AI Agents · Procurement Workflow
Procurement agents that produce real deliverables.
Various agents built around our tender and contract workflow. They take the inputs you provide, produce drafts in the format your team can issue, and stop where judgement starts.
Request For Quotation
Transforms project briefs into draft tender packages.
The Agent takes a project brief and a package scope, then drafts the full RFQ pack: instructions to bidders, scope of works structure, evaluation criteria, conditions of contract, and the standard return schedules. Output is editable, ready for review and issue.
Inputs
- Project brief and programme schedule
- Package scope notes
- Contract framework (FIDIC, bespoke, NEC)
- Bidder list
Outputs
- RFQ cover instructions
- Scope of works template aligned to the package
- Evaluation criteria and scoring framework
- Particular conditions of contract (where framework allows)
- Bidder return schedules (BOQ template, technical schedule, programme schedule)
Where it fits
Approach Step 02: Structure
The agent gets the RFQ to 80% in an afternoon. The remaining 20%, the package-specific judgement calls, stays with us.
Drafted RFQ pack.
Bid Evaluation
Evaluates bidder submissions against your scoring framework.
The Bid Evaluation Agent reads bidder submissions, scores each criterion against the rubric, writes the rationale per score, and produces the same evaluation report we would issue manually, only faster and consistent across packages.
Inputs
- Tender requirements pack
- Bidder submissions (PDF / DOCX / XLSX)
- Scoring rubric and criteria weights
Outputs
- Filled evaluation matrix per bidder
- Per-criterion rationale and evidence reference
- Award recommendation memo (draft)
- Clarification log: items the agent flagged for human follow-up
Where it fits
Approach Step 04: Evaluate
The agent does not award the contract. It builds the documented case so the team makes the call faster, with the rationale already written down.
Paired with the contractor pre-qualification checks we run before evaluation begins.
Structured evaluation output.
Contract Review
Marks up contracts in the position we hold.
The Contract Review Agent reads a contract draft, identifies clauses that work against the position we want to protect, and returns the document with inline comments written in the same voice we would use ourselves: direct, commercial, defensible.
Inputs
- Contract draft (DOCX / PDF)
- Position to protect (e.g., uncapped liability for fraud only, schedule protection, payment terms)
- Project context (lump sum vs. GMP, FIDIC variant, etc.)
Outputs
- Marked-up contract with inline comments
- Comment summary table: clause reference, issue, proposed wording
- Negotiation note: which positions are firm, which are subject to trade
Where it fits
Approach Step 02: Structure (pre-issue review) and post-bid contract negotiation
The agent does not sign anything. It produces the redline that the team would have produced: same voice, same priorities, in a fraction of the time.
Applied across the capital programmes we run, under FIDIC or bespoke frameworks.
Marked-up contract redline.
Workflow
How the agents fit into the workflow.
- 01
Structure
RFQ Agent drafts the tender pack.
- 02
Tender
Run by your team. Market interaction stays human.
- 03
Evaluate
Bid Evaluation Agent scores and reports.
- 04
Govern
Contract Review Agent marks up the resulting contract.
Prove it on yours
See an agent run against your tender.
Share a brief and a representative bidder submission. We’ll return the worked output in your format. It’s the same deliverable our team would issue on a live engagement, sent under NDA. The judgement stays with you.