Skim AI is built for a specific gap in the public procurement market. Traditional tender alert platforms are good at surfacing volume — more and more notices — but they leave the hard work (shortlisting, deciding whether to bid, drafting a response that scores, and learning from loss notices) with the supplier. Specialist consultancies do that work but at consultancy prices. Skim positions itself between the two: software that actually does the work, with a veteran bid team behind it, and DPS positioning it as an AI-native platform for UK and EU SMEs.
The product is organised around a set of agents. Opportunity Discovery acts as a bid researcher scanning every portal; Bid Analysis plays bid/no-bid analyst and reports decisions in minutes rather than the roughly 30 to 40 minutes of manual review it replaces; Competitor Analysis tracks rivals; Buyer Intelligence profiles the buyer and maps stakeholders to decision-maker contacts; Business Strategy and Partnership Intelligence cover strategy and consortium mapping. Tony, the flagship agent, is trained on twenty-plus years of UK and EU award notices, tender documents and supplier histories, and the platform claims 90% classification accuracy for tender matching against an industry average it cites as 8%.
On top of discovery, Skim offers pre-tender intelligence such as pipeline forecasting that surfaces contracts 12 to 18 months before they go to tender, a contract expiry tracker and framework renewal alerts, and incumbent lookup. On the writing side it retrieves tender files automatically from UK and Irish buyer portals, organises the data needed to complete a bid, advises on pricing from award values on similar contracts, and drafts responses. It also syncs to CRMs and exposes a machine-readable pricing page plus an MCP integration so the agents can be used inside Claude, ChatGPT, Copilot and Gemini. Skim publishes original research on UK procurement and offers a win rate calculator, and its public pricing runs from a free tier through Starter, Growth and Enterprise plans.
