“We have an AI-built product, but we can’t confidently launch it.”
Last verified: October 2026· engagement
A Lovable, Bolt, v0 or Replit prototype gets a product in front of people quickly. Launching it to paying customers, or through an enterprise security review, is a different bar: auth, data boundaries, secrets, tests, deploys and rollback. The useful question is which of those your specific launch actually requires — and whether there is a launch at all yet.
Is it worth fixing?
It usually is when several of these are true:
- There is a committed launch: a date, a customer, a security review or an investor milestone.
- Real user data will flow through the app, or already does.
- Someone internal will own the product after launch.
- The gaps are known to exist but nobody can say which ones block launch and which can wait.
The decision, one question at a time
Each step links to the page that answers it. You can get most of the way without talking to anyone.
- 1
- 2
- 3
- 4
- 5
What a first engagement looks like
Free read
A written list of the gaps that matter most for your stated launch — auth, data boundaries, secrets, tests.
Paid diagnostic
A fixed-price readiness assessment against the requirements of your launch, with a prioritised, priced hardening plan. Fixed price, quoted after the free read.
Stop or go
If there is no committed launch or owner yet, we recommend waiting and tell you what to have in place first.
The accepted outcome is the test above: the production requirements agreed for your launch pass, and a deploy and rollback are demonstrated in your environment.The full acceptance criteria, exclusions and pricing terms are written out onthe engagement page, so you and procurement can see the finish line before any call.
Access and security
- Repository access once there is an NDA; a public demo link is enough to assess fit.
- No production credentials are needed for the assessment.
When we’d tell you not to buy
Our scarce resource is senior engineering attention, so we would rather say no early. We decline or refer when:
- It is a demo with no committed launch, customer or owner.
- The ask is to keep adding features rather than to harden for a launch.
- Nobody will own or operate the product after handover.
Not there yet? Start withProductionizing AI-built apps, The productionization playbook or Supabase production hardening.
The engagements behind it
AI-Generated Code: From Demo to Production
$25–80k fixed, or 10% of first-year production revenue uplift · 3–8 weeks
AI Code Security & Supply Chain
$20–50k audit, $5–12k/mo monitoring · 3–5 weeks
Questions
Do we need to rebuild it?
Usually not. Most launch blockers are auth, data boundaries, secrets, tests and deploys, which can be hardened in place. The assessment says plainly if a rebuild is cheaper.
We just have a demo. Should we talk to you?
Not yet. Use the playbook, and come back when there is a launch, a customer or an owner. We will say the same thing on a call.
Can you get us through a customer’s security review?
We can harden the app and produce the engineering evidence a reviewer asks for. Formal certification such as SOC 2 is a separate audit.
Related guides
Is your codebase ready for AI agents?
Agents don't fail randomly on a repo — they fail predictably on the same things that slow down human engineers: fuzzy boundaries, missing tests, undocumented conventions. Readiness is just those weaknesses paid down deliberately.
Agentic codingWhat is agentic coding?
Agentic coding is letting an AI coding agent plan, edit across files, run tools and tests, and iterate in a loop toward a goal — powerful for the first 80% of a change, and quietly expensive on the last 20% if you skip standards and review.
Find out whether it’s worth fixing
A few structured questions, no call required. You get a written decision brief back — including whether we think you should buy anything at all.
Start the diagnostic