Work
Engagements, written like the work actually happened.
Anonymized production AI work: claims intake AI automation for insurance, CRM hygiene agent for B2B SaaS, and what a live system looks like six months after handoff. Each write-up names the constraint, what we refused, the measurable outcome, and what transferred to the client’s team.
8-week engagement · Founder-led
Claims intake that does not stall at the inbox
62% less manual review
Claims intake AI automation for a mid-market insurance carrier: classification, extraction, and a human-in-the-loop queue. Wrong routing delayed a payout, so the model only routes what it can defend.
Read the engagement6-week engagement · Founder-led
A sales-ops agent for CRM hygiene
18 hours / week recovered
A CRM hygiene agent for B2B SaaS sales-ops: it reads Salesforce call notes, updates fields, and flags exceptions—built against real CRM rules, not a sandbox. Forecast category never went automatic; the agent does not invent a close date.
Read the engagementWhat the system does on a normal Tuesday
Six months after go-live, this production AI claims intake system runs without us. The ops lead changed a routing threshold herself—no ticket, no retainer.
Read the engagementHow to read these
We do not publish vanity logos.
A case study that only shows a percentage is advertising. Each write-up names the constraint, the approach, what we would not build, what transferred to the client’s team, and what we deliberately left alone.
If your problem does not look like these, that is useful signal. Bring the actual job. We will tell you if it is ours.
Tell us the job that already hurts.
We’ll tell you honestly if we can help and what it would take.