Binti pilots AI in Florida

Child welfare software is not a flashy AI category, but that’s exactly why this Florida pilot matters. The work is brutally paperwork-heavy, the stakes are high, and most of the value comes from shaving hours off forms without messing with judgment. That is the setup for Binti’s new AI push in Florida. The company’s tools are being piloted with agencies including Heartland for Children in Polk County, where caseworkers are using AI to help draft adoption and licensing paperwork instead of starting from a blank page every time. (baynews9.com) ### What is Binti actually doing? Binti sells software to child-welfare agencies — foster care, adoption, licensing, placements, referrals, the whole admin spine. Its AI layer sits on top of that workflow and does three main things: transcribes meetings, drafts forms and case notes, and lets workers search across case files and agency information in plain language. The important d(baynews9.com)eviews. The worker decides. (binti.com) ### What changed in Florida? The concrete news is that Florida agencies are now using those tools in live pilot settings. Spectrum’s Tampa Bay report says Heartland for Children, which serves Polk County, was selected for the pilot. A separate Florida TV report earlier this year showed Family Support Services using Binti software in Florida casework more broadly. So this is no longer a lab demo or vendor promise — it is showing up inside day-to-day agency operations. (baynews9.com) ### Why is paperwork the bottleneck? Because child welfare runs on repeated documentation. Home studies, licensing packets, adoption files, case notes, status updates — the same facts get entered again and again for different people and systems. One adoptive parent in the Polk County story described being asked for the same information by separate groups that were not connected. T(baynews9.com) one interview or one set of notes into a reusable draft. (baynews9.com) ### What does the AI part actually automate? Mostly translation from messy inputs into structured drafts. Binti says workers can record family meetings, transcribe them, and generate draft paperwork or case notes. They can also upload handwritten notes. Then there is a search tool that pulls answers from forms, notes, case data, and policies. Think less “robot social worker” and more “autocomplete for bureaucratic memory.” That is a boring description — but boring is good here. (binti.com) ### Why does the human-review piece matter so much? Because child welfare is one of the worst places to let AI improvise. A hallucinated fact in a marketing memo is annoying. A hallucinated fact in a family case file is dangerous. Binti leans hard on that boundary: no decisions, no recommendations, no replacing caseworkers. Everything lands in draft mode for human approval. It also says the system is HIPAA-compliant and that customer data (binti.com) ### Is this a tiny experiment or something bigger? Bigger. Binti says it works with more than 550 agencies across 37 states and DC, serving agencies responsible for roughly 49% of children in care nationwide. When it launched its AI offering with Anthropic in August 2025, it said pilots had already run with 20 child-welfare agencies. So a Florida rollout matters less as a one-off local story and more as evidence that this product is moving(binti.com)rocurement and use. (binti.com) ### What is the real upside? Time. That sounds small, but in this field time is the scarce resource. Binti says many social workers spend more than 50% of their time on administrative tasks, and one user testimonial on its site says AI cut report-writing from 3 to 4 hours down to about 2. Even if those numbers vary by agency, the thesis is clear: fewer hours formatting paperwork means more hours with children, families, and placements. (prnewswire([binti.com)news-releases/binti-launches-first-of-its-kind-ai-for-social-services-offering-with-anthropic-unleashing-ai-to-its-customer-base-of-agencies-serving-46-of-child-welfare-across-the-us-302533917.html)) ### What’s the bottom line? This Florida pilot is a useful preview of where government AI is probably headed first — not decision engines, but note-to-paperwork machines. That is less dramatic than the hype cycle wants. But turns out it may be the more realistic win. In systems like child welfare, getting the admin burden down is not peripheral work — it is the work that makes the human work possible. (baynews9.com)