What we don't build, published before you pay us.
Most development shops find out what they cannot do halfway through your project, and then it becomes a change order. We would rather lose the deal on call one. Here is the actual list.
- EDI with retail or logistics trading partnersTrading-partner certification is a project of its own, on their timeline, not ours.
- HL7 / FHIR clinical record integrationClinical interoperability is a specialty. We will build the workflow around it and hand the interface to someone who does this full-time.
- Telematics and ELD fleet integrationsDevice-level integration with DOT compliance implications. Different risk profile, different vendor.
- Dealer, agency or custodian management systemsDMS, AMS and custodian APIs are the legacy systems of record for their industries. Building against them turns a 20% problem into a 45% problem.
- Licensed industry data platformsXactimate, CoStar, MLS and similar. The data is licensed, the terms are restrictive, and the cost is not ours to absorb.
- Real-time bidding or live auction mechanicsConcurrency and settlement correctness under load. This is not a feature, it is an architecture.
- Offline-first mobile with conflict resolutionTrue offline sync is one of the hardest problems in application development. Online-first with a sync queue is usually what you actually need.
- Payments as the productStripe checkout is in scope. Processing, escrow, trust accounting and payouts are a regulated business you would be starting, not a feature you would be adding.
- Anything requiring a trained machine learning modelUsing a language model is in scope. Training a model on your data to a required accuracy threshold is a research project with an unknown end date.
- SOC 2 or HIPAA attestation as a launch requirementWe will build to the posture. The formal audit is a months-long engagement with a third-party auditor and it cannot be compressed.
Does your business currently run on a spreadsheet, a group text, and a PDF that gets emailed? Then we can build it. Are you asking us to talk to a thirty-year-old system of record that somebody else controls? Then we can't — not at this price, not on this timeline, and we would rather tell you that now.
What is always in the build, in every project
Multi-tenancy. Real authentication and role permissions. Audit logging. Migrating you off the spreadsheet you are on today. These four appear in every single project regardless of industry, they are the least AI-tractable part of the work, and they are the reason a human development firm is involved at all. They are priced into every proposal. Anyone quoting you a build without them is quoting you a demo.
What to do if you're on this list
Book the call anyway. Almost every idea has a version that routes around the integration and still delivers most of the value — build alongside the legacy system, export to it rather than sync with it, and revisit once the software has proven it is worth integrating. That is a conversation worth forty-five minutes, and it is free.
Because the alternative is a change order in week six.
Every item on that list is something a development shop could accept, start, and then discover the hard way. That discovery has a name in this industry: it is called a change order, and it is where a $40,000 project becomes a $90,000 project while everyone acts surprised.
Our margin depends on AI covering the great majority of the build and a human development firm covering the rest. That ratio holds for records, roles, statuses and generated documents. It collapses the moment we have to negotiate with somebody else's thirty-year-old system of record. So we say no on call one, out loud, and we would rather lose the deal there than find out in week six.
If a shop has never told you what they cannot do, you have not finished the conversation.
Find out on call one, not in week six.
Forty-five minutes, free, and if your project is on that list we will say so before you have spent anything.
No cost. No commitment. The proposal is yours to keep.