Real operational problems.
Practical, custom-built software.

Consumer wellness & productivity software · ADHD and neurodivergent support

A task app people with ADHD couldn't stick with, and couldn't get their data out of

Most to-do apps assume people can plan ahead and follow through consistently. That's exactly what's hardest for many people with ADHD, so the usual pattern was a great first week followed by giving up once real life got in the way. On top of that, once someone put their tasks into this app, there was no way to move them back out to the apps everyone else uses, like Apple Reminders or Google Tasks. And every update to the app went straight out to all users at once, with no safe way to try a change first.

What we built: A task app designed around how ADHD actually works, including tools to reduce overwhelm, virtual co-working sessions to help people focus, habit tracking that doesn't punish a missed day, and a way to share progress with a therapist or family member. We also built a two-way bridge to popular task apps, and a safe way to test updates before they reach real users.

• Locked in one app → works with 6 popular task appsTasks can now move in and out of Apple Reminders, Google Tasks, Microsoft To Do, Todoist, Things, and TickTick, instead of being stuck with no way to leave.
• No safety checks → 26 automatic checks before every updateBefore, someone had to manually check that moving tasks between apps still worked, and that each person could only see their own data. Now a set of automatic checks does that every time, catching mistakes before customers ever see them.
• Updates went live instantly → updates are tested firstChanges used to go straight to the real app with no safety net. Now every update is tried out in a separate practice version first, with a plan to undo it if something breaks, before it reaches real users.
• No way to confirm an update worked → every release is checkedThere used to be no reliable way to know if an update actually worked once it went live. Now every release is checked to make sure it's running correctly before it's called done.

Mobile pet services · in-home dog training practice

A scheduling and records system held together by spreadsheets and paper

The business ran on a patchwork of Square Appointments, spreadsheets, and paper consult notes. Nothing accounted for drive time between in-home sessions, so the schedule routinely overbooked the trainer's day. Vaccination records for 80+ dogs lived in scattered email attachments and paper folders with no expiry tracking. Invoicing was manual and disconnected from payment collection, and clients had no way to see their own invoices, appointments, or documents.

What we built: One system that handles scheduling (accounting for drive time between visits), automatic invoicing with online payments, AI-assisted scanning of vaccination records, and a self-service portal where clients can check their own appointments and invoices.

• Scattered records → one searchable systemConsolidated $146K and 900+ invoices of historical business data into one place.
• Days → secondsVaccine records that used to take days to track down now get scanned and filed by AI in seconds.
• Cash/check only → live online paymentsCard payments have been processing online since June 2026.
• Constant phone calls → self-service portalClients can check their own appointments and balances, cutting out most "when's my appointment" calls.

Career services · independent career coach and early-stage HR-tech founder

A resume tool that couldn't be allowed to make things up

Job seekers were reusing one resume for dozens of different roles, even though each role really called for a different framing. Tailoring every resume by hand didn't scale, but off-the-shelf AI resume tools weren't a safe substitute either: they could invent achievements, invent job titles, or word things in a way that raised discrimination concerns, none of which the client could risk putting their name behind. Candidates also had no way to catch a deal-breaker requirement, like a needed clearance, degree, or location, before spending hours tailoring a resume for a job they were never going to get.

What we built: A job-search platform (resume tailoring, application tracking, interview prep, and a check for deal-breaker requirements) built around one strict rule: the AI can never say something about a candidate that it can't trace back to that candidate's own verified history.

• One resume for everything → a different, traceable resume for every jobEach tailored resume can be traced back to the exact line in the candidate's verified work history it came from, so instead of asking someone to just trust it, the app can show exactly where each claim came from.
• Hours wasted on jobs they didn't qualify for → filtered out up frontThe app checks for deal-breaker requirements before a candidate spends time tailoring a resume for a job they were never going to get.

Healthcare · privacy and compliance team

Privacy reviews that only one expert on the team knew how to do

Every time the company collected or used sensitive data in a new way, someone on the legal or compliance team had to fill out a privacy review by hand, in a spreadsheet or document. To do it correctly, they had to already know which privacy law applied (there are several, and they differ by country and state), remember every required question, and format the final report themselves. This took a long time, different people did it differently, and it was easy for an incomplete answer to slip through unnoticed.

What we built: Privacy Navigator, a guided tool that walks anyone, not just a legal expert, through a 7-step review. It figures out which privacy law applies based on where the data is from and what kind of data it is, warns about common mistakes as you go, offers to improve unclear answers, and produces a properly formatted, correctly named report automatically.

• 30 hours a week → 2 hours a weekTime the privacy and compliance team spends on these reviews each week.
• Regularly missing required information → almost neverThe tool now catches missing or incomplete answers before a review can be submitted.
• Applying the wrong privacy law → no longer happensThe tool automatically figures out which law applies, so reviewers don't have to guess.
• Formatting and naming mistakes on reports → no longer happensEvery report is generated and named automatically, the same way every time.

See how we help with HIPAA privacy & compliance →