Commissary Club: An AI Job Network for People Leaving Prison
Richard Bronson has launched Commissary Club, an employment-focused social network for formerly incarcerated people that uses AI to help with job and housing applications.
Bronson does not hide his own prison history. He worked for Stratton Oakmont in the 1980s, became a partner, then left after a year to start his own firm in South Florida, according to TechCrunch. Stratton Oakmont was run by Jordan Belfort and collapsed in the mid-1990s amid misconduct accusations. Bronson later spent a couple of years in federal prison for violating securities laws, the rules that govern buying and selling investments.
The name comes from the store inside a prison where inmates buy snacks and toiletries. The current team is listed as two people: Bronson and co-founder Roman Kissin. Kissin previously worked for eBay and IBM and served as chief technology officer of LexisNexis between 2022 and 2025.
Users build a profile on the platform and use AI help to complete job and housing applications. The first 10 applications are free. After that, Commissary Club charges $149 for 10 applications or $14.99 each.
Bronson is also the founder of 70 Million Jobs. When TechCrunch covered Commissary Club on October 29, 2020, the emphasis was on community through topic-specific clubs. The company is listed as a Startup Battlefield company, is based in Los Angeles, and debuted in beta with thousands of people on its wait list. Bronson now describes it as an AI-native social and identity platform, meaning AI is built into its core functions, that creates a new trust layer, a shared system for proving reliability, for people excluded from traditional systems.
The broader context here is identity and verification. Hiring platforms already handle resume parsing, software that reads resumes for key details, skills matching and background checks. They assume a continuous work history, a stable address and references that respond to email. For people reentering after incarceration, those assumptions break. A system built for gaps in employment, restricted documentation and housing instability has a different onboarding problem to solve. It must establish credibility without relying on the standard signals employers use to filter applicants.
Looking at what this means for builders, the choice to combine a social graph, the network of connections between users, with a transactional workflow, the step-by-step process for completing applications, stands out. Community features can reduce isolation after release and keep users returning. Application assistance provides the action that brings in revenue. The pricing places cost on the job seeker rather than the employer. That inverts the usual marketplace model where employers pay for sourcing. It may reflect the difficulty of getting employers to pay for a talent pool they have historically screened out, but it also creates friction for users with limited income.
In my view, the bet worth watching is whether a dedicated trust layer can port across contexts. If work history, housing applications and peer vouching can be structured into a persistent, portable profile, that profile becomes more useful with each successful placement. Technology has, over the long arc, lowered the cost of matching people to opportunity. Applying that pattern to reentry is a practical use of AI assistance, less about automation for its own sake than about translating lived experience into forms hiring systems can read.
