Software for state and local government

Public decisions that hold up.

Agencies make calls every day that they later have to explain: which AI tools to allow, which projects to fund, how to help someone back into work. We build the software for those calls, so the reasons are written down, a person makes the decision, and the public can see what was decided.

See what we make Talk to us

AssayWhich uses of AI are allowed
BearingWhich projects get funded
PathwaysWhich path gets someone back to work

Three products, one way of working

Each one is built for a decision an agency has to make and then stand behind, in front of its board, its auditors and the public.

Assay

Taking pilot agencies

Know every AI use. Show your work.

For CIOs, privacy officers and AI program leads

Assay finds the AI your staff already use, helps you decide what's allowed and why, and publishes a plain-language register for residents.

Explore Assay

Bearing

Demo on request

Fund the projects that matter most, and show why.

For transportation agencies, cities, counties, utilities and grant programs

Bearing runs a call for projects, a capital program or a grant round from submission to adoption, in your agency's own terms: scoring against your criteria, a ranked list with a funding line by year and source, scenarios to compare, and a ranking the public can check. AI can suggest a score and show the passage it's based on; a reviewer decides.

Explore Bearing

Pathways

In development

Help people find work that lasts.

For workforce agencies and their caseworkers

Pathways gives caseworkers one place to see the people they serve and what each person has told them, with every decision about their path kept on the record.

Explore Pathways

What every product shares

All three are built on the same foundation, so a safeguard proven in one is carried to the others, and your IT and security teams only have to learn it once.

A person decides: Where our software uses AI, it suggests and shows where the suggestion came from. A person makes every call, and the record says who.
A record that holds up: Every change and decision is kept in order and sealed to the one before it, so nobody can quietly rewrite history. Your auditors can check it themselves.
The public can see it: What residents should know gets published in plain language, from the AI an agency uses to the projects it funds.
Fits how agencies work: Staff sign in with the accounts they already have. We host it in US regions, or your IT team runs it in your own cloud.
Built to government security controls: We build to SOC 2, CJIS Security Policy and NIST SP 800-171 controls. That's how we build, not a certificate, and we won't claim one we don't hold.
Usable by everyone: Every release is checked against WCAG 2.1 AA, so staff and residents using screen readers or keyboards can get their work done.

Evidence first, including the evidence against us.

We set the test before we see the result.
When we measure something, the rules are written down before we look at the data. Nothing gets tuned until it looks right.
We say where it falls short.
A ranking or a model comes with the conditions under which it shouldn't be trusted. An integration we list has a published test and the date it passed.
We keep what we know apart from what we assume.
Every claim carries its evidence. Unknowns stay marked as unknown.

Agencies answer for their decisions. The tools behind them should be able to answer too.

Talk to us

Tell us what your agency is working on. We read every message ourselves and reply by email.

A prime or subcontractor? Start a teaming conversation.