What Binti’s co-founder has learned from building SaaS for social workers

After Felicia Curcuru watched her sister go through the arduous process of adopting kids, she set out to make it simpler for other would-be parents.

She thought that mission would take the form of a consumer-facing startup to guide users through labyrinthine bureaucracies. She raised pre-seed and seed rounds in 2015 on that premise and spent two and a half years helping “a few hundred families adopt.” But it eventually became clear there was only so much she could do from the outside.

“I realized that the government really controlled the whole process,” Curcuru said. “I was holding people’s hands through a process they didn’t really control.”

Curcuru decided then that a solution would have to start inside the government—but the rigid world of public sector red tape can be a challenging place for a young startup.

The realization set Curcuru on a path to what Binti is today—a software solution for child welfare workers currently deployed at more than 550 agencies across the country, according to the company. Binti has raised more than $60 million, including a recent cash infusion of $3 million from Melinda French Gates’s Pivotal Ventures.

Once Curcuru had that epiphany, she spent four months shadowing child welfare social workers in San Francisco County to see what their days looked like. What she found challenged some of her assumptions about what was happening behind the scenes.

“It’s a little embarrassing to say this, but at first, coming from the family side, I was like, ‘Agencies are the barrier,’” she said. “‘Families are struggling because agencies are not serving families well.’”

Instead, she saw well-meaning but overextended public servants whose office tools were no match for the scale of their work, she said. (“Imagine using a 70-column Excel spreadsheet with thousands of rows shared across 30 people.”) She started to see why children might fall through the cracks of this system.

“They’re the nicest people in the world, and they tend to have really bad software,” Curcuru said.

Much of the software used by local and state governments comes through customized deployments run by the likes of Accenture and Deloitte, Curcuru said. It can be tricky for the same software vendors that dominate corporate offices to meet specialized compliance needs or integrate with various legacy systems.

“What we did was we really studied what’s the same but what’s different across government agencies, and then ‘How can we build software where everyone can use the same software, and we can quickly configure the things that vary upon launch?’” Curcuru said.

To overcome those hurdles, Binti has to do more up-front work on each new feature, consulting with agencies at the state, county, and private level across the country ahead of release, Curcuru said.

“But then the benefit is, when we launch [with] a new agency, we’re not doing any engineering work. We can launch very quickly, instead of over the course of years. And then [there are] all the other benefits of SaaS, where we roll out new features, and everybody benefits from them,” Curcuru said.

After building the first version of the software platform, Binti was able to sign 21 government agencies over the course of several months.

“It shifted from feeling like I was pushing the boulder uphill to the boulder running downhill, because I’m actually solving a problem people care about [and] they’re willing to pay for,” Curcuru said.

Now, like almost every startup these days, Binti is exploring how it can best tap the latest generation of AI tools. The company has been using AI for translation and chat interfaces for case information. Binti also offers a tool that can record hours-long meetings between social workers and families and use the transcripts to pre-fill paperwork, she said.

Using AI in such high-stake settings with strict public-sector rules can require more vigilance than other applications might, she said.

“AI should not be deciding, like, ‘Do we remove this child?’ Or really sensitive decisions like that. But it can be used to save administrative work,” Curcuru said. “Then anything that AI does has a human in the loop to review it, and there’s audit trails—what did AI do versus what did the human do?”

Curcuru said Binti is currently working on new AI features as well as eventually expanding into other adjacent government verticals beyond child welfare.

For other founders who want to build products for governments, she recommended starting small with a discrete problem that needs solving. Binti spent three years working with counties and private agencies before leveling up to statewide offices.

“Spend time shadowing, if you can, and understanding the problem. And then find a wedge. It might be hard to go in and compete for a huge project out of the gate,” Curcuru said. “We found our first wedge of licensing foster parents, which is a very specific need and a smaller product…maybe you can start selling to counties or cities, and then get a lot of proof points, and then you can graduate to states.”

Focusing on a problem that you’re passionate about can also sustain you through more challenging times. “Startups are really hard,” Curcuru said. “And I do think that two and a half years in, when things weren’t working, I might have quit if it wasn’t something that I really cared about solving.”