Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.

We've shipped something we've been quietly excited about for a while: a way to talk to your Tenderlake workgroup in plain English, from inside Claude.ai.

See it in action

How to set it up

It works through a new MCP (Model Context Protocol) server we host at https://mcp.tenderlake.com/mcp. Once it's connected to your Claude account, Claude can reach into your live Tenderlake data on demand and answer questions from it, no SQL, no exports, no copy-paste from the UI.

What you can ask it

Once the connector is live in your Claude.ai account, the kinds of questions that work today look like this:

"What new notices landed in my folders this week?"
"Summarise the buyer contacts on this tender."
"Which of my AI Searches matched the most notices in the last 30 days?"
"Show me the recent Demand Signal hits across all my folders."

Behind the scenes, Claude is calling a coherent set of Tenderlake tools, folders, AI Searches, Demand Signals, the recent-notices feed, full notice details, and stitching the results into a plain-English answer. You get synthesis, not just rows: comparison across notices, summarisation across folders, and follow-up questions that build on what came before.

How it works

The connector authenticates via OAuth and rides on your existing Tenderlake API key. That means all your permissions, workgroup boundaries, and audit trails are exactly as they were; nothing is bypassed, nothing is duplicated. The connection refreshes itself silently in the background; you'll only need to reauthorize if you leave it idle for more than 90 days.

If you're a Tenderlake Admin, you manage the underlying API keys the way you always have. Everything Claude does on a user's behalf flows through the permissions attached to that key.

Why we think it matters

Three things, mainly.

Zero learning curve. Anyone in your organisation who can write an email can now query Tenderlake. They don't need to learn the UI, remember which filter is on which screen, or know how AI Searches differ from Demand Signals. They just ask.

Conversational synthesis. Claude doesn't only fetch data, it reads across it. Ask it to compare two tenders, summarise a buyer's recent activity, or pick out the notices most relevant to a specific capability, and you get a written answer rather than a results list.

Same data, new surface. Nothing about your Tenderlake data, scoring, or matching changes. This is the workgroup your team already trusts, exposed through a chat interface for the moments when chat is the right tool for the job.

A note on what this is - and isn't

We want to be straightforward: this first release is a demonstrator, not a finished product.

We've shipped a deliberately coherent set of tools that we think covers a lot of useful ground. But we're at the start of figuring out which conversations actually pay off in real procurement work, and which ones expose gaps we need to close. We expect the most valuable use cases to emerge from how customers actually use it, not from what we predicted at the whiteboard.

So we're asking for your help. Specifically, we'd love to hear:

Which questions did Claude answer well? Concrete examples are gold.
Which questions did it struggle with? Even more useful, these tell us where to invest.
What data is missing? If there's something in Tenderlake you wanted to ask about and couldn't, we want to know.
What workflows would you want to automate next? The chat interface is one surface; what should come after it?

Email us with examples, screenshots, transcripts - anything. The good, the bad, and especially the surprising.

Getting started

If you're an existing Tenderlake customer, the in-app explainer walks through both a short demo and the connection steps, with two videos to get you going. You'll find it under Dev → Claude when you are logged in to Tenderlake. The full history of what the underlying API can do (and what we've added recently) lives in the Dev → Changelog, and Tenderlake Admins can manage API keys at Dev → Manage API Keys.

We're looking forward to seeing what you ask it.

Tenderlake Claude.ai connector