So, there’s a quota of three pieces of feedback, no one can leave until we hit three So, can you guys give me some feedback on ways that we could do this better next time? Yeah >> Male Speaker: [inaudible] >> Bruce Greenstein: Could we get a — can you hold on one second? We’ll grab a mike So, we need two more comments feedback, so keep thinking >> Male Speaker: Sure, thank you And, like I said, I’m one of the co-founders of Regendis [spelled phonetically], and obviously — and John, thank you for your really tough questions because we’re obviously clearly earlier-stage than most of the companies So, the feedback would be is there room to kind of separate out two different groups You’ve got sort of “startups” — and I’ll put air quotes — that are doing 10, $20 million dollars in revenue That’s very different than a handful of the companies here, myself included, who are early We have de minimis revenue, we’re looking for that entry point And I think that the needs of those two groups are different, that the questions being asked are very different And so, having that in maybe two different sessions, or — >> Bruce Greenstein: Yeah, I like that I like that One for those with revenue, and the others that are still working in their mom’s basement I love it [laughter] I’m kidding But I get it Good Yeah >> Male Speaker: Can I respond to that a little bit? >> Bruce Greenstein: Yeah >> Male Speaker: Yeah, thank you for saying that Because we talked a little bit earlier during my visit about preparation for Olympic wrestlers

and I coached in the corner for 20 UFC World Championship fights And the reason why I bring that up is the preparation piece is a big deal In other words, putting yourself in a position to understand what the potential risks and outcomes are And there are two things I would always share with folks building startups First of all, as an overlay, no matter what you’re doing I admire you Because anybody that takes the risk to go build something on their own puts themselves right out there Two, it’s very important early on to develop two understandings — to the best of your ability, these will evolve over time What is your quantifiable value proposition, right? The best you can, quantify your value proposition so you can say in clear terms that if we get this right, we think the value we create for our client looks like this And then two, try to quantify the addressable market Again, your view of that will evolve as well, particularly as you morph your product and your services, maybe combine the two But future investors or future clients need to understand first what the value is, you know, the best of your ability And two, what’s the scale of your opportunity And I think if you can speak to those two, you’re in a great place because we understand all along the way the pieces and parts will interchange and move Does that make sense? >> Male Speaker: [inaudible] You know, like I was saying, I just think the ask of the companies that are earlier on is the different than the ask of the companies who are later on And so, having the resources and the proper — as I walk my mic up to you if [inaudible] [talking simultaneously] >> Bruce Greenstein: Actually, if you could hand it to the next person over here >> Male Speaker: [inaudible] >> Bruce Greenstein: Okay, go for it >> Dominic Bonaduce: [inaudible] Dominic Bonaduce, I run a co-working space in northwest D.C called Alley power by Verizon We had a number of our companies that work in our space actually present to you guys today Two, maybe three, pieces of feedback So, one, it’s maintaining attention We do — I do an event like this almost every week, or I go to an event like this a couple times a week And one of the things that I thought would be extremely beneficial to this program — I might have come here to listen to you speak, Bruce, but not you, John, sorry But one thing to kind of say is one thing you can do is you can have all of this going on at once into breakout rooms, and then create an environment in the middle to mix and mingle One of the biggest proponents or solutions for providing a space for innovation is to have a chance for everyone to get to know each other in the room I didn’t — I, you know, I talked to a few people in the space today that I knew, introduced myself to a couple people, but there wasn’t really a time that was designated for all of us to say, “Hey Nice to meet you, I’m Dom What do you do?” And even to engage yourselves up on the stage I felt like there was a veil in between us and you And I think the idea behind this whole program was to kind of lift the veil and demystify it respectively, but also to make yourselves a little more available to the startup community off of just a stage from presenter to listener There’s just some feedback >> Bruce Greenstein: That�s great feedback I love it, and we’ll take that to heart for sure Now I feel guilty I stayed here for, like, an hour I missed lunch, I ate a PowerBar up here But that is very, very good feedback So, thank you very much, I appreciate that There was another — there’s a couple more Yeah, go for it >> Male Speaker: Can you guys hear me? Okay Second that notion I also have a very easy one Just everybody should just have a handful of slides, there’s just a ton of information So, just having links or anything that the crowd can take pictures of Super easy suggestion, but it would just be helpful to take the information, and bring it home, and be able to just utilize that, right? Because there’s no way to capture, just too much information >> Bruce Greenstein: Yeah, great — >> Male Speaker: Easy fix Even if they’re crappy slides, just something would be super helpful >> Bruce Greenstein: I love it Good All right, that’s an easy one Good Up here We’re busting the quota >> Erin Hawley: Erin Hawley with DataRobot So, I think what’s interesting for a lot of companies our size is that we get this opportunity But, like, on “Shark Tank” they make a decision They sort of say, you know, “We like your idea,” or, “We hate your idea,” or, “We’re going to do something with it.” I know this is not the way the government works, but one of the things I loved about DHS and their Silicon Valley Innovation Program — because everybody now has something going on in Silicon Valley But what I like about it is they put out on their website what they’re looking for And they kind of lay out the requirement and then you’re given a chance to — 10 pages or less — submit And it’s kind of unheard of because you actually get feedback, and I think [laughs] that a lot of times you don’t get anything back So, they actually respond to you, and if they like what you said then they give you a date and you have a 15-minute oral presentation And you can, you know, crush it or you can get yourself killed and, you know, be sent

off the island But after that they ask you a bunch of questions And then they let you know within 48-hours whether or not they’re going to fund that requirement And you knew going into it that each phase is worth between 50,000 and $200,000 So, we’re putting skin in the game The agencies are putting skin in the game And frankly, any of us who are in the software industry, God love our integrators — trust me I love all of you, but it’s really tough sometimes to get the thought leaders inside the agencies to identify some of these emerging technologies, whether we’re Silicon Valley or Boston-based like we are And to have that interaction with the government, and then to have something like that program which is that quick to award these emerging ideas and technologies based on a requirement you had, I think means a lot And it helps us move the needle to help your agency So, that’s my — >> Bruce Greenstein: Okay, that’s very interesting Couple things One is you implied that we don’t make decisions We might not be announcing them but trust me, these are some very smart people with very big checkbooks and large programs that they’re running If you left a solid and positive impression, people make decisions And so, hopefully you’ve got everybody’s phone numbers and cards, and you’ll follow up Each of the teams ought to be following up Because if you miss this as a beady [spelled phonetically] opportunity, we need to have something very rudimentary that precedes events like this But your second piece is really — is interesting It would change the way that we do procurements or something in a more iterative way And that is not outside the realm of possibility The more that we learn — John heads up this reimagine the way we procure across all of HHS And so, my suspicion is he doesn’t come up with a redesigned plan that’s just, like whatever, 98 percent of what it was before but really, completely rethinking that process So, if you have a best practice to share, I’m sure he’d appreciate hearing more about it But it also — you all have laid seeds in the field today that, you know, may take one season to sprout or maybe take longer — or you might get a call next week So, for us, again, that’s the “big cheese piece” that we get to learn from those that come in as well Is there one last comment? Yeah >> Male Speaker: So, appreciate the opportunity I thought the “Shark Tank” portion of this was really solid I think the speakers beforehand would have been more helpful to us if you guys had just hard and fast gave us your spending priorities for the year You guys have big budgets, we know you have big budgets If you say, “You know what? Startups I want x, y, and z,” then we’ll come back to you with x, y, and z A lot of times it’s really hard when the differential in size, right — 100 employees versus 70,000 employees, 80,000 employees — to know what the real spending priorities are So, anyhow, that’s what I’ve got >> Bruce Greenstein: Yeah, that’s a good piece of feedback as well All right, so this was very helpful for me to prepare for the next one Don’t feel like the feedback has to end Probably what, best way is [email protected] So, that’s the email address — [email protected] And you could do it anonymously or, you know, let us know who you are We really appreciate the feedback So, let me thank three, four — first of all let me thank Kevin for putting this together today, really, really appreciate it [applause] Let me thank the companies that came to present You guys were button-down and tight, and it was really, really informative and entertaining [applause] Let me thank the judges for giving insightful feedback, that was really good [applause] And then let me thank everyone that came to visit with us today, those that are here in the room, and those that are watching the webcast Thanks very much, and we’ll see you next time [applause] >> Male Speaker: Produced by the U.S. Department of Health and Human Services at taxpayer expense