Closing a company means writing off years of work. The code history, tickets, docs and internal knowledge you built can still have value. AI companies license this kind of historical business data to train and test their models.
Start a conversationAmka Data works with founders of software startups that are closing or have recently closed, mainly in the UK and US. If your data is likely to interest a buyer, I make the introduction, so the work your team did earns something before the company is gone.
We look at which systems your company used and whether the data is likely to interest buyers.
If it's a fit, I introduce you to a buyer. They assess the data, handle de-identification, and make you an offer.
The licence is between you and the buyer. You stay in control throughout, and nothing moves without your agreement.
Buyers are most interested in real working history: how problems were discussed, planned, built and fixed. If your company used tools like these for a few years, there may be something worth looking at.
GitHub, GitLab, Bitbucket: commits, pull requests, reviews
Jira, Linear, Asana: issues, sprints, project history
Notion, Confluence, Google Drive: specs, wikis, processes
Slack, Teams: working conversations between the team
Zendesk, Intercom: customer issues and how they were solved
HubSpot, Salesforce: pipeline and account workflows
Analytics events, logs, design files
Ask. Specialised datasets can be valuable too.
Customer names and personal details are stripped out by the buyer before any data is used. Companies in regulated or sensitive sectors are only considered with their own lawyer's sign-off.
I believe data matters. For years, the Slack threads, tickets and code of a closed company were worth nothing once the doors shut. They sat in an archive until someone deleted them.
AI has changed that almost overnight. The records of how real teams worked are now some of the most useful material for training and testing AI, and companies are paying for them. I want to play a small part in making sure founders get something back from that work, instead of it quietly disappearing.
I'm based in Nairobi and work with founders in the UK, US and East Africa. I'm also Kenya Lead at Intro Africa. Amka means "wake up" in Swahili: waking up data that would otherwise go to waste.
It depends on how much history there is, how complete it is, and what buyers want at the time. Published buyer ranges start in the tens of thousands of dollars. Some companies won't qualify, and I'll tell you honestly if yours is unlikely to.
Companies that license business data to AI developers. I talk to several, so I can introduce you to whichever fits your data best. You'll know exactly who they are before you share anything.
Buyers de-identify the data, removing names, emails and customer details, before it's used. You agree the terms of the licence, including what's excluded. If something is too sensitive to share, it stays out.
Nothing upfront. If a sale happens, any commission is agreed with you in writing before I make the introduction. If you decide not to go ahead, you owe nothing.
Not if you move quickly. Once a company is dissolved, its assets pass to the Crown, so a sale needs to complete before then. Dissolution usually follows about two months after the Gazette notice. The earlier we talk, the more options you have.
A short email is enough to start. Tell me what your company built and which tools you used.
paul@amkadata.com