A surprising number of UK businesses commission custom software from a small studio, get most of the way through, and then suddenly cannot reach anyone. Sometimes the studio went under. Sometimes they pivoted. Sometimes they just stopped replying to emails. Whatever the reason, this is a recoverable situation — but the first 14 days matter. Here’s the practical UK playbook.

Step 1: do you actually own the code? (UK contract law)

First thing to check: the IP assignment language in your SOW or master services agreement. UK contract law does not automatically transfer copyright in commissioned software to the commissioning party. Without explicit assignment, the agency owns the code and you have an implied licence to use it — which usually does not cover modification by a third party.

Most reputable agency contracts have clear assignment language (‘all IP in the deliverables is hereby assigned to the client on payment’). Cheap contracts and amendments-via-email often don’t. If your situation is the second, there is still a path forward — see step 2 — but the order of operations matters and a 30-minute call with a UK IP solicitor before you do anything else is cheap insurance.

Step 2: getting access to the deployment, the data, the domains

Operationally, you need three things back: the production deployment (servers, hosting, container registry), the data (databases, storage buckets), and the domains (DNS, SSL certificates, email). These often live in the agency’s accounts, not yours. Whether you have legal access depends on whose name is on what — and that’s usually a mess.

Practical move: list every credential, every domain, every cloud account where your build lives. For each, identify whose name is on the account. Where the name is yours or your company’s, you can recover access via the standard ‘I forgot my password’ flow. Where it’s the agency’s, you’ll need them to transfer ownership. Where they’ve disappeared, the cloud provider’s official transfer-of-ownership process applies — this exists for AWS, GCP, Azure, and most UK domain registrars, and it works, but it is paperwork-heavy and takes 2–4 weeks.

Step 3: what’s salvageable versus what isn’t

Once you have the code and the deployment access, the next question is whether what’s been built is worth keeping. We see three patterns. First: the build is mostly sound, the agency was just bad at communicating, and 70–90% of the code can be picked up by a new team. Second: the build has architectural decisions that look reasonable but cost the next team 2x to maintain — a different framework choice, a missing test suite, no CI. Third: the build is genuinely worse than restarting from scratch, usually because the previous team did not have the seniority the work needed.

Telling these apart is what a rescue triage week is for. We’ve come into all three situations. In our experience, pattern 1 is the most common (about 60% of cases), pattern 2 is next (~30%), pattern 3 is rare but real (~10%). The honest answer matters, because rescuing pattern 3 ends up more expensive than restarting.

Step 4: the cost of restart vs. resume

Once you know which pattern you’re in, the cost calculus is concrete. Resuming a pattern-1 build with a new team is usually 30–60% of the original budget — the new team is paying a context tax, not a build-from-scratch tax. Resuming a pattern-2 build is 60–100% of original budget, because the new team has to fix architectural debt while shipping. Restarting a pattern-3 from scratch is 100–120% of original budget but with much higher confidence in the outcome.

These ratios are real. We track them across our rescue engagements. The decision is rarely about pure cost — it’s about cost-vs-confidence. Most clients pick the resume path even when restart is cheaper, because they have a deadline and ‘we tore up the build and started over’ is a hard story to tell stakeholders.

Step 5: who to bring in next

Two options. Bring the work in-house, hiring 1–3 engineers depending on scope. Six months minimum to ramp them, and only realistic if the company has a long-term software roadmap that justifies the team. Or bring in another agency or studio to take it over. Faster, more expensive per month, faster ship date.

We do the second often — see project rescue. We’ve come into builds where the previous agency vanished and have shipped to production within 8–12 weeks, depending on what was rescued. Book a triage and we’ll come back inside a week with an honest read.