AI Relevance Scoring
Every tender scored 1-10 against your specific business profile. Stop reading notices that do not fit — focus only on contracts you can win.
Keyword search finds tenders containing your words. AI scoring finds tenders that match your actual capabilities — even when buyers use completely different terminology.
What scored results look like
Example for a cybersecurity consultancy — same search, scored by AI
Cybersecurity consultancy — NCSC CAF assessment, Cabinet Office
Direct match: NCSC CAF assessment is your stated core service. Cabinet Office buyer aligns with your central government focus.
Security operations centre — NHS South London ICB
Strong match: SOC delivery is in scope. NHS buyer fits your public sector experience. Value within your typical range.
IT managed services — London Borough of Southwark
Partial match: IT managed services overlaps your capability but does not specifically require security expertise. Worth reading.
Telecoms infrastructure — Network Rail
Weak match: Telecoms infrastructure is outside your described capability. Security element is incidental, not the core requirement.
Print and document management — HMRC
No match: Document management has no overlap with your cybersecurity and advisory profile.
How AI scoring works
AI scoring vs keyword search
Available on Growth and above
AI relevance scoring is included on Growth (£599/mo), Professional (£1,199/mo), Bid Agency, and Enterprise plans.
View all plans →How to get the most accurate scores
Be specific in your profile
Write 3-4 sentences describing exactly what you do, for whom, and what types of contracts you target. Vague profiles produce generic scores.
Include your sector focus
Specify which public sector segments you target — NHS, local government, central government, defence, education. AI uses this to weight sector fit.
Add contract type keywords
Include the specific service lines you win — not just broad categories. "Penetration testing" scores better than "security".
Update when you change
If your capabilities or target sectors change, update your profile. AI scores are only as current as your profile description.
Frequently asked questions
Which AI model does PSIP use for scoring?
PSIP uses Anthropic's Claude — one of the most capable large language models available. Claude reads each tender notice and compares it against your business profile to produce a 1-10 relevance score with a reasoning statement.
How accurate is AI tender scoring?
Accuracy improves significantly with a detailed, specific business profile. A vague profile produces generic scores. A profile that describes exactly what you do, which sectors you target, and what types of contracts you typically win produces highly accurate scores. Most users find scores of 7+ reliably indicate contracts worth reading.
Does AI scoring replace reading tender notices?
No — it replaces reading irrelevant tender notices. Tenders scoring 8-10 should be read in full. Tenders scoring 1-3 can typically be dismissed. Tenders scoring 4-7 warrant a quick scan. The goal is to focus your reading time on what matters.
Which plans include AI scoring?
AI relevance scoring is available on Growth (£599/month), Professional (£1,199/month), Bid Agency, and Enterprise plans. It is not included on the Starter plan.
Can bid agencies score tenders for multiple clients?
Yes. The Bid Agency plan includes client workspaces. When you switch to a client workspace, AI scoring automatically uses that client's business profile — so every search is scored against the right context for each client.