Somewhere in California right now, a single mother is opening a past-due utility notice. The font is hostile. The math is worse. There is a phone number that, if she dials it, will lead to ninety minutes of hold music and a person who is not allowed to help her. This is the customer Promise built its company for - and the agency Promise built its software for.
The strange thing about Promise is that it looks, from the outside, like the most boring kind of B2B software. It is. It is also one of the most consequential companies to come out of Oakland in the last decade. The product handles government billing. The market is invisible until you owe it money. The mission is to make sure that one bad month does not become one bad decade.
The company sits at the intersection of three things American technology usually avoids: public utilities, public agencies, and the public. Its customers are city water departments, county courts, and state benefits offices. Its end users are the people those institutions have always struggled to serve well - and increasingly, struggled to serve at all.
The problem they saw
Public-sector billing systems in the United States are, in the polite phrase used by consultants, mature. In the less polite phrase used by everyone else, they are antiques. A meaningful percentage of municipal water utilities still run on software older than the people using it. Late fees compound. Shut-off notices print on schedule. The penalty structures are designed for an America where most households had a savings cushion - which is no longer the America most households live in.
Phaedra Ellis-Lamkins and Diana Frappier did not come to this problem through a software lens. Ellis-Lamkins spent her career in labor organizing and economic justice work. Frappier spent twenty years as a criminal defense attorney. They met through Van Jones, worked together at Green For All, and watched, repeatedly, the same scene play out: a working family loses everything not because of a single catastrophe but because of a small bill that became a bigger bill that became a court date.
The diagnosis was unromantic. The systems were not malicious. They were simply built for one kind of payer and ignored the rest. So Ellis-Lamkins and Frappier did something that, in retrospect, looks obvious and at the time looked deranged: they decided to build software for the rest.
The founders' bet
The bet was that government would pay for empathy if you could prove it also produced revenue. This is not the way empathy is usually pitched. It is, however, the way procurement officers think. And Ellis-Lamkins, having spent fifteen years organizing alongside public officials, knew how procurement officers think.
Promise joined Y Combinator's Winter 2018 batch as one of the cohort's least Silicon Valley-shaped companies. The standard YC arc - consumer app, viral growth, monetize later - did not apply. Promise's growth would come one city contract at a time, each one negotiated against the longest sales cycle in technology. The upside was that once a city signed, it tended to stay signed.
The funding came in stranger places than usual. First Round Capital led an early seed. Jay-Z's Roc Nation participated. Kapor Capital, which has been writing checks for mission-aligned founders longer than the term was fashionable, came in. In 2021 the Series A closed at $20M - one of the largest rounds ever raised by a Black, women-led founding team. A year later they extended it by another $25M.
The product
Promise's platform now spans four named modules. Each one targets a specific point of failure in how public money moves.
PromisePay
The flagship. Flexible, interest-free installment plans residents can set up online without a phone call, a notary, or a court appearance. Agencies recover revenue they were otherwise writing off. Residents keep the lights on. The economics work because partial recovery beats nothing, and because shut-offs are themselves expensive.
PromiseBenefits
Software that moves emergency relief and benefits money out the door faster - the kind of system that, during the early pandemic, would have prevented several of the slow-motion administrative disasters that we all read about.
PromiseVerified
Income verification, automated. Faster eligibility decisions for the assistance programs whose front doors are usually a fax machine.
PromiseAudit
AI-powered fraud and waste detection - the rare instance of "AI in government" that is not a press release.
A short, useful timeline
The proof
Promise is unusual among mission-driven software companies because its mission claims are testable. Either the city's revenue numbers improve or they don't. Either fewer residents lose service or they do. The contract renewals - the most honest signal in B2B software - have continued.
Promise, by the rough numbers
The team has grown to roughly two hundred, drawing engineers from Google, Stripe, and Palantir and pairing them with operators who have actually run public benefits programs. This combination - Silicon Valley scale skills, plus people who know what a Medicaid renewal cycle actually feels like - is rarer than it sounds.
The mission
Promise's stated mission - modernize and humanize government payments - is, on paper, the kind of phrase that loses meaning the moment it leaves a deck. In practice it has a sharper edge. The company has been explicit that it will not work on programs designed primarily to penalize the people they touch. Predatory collections work is off the table. The product roadmap is constrained by what the founders are willing to ship.
This is, depending on your priors, either a clever positioning or a real constraint. Both can be true. What is unambiguously true is that it has made Promise a magnet for a particular kind of employee - the senior engineer with a public-interest itch, the policy operator tired of writing memos no one acts on - that competitors have not been able to recruit.
Ellis-Lamkins has been blunt that she did not come to Silicon Valley to make a slightly faster Twitter. She came because the most powerful tool she had ever seen for fixing things at scale was code, and the things she wanted fixed were not being fixed by anyone else.
Why it matters tomorrow
The next decade of American public services is going to look unlike the last one. Climate-driven utility crises - heat waves, water shortages, grid failures - will push more households into arrears more often. Federal benefit programs will keep churning through new compliance regimes. Cities will keep being asked to do more with budgets that compound at the rate of inflation while their costs compound faster.
In that environment, the agencies that can move money - in and out - quickly, accurately, and humanely will function. The ones that can't, won't. Promise has positioned itself as the picks-and-shovels vendor for the agencies that intend to function. It is not a glamorous bet. It may turn out to be a very good one.
Somewhere in California right now, a single mother is opening a past-due notice. This one is different. It has a link. The link works. The plan adjusts to what she can pay this month, this week, today. No phone call. No notary. No court date. She closes the laptop and goes to bed. That, in the most undramatic way possible, is what Promise built.