How it works, in the open
Coverage & methodology
Maintained by BidSquirrel · last reviewed July 2026
We would rather tell you exactly what we cover and how we score than leave you guessing. Here is the honest version: the sources we scan, the ones we do not, how often the data refreshes, and how every tender is assessed.
Coverage map
Exactly what we ingest every day, and what sits outside our scope - so you always know where the feed begins and ends.
In scope - scanned daily
Find a Tender (Central Digital Platform)
The UK government's central service for new above-threshold tenders under the Procurement Act 2023, including Welsh above-threshold work.
Contracts Finder
England's register of public contracts and lower-value opportunities, plus notices from procurements begun under the older PCR 2015 rules.
Public Contracts Scotland (PCS)
Scotland runs its own regime; almost every regulated Scottish opportunity is advertised here.
ProContract council portals
The local-authority systems carrying the below-threshold and quick-quote council work most national-only tools never see.
YORtender
The Yorkshire & Humber council quote-portal network, carrying the same kind of below-threshold local work as ProContract.
Atamis (NHS)
The NHS Health Family's own procurement system, for NHS trust and integrated care board opportunities outside the mainstream notice services.
Paused - historical notices only
Sell2Wales
Source paused — Welsh Government OCDS API unavailable since Mar 2026. Historical notices remain browseable.
Out of scope
Outside Great Britain
BidSquirrel covers Great Britain - England, Scotland and Wales - only. Opportunities procured under separate arrangements elsewhere sit outside our scope.
Private-sector tenders
We cover public-sector and public-body procurement, not private commercial RFPs.
Coverage spans Great Britain - England, Scotland and Wales, from the national portals down to the low-value local-authority work that often never reaches a national site.
How often the data refreshes
Our connectors run every day, pulling new and updated notices from each source and re-scoring them against the businesses on the platform. New coverage and re-scored matches surface in your dashboard automatically. Live counts shown on the site are read from the database at page load, not hard-coded.
How each tender is scored
Every tender is assessed by an AI built and regression-tested as a UK procurement specialist - not a general-purpose chatbot. It reads each notice under the right regime (Procurement Act 2023, the legacy rules, NHS Provider Selection Regime, and the devolved Scottish and Welsh regimes), then scores the opportunity against your business, scale, location and growth goals:
- a genuine fit you should seriously consider
- workable, with caveats worth weighing
- an on-strategy stretch aligned to your stated growth goals
You always see the reasoning behind a score, never a black-box number, so you stay in control of the call. For the full detail on how the AI is built and tested, see how our AI works.
Sector expertise, where we have it
On top of the procurement-law reading, BidSquirrel reads each notice through a sector-experienced lens. For the fields we cover deepest, it applies the standards, accreditations and evaluation patterns a specialist bid lead in that field would recognise - raising the questions an experienced practitioner would ask, rather than matching keywords. Five sectors are covered in depth today:
Construction
The statutory health-and-safety duty regime (CDM Regulations 2015), the Common Assessment Standard prequalification route, fire-performance evidence for external walls (BS 8414), and the workmanship series (BS 8000).
Cleaning & Facilities Management
Output-based cleaning-quality measurement (BS EN 13549), planned-maintenance task libraries (SFG20), operative competency (BICSc), substance risk assessment (COSHH 2002), and the NHS National Standards of Healthcare Cleanliness 2025 on clinical sites.
Healthcare & Care
CQC Regulation 12 (safe care and treatment), the Data Security and Protection Toolkit for services touching NHS patient data, and the clinical-risk-management standards for health IT (DCB0129 and DCB0160).
IT & Digital
Cyber Essentials and Cyber Essentials Plus, information-security management (ISO/IEC 27001), and the accessibility level public-sector digital services must meet (WCAG 2.2 AA).
Security Services
SIA licensing matched to the actual licensable activity, security screening to BS 7858, and static-guarding practice (BS 7499).
Outside these five, the AI stays capable but general - it never fakes depth it does not have. That partial-but-honest coverage is deliberate: we would rather cover a handful of sectors properly than claim expertise everywhere.
The standards it references are named, current and re-checked - each verified against its primary source, with a monthly currency check so it never leans on a standard that has since lapsed or been superseded. And it stays complementary to the procurement-law reading above, not a replacement for it: the regime tells you which rules govern the award (Procurement Act 2023, the legacy rules, NHS Provider Selection Regime, the devolved regimes); the sector lens tells you what an experienced practitioner in that field would probe. It raises what a specialist would raise - you still answer from your own delivery knowledge.
A worked example, end to end
Talking about scoring is easy; here is one run through it. The tender below is an anonymised example - the figures and wording are representative, not a real notice - but it shows exactly what you would see, and why.
Example · anonymised
Grounds maintenance and landscaping - a Midlands district council
- SourceContracts Finder, published under the Procurement Act 2023
- Value£1.8m over four years (one year, plus three one-year extensions)
- BuyerA district council, scoring 60% quality / 40% price
- The business profiledA 14-person grounds and landscaping firm, ~£900k turnover, based 18 miles away, wanting to move from sub-contracting into prime council contracts
Eligibility checks first. Before any fit score, the notice is read for the pass/fail gates that decide whether a bid is even worth writing. Here is what that surfaces:
ISO 9001 quality management
Asked for, but the notice accepts an equivalent documented quality system - the firm's existing process maps qualify, so this is met without the certificate.
SSIP / CHAS accreditation
Required for health and safety pre-qualification. The firm already holds CHAS, so this clears.
Employer's liability insurance ≥ £10m
The firm currently carries £5m. This is a hard gate - the bid cannot proceed until cover is raised, though that is a phone call to a broker, not a barrier to entry.
Minimum annual turnover £600k
The notice sets a turnover floor of roughly two times the annual contract value. At ~£900k the firm is comfortably over it.
Then the verdict and the reasoning. With one gate flagged but fixable, the work itself a clean match, and the buyer within sensible travel distance, this scores:
Strong - with one thing to sort first
The scope sits squarely in the firm's core trade, the value is the right size to step up from sub-contracting to prime, and the site is within an easy operating radius. A 60/40 quality-led evaluation rewards a credible method statement and local track record over the lowest price, which suits a firm that competes on service rather than on being cheapest. The single caveat is the insurance gate - so the score is “Strong”, not “you have already won”: it is a genuine fit worth pursuing, provided employer's liability cover is raised to £10m before submission.
Had the firm been £200k under the turnover floor, the same notice would have come back Fair- workable, but only via a partner or a parent-company guarantee. Had it been a £30m county-wide contract well beyond the firm's current scale, it would have been flagged Development: off today's capacity, but on-strategy if growth into larger prime contracts is the stated goal.
Where a human still decides
The example above does the triage - the part that otherwise eats an afternoon per notice. It does not make the decision for you, and it is worth being plain about the line:
The bid / no-bid call is yours
A "Strong" score says this is worth your time, not that you should drop everything. Your pipeline, your capacity this quarter and your appetite for the buyer are things only you can weigh.
We draft the bid with you, not for you
BidSquirrel helps you structure the response, scores your draft and feeds back where it is thin - but the evidence, the references and the final words are yours. It will not invent case studies, fabricate accreditations or ghostwrite a submission you cannot stand behind.
We help you price it, but you own the number
Price and value-for-money can be most of a public score, so we build a costed breakdown from the scope, recommend a price band and write the value-for-money case - working from your own rates or clearly-labelled estimates you confirm. We read recent comparable award values to sense-check the band, but we present them as comparables, never a precise market rate, and we never hand you a single oracle number. Every figure is yours to set and defend.
Borderline eligibility needs a human read
Where a requirement turns on the exact wording of the notice - an "or equivalent" clause, a contested turnover test, a framework lot definition - we flag it and point you to the official notice rather than ruling it in or out for you.
We do not verify your figures or submit on your behalf
We never file a bid, never sign anything, and cannot confirm your insurance, certificates or accounts are current. Always check each opportunity, deadline and requirement against the official notice before you act.
Where the data comes from
The tender notices we aggregate are public information published by contracting authorities. This service contains public sector information licensed under the Open Government Licence v3.0. We aggregate and add guidance on top; we are not the publisher of the underlying notices.
Accuracy and limits
BidSquirrel is decision support, not legal or procurement advice, and not a substitute for the official notice. Sources can change their data, and AI guidance can be wrong - so always verify each opportunity, deadline and requirement against the official notice before you act. Our Terms set out the full position, and our Privacy Policycovers how we handle your data (we never use your data to train AI models; an ITT you add is the buyer's common tender pack, kept as a shared reference for that tender so other bidders don't re-upload it, while your own answers, drafts and evidence stay private to you).
How we check our own scoring
Anyone can say their matching is good. What follows is the method we grade ourselves against, so you can judge the method rather than take the adjective. We publish it whether or not the result flatters us.
What we grade against
Scored matches are sampled and graded on eight things, each of which can fail independently:
- Regime correctness. Is the notice classified under the right rules - Procurement Act 2023, below-threshold, a framework call-off - and does the reasoning follow from that classification?
- Eligibility gates. Where a hard requirement exists and the supplier cannot meet it, is it named as a blocker rather than softened into a caveat?
- Blanks are not penalties. A field you have not filled in is an unknown, not a mark against you. Reasoning that treats a blank as a negative is a defect, and we count it as one.
- Scale. Is the contract size judged against what you have actually delivered, rather than against the value alone?
- Geography. Is the delivery location real, and inside the area you told us you cover?
- Calibration. Do the scores separate? A page where everything is a 6 is a failure even if no single score is defensibly wrong.
- Evidence honesty. Does the reasoning cite the notice, or does it assert something the notice does not say?
- Coherence. Does the verdict follow from the reasoning given? A Strong verdict under three stated blockers is incoherent regardless of which is right.
How a defect is traced
A grade on its own changes nothing. Every defect is traced back to the specific cause - a clause in the scoring prompt, a branch in the regime classifier, or a hard gate - so the fix is to the mechanism rather than to the individual score. Re-grading the same sample after a change is how we tell a fix from a coincidence.
What we will not claim
We do not say the system is “trained on” your data or that it “learns from” you: it does not, and your data is never used to train a model. We do not publish an accuracy percentage, because a single number across wildly different trades and contract sizes would mean very little and would be impossible for you to check. And we do not impute a contract value a buyer has not published - where a notice states no value, we show “Value not stated” rather than estimating one and presenting it as fact.
Where a score depends on something your profile does not tell us, we say so on the match itself rather than quietly guessing. That is the honest version of a quality claim, and it is the one we hold ourselves to.
How we keep our guidance accurate
Our guides are written and maintained in-house and checked against primary sources - legislation.gov.uk (the Procurement Act 2023 and the legacy Public Contracts Regulations), gov.uk guidance, and the official portals (Find a Tender, Contracts Finder and Public Contracts Scotland). Every guide shows the date it was last reviewed, and we re-check the set on a regular cadence and whenever the rules change - a threshold refresh, a new Procurement Policy Note, or a change in how a regime applies.
It remains decision support, not legal advice. Where a point turns on a specific contract or a borderline legal question, we say so and point you to the official notice or qualified procurement counsel rather than present it as settled.
Who maintains this
BidSquirrel is built and run by Orcana Ltd (company no. 17263330), registered in England & Wales and with the ICO (no. ZC171438), based in Hull. It was built by people who have run and evaluated public-sector procurements - which is why the scoring works like an evaluation panel rather than a keyword search. Questions about our coverage or method? Email hello@bidsquirrel.co.uk.
See it scored to your business
Describe your business in plain English and see your top matches scored in minutes - free, no card.
Get started free
