The artifact that redirected Felicia Curcuru’s career was not a pitch deck. It was a spreadsheet wide enough to make the eye surrender. Seventy columns. Roughly a thousand rows. Shared by a couple of dozen social workers trying to track every form, training, inspection and background check required to approve foster families. When the grid became too much, sticky notes took over. The unofficial system had quietly become the usable one.
Curcuru found it in 2016, during four months spent shadowing the people who licensed foster families for San Francisco. She had arrived with a company, a mission and a product. What she did not yet have was the right point of view. Binti’s first version helped families complete foster and adoption applications online. It was tidy. It was sensible. It was also aimed at only one side of a process controlled by public agencies.
The spreadsheet supplied a different diagnosis. Families struggled because workers struggled, too. Caseworkers were not short on care or effort. They were short on tools that could show, plainly and in one place, which family had completed which step and what needed attention next. The bottleneck had a human face and the dimensions of a master grid.
Three people inside one customer
Binti’s first attempt exposed a familiar confusion in software: the person in pain, the person using the tool and the person buying it may not be the same person. Prospective parents felt the burden. Social workers moved applications through the system. County and state agencies bought the infrastructure. A product could delight the applicant and still remain peripheral if it did not help the worker or fit the agency’s requirements.
Curcuru’s pivot did not replace one customer with another. It expanded her definition of the customer. The family still needed a mobile application, electronic signatures and a clear way to submit documents. The social worker needed a dashboard, task tracking and records that survived staff turnover. Supervisors needed reports. Agencies needed software that could adapt to different rules without becoming another expensive custom project.
That distinction gave the mission operational form. Helping a child find a family could mean removing a duplicate field from an application. It could mean reminding a worker that a certification was about to expire. It could mean keeping a family visible after a colleague changed jobs. The moral purpose was large; the product work was minute. Curcuru learned to move comfortably between the two.
The design also had to leave room for people who preferred paper. In a 2026 conversation, Curcuru estimated that about four in five families at participating agencies choose Binti’s digital application, while the remainder can stay on paper and have a worker upload the documents. Adoption of a modern system did not require pretending that every user owned the same device, spoke the same language or wanted the same interface.
This is an underappreciated product choice. The digital path becomes useful because it is easier, not because the analog door has been nailed shut. For a system serving families under varied circumstances, flexibility is not a concession. It is part of the service.
A mission survives its first product
The story began years earlier. Curcuru was in high school when her older sister adopted two children. She saw a process that was difficult, complicated and stressful, even while the system needed more families. In her mid-twenties, she became a Court Appointed Special Advocate for foster youth in San Francisco. The contradiction stayed with her: children needed stable families, willing adults existed, and the path between them could still feel punishingly hard.
Her professional apprenticeship was unusually suited to complicated institutions. She studied economics at Wharton, spent three years at McKinsey working with large government agencies, tried a gifting startup, then joined the online venture platform FundersClub as its first employee. There she led user-experience and business-development work and learned the texture of an early company from close range.
When she returned to entrepreneurship, child welfare was the problem that mattered enough to endure. Curcuru raised $350,000 from angels, hired an engineering consultant and launched the first Binti. The software worked like an online guide for applicants: forms, IDs and collaboration in one place. She personally helped about 100 families through the process before a $2 million seed round in 2015.
Then growth plateaued at roughly 20 families a month. Adoption is generally not a repeat purchase. More importantly, agencies set the rules and ran the workflow. Binti could guide a family through a maze without being able to move any of its walls.
“I wanted to make Binti un-killable.”Felicia Curcuru
The correction was expensive. A team that had grown to six contracted back to Curcuru alone. A co-founder and close friend stepped away when their interests diverged. Curcuru lowered her salary to $30,000 and kept enough of the seed round in reserve to buy time. The joke was that she now had decades to solve it. Beneath the joke sat a disciplined choice: protect the mission, release the implementation.
The founder takes a chair
San Francisco’s Startup in Residence program created an opening. Curcuru joined the child welfare office and watched the work at full resolution. A home inspection started in the field with a safety checklist. The day ended at a desk with the master grid. Missing updates could turn into missing families. When a worker left, a case recorded only on paper risked vanishing from the team’s working memory.
The important observation was not “government uses old software.” It was finer grained. Workers needed a morning view of what demanded attention, a mobile tool for inspections, a reliable record of background checks, reminders for annual recertification and reports that supervisors could understand. Each sticky note was a tiny feature request written under pressure.
Curcuru called back two engineers from the first version and began building alongside the people who would use the second. She also sought social workers from eight states, asking them to map their processes end to end. The aim was software broad enough to share and flexible enough to respect different state and county rules. This was not customization by surrender. It was configuration informed by pattern recognition.
At the end of four months, San Francisco became the first government customer. Within eight months of building the agency product, Binti had contracts with 21 of California’s 58 counties. The sales motion changed character. Curcuru no longer had to push a consumer tool uphill. Social workers referred colleagues in neighboring agencies, and new customers arrived already understanding the problem.
Pricing required another translation exercise. Curcuru worked with San Francisco staff to calculate the time and money the platform could save, then used a fraction of that value as the starting price. It was a practical way to make an unfamiliar purchase legible: connect the software bill to work the agency already understood. She kept the rebuilt company lean, too. Binti operated with Curcuru and her technical co-founder until it reached roughly $500,000 in annual recurring revenue. The restraint reflected a lesson from the first attempt, when hiring had moved faster than certainty.
Government sales are often caricatured as procurement purgatory. Curcuru found that credibility could travel through the same professional network as the problem. A San Francisco social worker knew someone in Marin, Alameda or San Mateo. One successful deployment became a demonstration that a neighboring office could inspect. Word of mouth did not eliminate contracts or compliance. It made the conversation concrete.
From one office to a national footprint
Milestones reported in public Binti profiles and interviews. Bars show trajectory, not a linear numerical scale.Binti’s products now cover caregiver licensing, family finding and engagement, service referrals and administrative work. The company says more than 12,000 social workers use its tools and more than 100,000 families have moved through its systems. In 2026, Curcuru described a footprint of more than 550 agencies in 36 states, representing nearly half of American child welfare. Binti says it has raised more than $60 million, including a $45 million Series B in 2021 and new investments from Pivotal Ventures and SJF Ventures in 2026.
A boundary is also a feature
The company’s latest chapter brings a more powerful tool into a more sensitive room. In 2025, Binti partnered with Anthropic through its AI for Good initiative. Its features can translate an application, search a case file, transcribe a meeting and draft documentation from notes or a recording. Curcuru’s framing is deliberately modest: AI is another tool in the toolbox.
The boundaries matter more than the flourish. Binti says the system does not decide whether a family is suitable or recommend whether a child should be removed. Generated work remains a draft for a person to review. Curcuru has written that modern technology can return time to social workers, with early users reporting substantial reductions in paperwork time. The desired unit of progress is not text generated. It is time returned to conversations with children and families.
Use AI for transcription, translation, search and draft documentation.
Do not delegate decisions about children, families or cases to a model.
Save generated material as a draft until a worker checks and approves it.
That philosophy continues the lesson of the spreadsheet. Technology earns trust by locating the exact burden it can remove and declining authority it should not hold. A mature product does not need to perform every impressive trick. It needs to understand the room.
The glamour of the specific
Curcuru’s advice to founders is to choose a problem they care about enough to survive its hard periods. Her biography supports the claim better than any slogan could. She watched an early model stall, shrank the team, returned to fieldwork and asked strangers at conferences to narrate their documentation processes. This is persistence stripped of romance. It looks like listening to eight versions of a bureaucratic workflow until the common architecture becomes visible.
There is also a leadership preference hiding in Binti’s public values: let evidence improve an idea, create empathy through feedback, and give people enough information to decide well. Those themes echo Curcuru’s own pivot. She changed her mind because the work supplied better evidence. She did not confuse being the founder with being the oracle.
Her life outside Binti is described in gentler verbs: dancing, meditating, playing tennis, traveling, spending time in nature and with family and close friends. The list has an appealing lack of founder cosplay. Even the company’s grand ambition is expressed through ordinary moments: a worker finding the next task, a relative receiving a call, a form arriving in the language a family prefers.
The 70-column spreadsheet is gone from the center of this story, but its lesson remains. Important systems rarely announce their useful product opportunities. They leave clues in duplicated entries, improvised reminders and the patient expertise of people who know exactly where work falls through. Curcuru’s gift was to take those clues seriously. She followed the sticky notes until they formed a map.