As more tech entrepreneurs are applying their skills and energy to tackling issues like housing, transportation and public safety challenges in their communities, state and local governments are working to create ecosystems to support those efforts. GT talked to three leading gov tech startups (and checked in with a fourth) about the problems they are trying to solve and the challenges of starting and growing a business in this young but maturing space.
From 2006 to 2010, Jon Coss ran a Folsom, Calif.-based consulting firm that helped system integrators with large state government IT projects. During that time he saw gaps in those implementations that gave him ideas for his next business. “I worked on a massive California Medicaid contract,” he recalled, “and I was shocked by the large number of requirements and yet how few were related to fraud, waste and abuse.”
After Coss sold the consulting firm, he began looking around for new opportunities. A friend who worked at Google invited him to look at some of its analytical offerings running in the cloud. “Two things came together for me,” he said. “One was that you now have these cloud-based analytic technologies you can rent, and the other is that that there was an opportunity to apply this new technology to this old problem of fraud, waste and abuse that had been underserved in terms of attention in the past. While driving home from Menlo Park to Sacramento, I decided to start the company, and I incorporated a week later.”
In 2011, he launched Pondera Solutions, a Gold Hill, Calif.-based company that develops fraud detection, investigation and enforcement solutions for large government programs such as unemployment insurance, Medicaid and nutritional assistance. Its“fraud detection as a service” is a subscription-based, cloud solution that uses third-party data matching and prediction models to red-flag potential abuses for investigative follow-up.
Product development took some time and involved several iterations, admits Coss, a former Oracle Corp. executive. He wanted a product that would scale and be portable across states, and he wanted to hide all the complexity of the analytics and data sets and just deliver leads about potential fraud to clients. He hired a combination of predictive scientists, subject matter experts and technology specialists. “They don’t all speak the same language,” he says, “but they all have interesting things to bring to the table. Once they came up with good ideas, I thought they were original and not something a traditional tech company might come up with.”
One of the most promising graduates of the Startup in Residence (STiR) program’s inaugural 2014 class was Binti, which develops software to help speed up the process of approving applicants to become foster parents. Binti automated many steps that the city-county of San Francisco’s Human Services Agency used to complete manually, and established a system that foster care workers say saves time.
When Government Technology profiled the company in June 2017 it had signed up 21 of California’s 58 counties as customers. As of November, Binti had added four more counties in California, one in Colorado and a foster care agency in North Carolina. “We have expanded to private foster care agencies and are talking to a lot of agencies at the state level, but we don’t have our first state-level customer yet,” said co-founder and CEO Felicia Curcuru.
Binti had raised money a few years ago for an earlier business model to help families navigate adoption. It still had money from that, and grew its revenue from working with California counties. But it also took a $500,000 investment from the GovTech Fund. “We didn’t need it, per se, but Ron Bouganim, the managing partner, became a mentor to me because I was new to government,” Curcuru explains. “He was excited about our progress and offered to invest.”
The company has grown from five full-time staffers and several part-timers to 12 employees. “We sublease space in the basement of Code for America’s offices,” she said, “but we just made an offer on our own space in Oakland, and it has windows.”
Looking ahead, Curcuru says Binti wants to expand its impact. “I think we are on the brink of having customers in a lot of other states, so that will be exciting,” she said. The company also has a road map of different products related to foster care it plans to roll out.
“We also want to measure our impact more,” Curcuru said. “We benchmarked our impact with San Francisco and we have increased the number of foster families applying by over 300 percent. There are other benchmarks we want to measure. In addition to the number applying, we want to increase the number that actually get approved and make the process faster.”
Pondera spent 2011 and 2012 building the solution and developing new predictive algorithms. It beta-tested the platform with the Florida Agency for Health Care Administration. Its first unemployment insurance customer was the Iowa Workforce Development Agency in 2013. After that, customer growth took off. Pondera is now operating in 30 programs in Florida, Georgia, North Carolina, Pennsylvania, New York, Iowa, Nevada and California. Pondera recently won its first local government contract with the city of Los Angeles, which will use the platform for audit tax selection to go after businesses that may be under-reporting income.
In its sales approach, Pondera reaches out to the executives fighting fraud, such as the program integrity units. “Department directors and deputy directors are being told by their governors that they need to do something about Medicaid fraud, or they need to become more business-friendly by lowering unemployment insurance rates,” Coss explained. “Those are the folks we approach because our offering matches what they are trying to do.”
To fuel its growth, in June 2017 Pondera received an investment from San Francisco-based Serent Capital, reportedly in the tens of millions of dollars. According to Coss, Pondera is currently contracting with states that allow it to analyze about 34 percent of the country’s Medicaid beneficiaries. Looking ahead to 2018, he would like to expand on that. “The way fraud detection works, as we add data, it informs our prediction and detection models and makes us more effective. Then we hope more states will want to work with us.”
CivicScape, a company that applies predictive analytics to policing, grew out of its founder’s experience in the Chicago Police Department.
Brett Goldstein, who began his career as an early employee of OpenTable, joined the Chicago Police Department in 2006. After a year as a beat cop, he was asked to head up a new group that used advanced analytics to enhance police efforts. “Initially I was disappointed to find we were using data in a remarkably basic way,” he said. “We were data-rich but information-poor. But I was given carte blanche to try to change things.”
One change he made was to include other city data streams such as neighborhood-based 911 and 311 calls. “We needed to look at disparate data sources to better understand what was happening, and we had to be better listeners to the data,” he said. “We also needed to move beyond descriptive statistics to advanced statistics using machine learning and data science.”
The system went live in 2010. Goldstein says that it helped lead to lower homicide rates. “There is something to be said for using your data intelligently, coupled with smart resource allocation. You can really make a difference.”
After serving as the city of Chicago’s chief data officer, Goldstein left government in 2013. With a grant from the Laura and John Arnold Foundation, he led a team at the University of Chicago refining the predictive analytics platform. “We were able to introduce artificial intelligence and neural networks into the model,” he says. “We found we could produce models that were exceptionally accurate and worked hour by hour for small spatial areas such as three square city blocks.”
Photo courtesy CivicScape
To make the system sustainable, in 2017 Goldstein arranged a technology transfer out of the University of Chicago and launched CivicScape as a startup. “We spent six months building a smart user interface,” he said. “We focused on making it mathematically as smart as possible but also useable for anyone. We have a SaaS-based product to help people be in the right place at the right time whatever their mission.”
The cities of Camden and Linden, N.J., and Dearborn, Mich., were among the early users, and now Detroit is in the implementation phase, Goldstein says. It was important to him that the solution be viable for small and medium-sized cities. “One of the things that worries me in the technology space in public safety is that we design a lot of things for big departments that have technology resources and lots of money,” he explained. “That doesn’t mean that the small and mid-sized departments shouldn’t have something that works just as well and that is easy to operate. Instead of the common approach of looking at the top 25 departments in terms of size, I wanted to be able to implement in any size department.”
The concept of predictive policing has received criticism from groups such as the American Civil Liberties Union, which expressed concern about racial profiling using analytics, but Goldstein argues that transparency about the system can help alleviate concerns. “We have a GitHub page where we allow for algorithmic scrutiny, and that has been huge,” he said. “It gives a level of transparency so chiefs of police are able to talk about their initiatives in a transparent way. People can test our math. We don’t have a secret sauce or a black box. Potential customers can check it out and use their data and make sure they feel good about it.”
CivicScape is funded by Ekistic Ventures, of which Goldstein is a managing partner. Ekistic also incubates new government-related tech startups. “We spend a lot of time scouting out whether this is a problem we can solve and listening to pitches. I like to build things and want to have a positive impact,” he said. “My first one was CivicScape, and I feel like the traction is good. That is exciting. Now I am starting to think about what the next idea is. I am testing a couple of things and I hope one of them has legs.”
The co-founders of Alpharetta, Ga.-based GovSense complement each other well. Gary McTall spent 10 years consulting state and local governments on IT projects; Paul Cammisa has more than 30 years of finance, technology and business operations experience. They brought that together to begin creating cloud-based permitting, licensing and finance software for state and local government. In September 2017, GovSense was a winner in the inaugural Government Experience Awards from the Center for Digital Government* for its economic development initiatives, using GIS data for non-traditional visualizations like planning and budgeting.
McTall said the idea for the company grew from things he saw as a consultant. “No matter the size or location, I saw an abundance of silos. A lot of it is the way government procures solutions, budgeting by department, but I also felt that vendors in the industry have created artificial barriers and built systems to capitalize on departmental budgets, and I got burned out on it. I thought there had to be a better way.”
Having founded a company called InnoVergent in 2006, Cammisa had 10 years of experience with enterprise cloud applications in the private sector. “I came from the SAP world, so I understood large ERP applications,” he said. “I bring the cloud and business acumen to the table, and Gary brings knowledge of the public sector. Between the two of us, we are a great one-two punch.”
GovSense founders Gary MTall and Paul Cammisa. / Photo courtesy GovSense
Together they developed a software-as-a-service platform leveraging the Oracle Cloud and designed to break down those agency silos. “Our mission is to provide leaders with a 360-degree view of their organization,” McTall said, “with one database and one unified system to achieve the goals of the organization rather than departmental systems.” Their SaaS subscription business model is charged based on the number of users.
The two executives started building GovSense in 2014 to expand on the success they had working for private-sector and nonprofit clients. They launched it as a sister company to InnoVergent in 2016.
McTall says their software scales to work with jurisdictions of all different sizes. “We work with as small a community as Basalt, Colo., (population: less than 4,000) to an entity as large as the state of Colorado. Being able to scale to those sizes and everywhere in between helped us really grow faster than the traditional startup would in this sector,” he added. Other customers include the state of Texas and the city of Evanston, Ill.
The GovSense sales pitch to cities and counties is that they can replace ERP systems that are expensive to support and upgrade with a less costly service where upgrades are painless.
Cammisa used government economic development agencies as an example of groups that gain from the 360-degree view of constituents they describe. “If I recruit a company and its 40 employees to my city, I want to track the tax incentives offered, the business licenses and permits,” he said. “What are the sales taxes and property taxes? I don’t need to go into six different applications or send out an email and wait six weeks to get that information. I can log in and get at all of that — from the beginning of the economic development life cycle right through to gross receipts. That is a real differentiator.”
Always trying to stay on the leading edge of technology, GovSense is about to announce it will have blockchain capability built right into the platform, McTall said. “We are agile. We can do that within our platform without relying on third-party services.”
*The Center for Digital Government is part of e.Republic, Government Technology’s parent company.