Predictive Sales: How AI Is Spotting Public Tenders Before They Happen

There's a scene I lived through in about 2009. I was pitching a council contract for an SMS marketing platform. Turned up, did the dog-and-pony show, felt quietly confident. Lost on evaluation. Afterwards, over truly terrible biscuits, the procurement officer told me something I've never forgotten: "We had a preferred direction before the spec was written."

Translation: the incumbent had already shaped the room. I was a supporting actor in someone else's play.

That dynamic hasn't changed. What has changed is that a machine can now tell you when the next play is being cast.

The Record Was Always There

Every UK public contract leaves a paper trail. Find a Tender, Contracts Finder, Public Contracts Scotland, Sell2Wales, the old TED archive: who bought what, from whom, and when it ends. The problem was never secrecy. It was assembly. Cross-referencing the same deal across four portals and resolving a supplier spelt three different ways isn't analysis. It's archaeology.

Love Me Tender does the digging. Its models read 1.6 million official procurement records, rebuilt every night, to spot contract end dates, extension options, reordering rhythms and typical contract terms. Then they turn the lot into something no sales director has ever had handed to them on a plate: a call sheet of public buyers likely to come back to market, with a forecast window and a confidence score on every line.

The Numbers That Matter

Here's the bit that made me sit up. Love Me Tender doesn't hide behind a vague "AI-powered" badge. It publishes its back-test, measured on holdout periods the model never saw.

• 62% of its 5,340 High-confidence forecasts were matched by the same buyer publishing a notice in the same category in or near the predicted window.
• Medium forecasts hit 22% (of 83,532); Low ones 12% (of 124,062). Even a "Low" lead is a buyer with a real buying signal and a phone number.

Then there's the new £295 AI Prediction layer, tested across 356,017 historical forecasts from 2023 to 2025. It flagged more than five times as many leads as top priority (54,425 against 9,925) and still lifted the hit rate from 58% to 63%. The share of forecasts landing within three months of the actual tender jumped from 50% to 61%, and the timing gap tightened from 3.0 months to 2.2.

Read that again. Five times the hot leads, and more accurate. In sales terms, that's a bigger pipeline that converts better. Nobody gets both.

Love Me Tender is careful to say a "match" means the same buyer advertised the same category within the window, not necessarily the identical contract. Fair enough. A buyer coming to market for exactly what you sell, within a few months of when you were told they would, is a sales call. Not a mailshot.

How It Works: Three Steps to a Call List

The how-it-works process is refreshingly blunt:

1. Tell it what you sell and where. Type "window blinds" or "fire alarm servicing"; it maps your words onto the categories public buyers actually use.
2. See who's due to buy. Live tenders, announced opportunities and predicted re-tenders, clearly labelled, so you know what to bid for now and what to prepare for.
3. Put names to them. Buyer names, current suppliers, published procurement contacts, phone numbers and websites where available, plus the evidence behind every forecast.

Search is free. You only pay when you want the names. The £99 UK-wide report gives buyer-level detail across every UK region for the next 12 months, plus a look at what's coming in months 13 to 24. It's delivered as a PDF and spreadsheet. No subscription, no sign-up, and a refresh within a year costs half price. Prefer your patch? The regional report is £49. Want the AI ranking of who to call first? That's the £295 tier. You can inspect real sample reports before spending a penny.

When the Incumbent Goes Bust

Here's where it gets properly clever. The incumbent advantage is structural: they know the stakeholders and helped write the last spec. But incumbents aren't immortal.

Love Me Tender doesn't just tell you who holds the contract. It checks the incumbent's company health, flagging their trading status and overdue filings using Companies House data. A supplier that has gone into liquidation, been dissolved or has stopped filing its accounts is a supplier that may not be there to deliver, let alone renew.

So picture it. A Midlands council's cleaning contract ends next spring. The incumbent has held it for years and would ordinarily walk the renewal. But the report flags that the company's accounts are overdue, or that it is no longer trading. That buyer is suddenly facing a forced re-tender, possibly an emergency one, with no comfortable favourite. That's not a long shot. That's an open door, and you're the one who knows it's open.

Why Reactive Bid-Chasing Loses

By the time a tender is published, the buyer's picture of the solution is already formed. You're writing a bid to fit a box someone else helped design. The Procurement Act 2023 nudged things towards transparency and SME access, but it hasn't changed the physics.

TenderSpotters puts that logic in plain words: public sector contracts, spotted before they're advertised. It works from contract end dates, renewal options, buying cycles and early notices, with a confidence score on every forecast. The point is to get you talking to the buyer in the window when a conversation is still a conversation, not an evaluation.

Honest Limits

Credit where due: the coverage has gaps and the makers say so. Northern Ireland's below-threshold notices aren't yet covered, as eTendersNI has no data feed. Contracts Finder keeps only one supplier per multi-supplier award, which blurs the incumbent picture on frameworks. And contracts quietly extended or rolled over will always trip up a model trained on history. Forecasts are estimates, not guarantees.

But imperfect intelligence used early beats perfect intelligence used too late. Every time.

The Founder's Verdict

I built something adjacent to this in 2007. The data sourcing alone nearly finished me. Now the data is public, structured and rebuilt nightly, and the economics have flipped.

For UK SMEs in facilities, cleaning, IT, fire and security or care, the playbook is simple. Stop refreshing portals for live tenders. Start working contract expiries and failing incumbents, 12 months out. Run a free search on Love Me Tender, grab the £99 UK-wide report, or read how TenderSpotters works. Then pick up the phone before the clock starts.

The data has always been there. Someone has finally made it sell.