|

Founder-led and owner-led businesses, typically 5 to 50 people, with revenue coming in and operations starting to feel like the bottleneck. The vertical doesn’t matter much. The shape does.

Initial builds typically run $5K–$15K, depending on scope. They’re usually a focused first project like migrating a base, rebuilding a workflow, or getting your data into a structure your team and AI can both work with. From there, the system grows through retainer work and follow-on projects rather than as one big upfront build. The exact number is locked at the end of the 1–2 week discovery phase, before any build work begins.

No. Most clients start with a focused first project rather than a full rebuild. That might be migrating one base, building a single internal tool, or restructuring data so you can use it. From there, the system grows through follow-on projects and retainer work as the business needs more. You’re not locking into a big upfront engagement to find out if it works.

4–8 weeks end to end, in three phases. Discovery (1–2 weeks): interviews, data audit, written scope, fixed quote. You can stop here if you want. Build: database, automations, AI agents, internal tools. Your team is running live work through it before handoff. Handoff includes training and runbooks. From there: take ownership and run it yourself, or continue on retainer.

Your existing system is the spec. We don’t ask you to rebuild what you already know works. Discovery captures every table, automation, and integration you currently use. The build translates that into a system you own. Your existing setup stays running until cutover, so nothing breaks while we work.

The promise of low-code (Airtable, Softr, Monday) was that anyone could maintain their own system. In practice, once a system gets complex enough, non-technical users stop touching it. The platform that was supposed to put them in control ends up locking them out, because every change feels risky and the entry point is higher than it looks. AI-native systems flip this. Your team describes what they want in plain language (‘add a status field to client records and let me filter by it’), the AI drafts the change, and you or we review before it ships. The result: your team makes more of their own changes than they ever would inside a no-code platform. Bigger structural shifts (new integrations, new AI agents) come back to us, or any developer who works in Next.js and Postgres.

AI in the system means your team can ask things like ‘show me which campaigns are slipping and why,’ ‘draft an invoice memo from last month’s project work,’ or ‘find every conversation with this client that mentioned pricing.’ The AI reads from the same records your team does (not pasted context, not yesterday’s CSV) and answers grounded in what’s actually happening. It can also write back to records when work is done, with you or us reviewing before changes commit.

Everything. The code (as a Git repo we hand over, not a zip), the database (on infrastructure you control), the documentation, and the credentials. Retainers run month to month with 30 days notice. There’s no penalty for leaving. If you walk, we do a clean handoff: docs sync, credential transfer, and a walkthrough for whoever takes over.

Where your data lives depends on your preference. Some clients want it on infrastructure they own and control: typically a Postgres database on Supabase, AWS, or self-hosted, with credentials handed over at handoff. Others, especially in an ongoing retainer relationship, prefer us to host it on our own infrastructure. AI model calls go to providers like OpenAI or Anthropic, and your data isn’t used for training under their API terms. If you have stricter requirements (HIPAA, NDA constraints, specific data residency), we configure for those during discovery.

Yes. Every system we build uses a stack any competent developer can pick up: Next.js, Postgres, and standard tooling. Documentation, runbooks, and the deployment setup are part of every handoff. If you want to bring in your own dev, hire in-house, or move to another firm, the system is ready for it. That’s a deliberate design choice, not a contingency.

Two common paths. Most clients continue on a monthly retainer that covers ongoing maintenance and new builds. Tiers run from steady maintenance through fractional capacity for active development. For clients who don’t want a retainer, we also do ad-hoc projects: scoped, priced, shipped. You can also bring in your own developer or another firm. The system is built to be extended by anyone who works with Next.js and Postgres (part of any competent developer’s repertoire), or by anyone willing to dig in themselves and use AI to draft and troubleshoot the changes.