CONTEXT(CLONE)
Sign in

Check tender bids against every mandatory and scored requirement

For: Sourcing manager or chair of a tender evaluation committee running a formal RFP

Pattern: Specialist panelNeeds live modelsDesigned for 8 to 300 agents
Get early accessSign in to use

The pain today

Each bid runs to hundreds of pages of technical, commercial, legal and sustainability annexes. Evaluators skim and score unevenly, and a missed mandatory requirement surfaces after award, when the decision can be challenged.

The ask

“I attached our RFP with the requirement list and scoring criteria, and the bids we received. For each bid, check every mandatory requirement and tell me where the answer is missing, partial or contradicted elsewhere in the same bid. Show the page you relied on.”

Plain words, as you would say it to a colleague. Edit it to fit your case before you send it.

What you attach or connect

  • RFP with numbered requirements and scoring criteria
  • Bid documents and annexes per bidder
  • Clarification questions and answers
  • Evaluation committee briefs per discipline

The unit of work

One worker task per one requirement in one bid, per reviewer brief.

Why a swarm fits

Each requirement is answered in a small part of one bid, and bidders are independent. Technical, commercial, legal and sustainability reviewers read the same passage against different checklists.

Not for

A quotation decided on price alone, or a few short bids: a comparison spreadsheet is cheaper and fully adequate.

The decision tree

5 typed decisions, each with an action for every answer

At fixed moments in a run, the engine puts one narrow question to a decision model. The decision model never writes text: it answers yes or no with a probability, picks from listed options, or gives a score, about a small slice of the material. The engine then does exactly what this tree says, which is what makes the run auditable. The thresholds are the template's design values, not measured results.

  1. Planner, while planning

    Scope checkYes or no, with a probability

    Before work starts on a unit

    Is this requirement marked mandatory or scored in the RFP, as opposed to background information for bidders?

    Sees only: One numbered RFP requirement with its heading and weighting column

    Why: Avoids assessing every bid against text that carries no evaluation weight.

    • Yes: 0.60 or higherthenAccept
    • Unsure: 0.30 up to 0.60thenAccept
    • No: below 0.30thenSkip this unit

  2. Before workers, before a task runs

    Small worker or strong modelYes or no, with a probability

    Before a task runs

    Does assessing this requirement need more than one reviewer brief, for example a technical answer that depends on a pricing or legal annex?

    Sees only: The requirement text and the bid's table of contents

    Why: Sends a requirement to several specialists only where one brief cannot judge it.

    • Yes: 0.60 or higherthenSplit the work
    • Unsure: 0.35 up to 0.60thenSplit the work
    • No: below 0.35thenAccept

  3. After workers, the judge checks

    Evidence checkYes or no, with a probability

    After a worker answers

    Does the quoted bid passage commit the bidder to meet this requirement, rather than describe a capability, an option or an intention?

    Sees only: The requirement, the quoted bid passage and its page reference

    Why: A capability statement read as a commitment is how a non-compliant bid gets through.

    • Yes: 0.85 or higherthenAccept
    • Unsure: 0.50 up to 0.85thenEscalate to a strong model
    • No: below 0.50thenReject and retry

  4. Reconciler, while merging

    Conflict checkA choice among options

    While reconciling

    Does this other passage of the same bid, such as a pricing assumption or a legal reservation, take back what the answer promises?

    Sees only: The answering passage and one candidate passage from another annex of the same bid

    Why: Bids often promise in the technical annex what they withdraw in the assumptions.

    • Takes back or limits the promisethenMark unresolved
    • Consistent with the promisethenAccept
    • About something elsethenAccept

  5. Accountable person, before anything is settled

    Person decidesYes or no, with a probability

    Before anything is reported as settled

    Would this assessment mark a mandatory requirement as not met or only partly met, which could exclude the bidder?

    Sees only: The assessment, its evidence and the requirement's mandatory flag

    Why: Exclusion is the committee's decision and can be challenged in law.

    Accountable: The evaluation committee owns every score and the award decision; the swarm prepares evidence and never ranks bidders.

    • Yes: 0.40 or higherthenAsk a person
    • Unsure: 0.15 up to 0.40thenAsk a person
    • No: below 0.15thenAccept

The fleet: who does what

Model tiers by role, not brands: you choose the models. Strong reasoning models plan and reconcile, small fast models do the wide work, and the judge is a decision model from a different family, so it does not share the workers' blind spots.

  1. Planner

    A strong reasoning model assigns each requirement to the reviewer briefs that own it and locates the answering sections per bid.

    Decisions here:1. Scope check

  2. Workers

    Small fast workers, each holding one brief (technical, commercial, legal, sustainability), assess one requirement in one bid.

    Designed for 8 to 300 agents, one worker task per one requirement in one bid, per reviewer brief. Each worker receives only its own unit.

    Decisions here:2. Small worker or strong model

  3. Judge, from a different model family

    A decision model from a different family checks that the cited bid passage supports 'met', 'partial' or 'not met'.

    Decisions here:3. Evidence check

  4. Reconciler

    A strong reasoning model builds the compliance matrix per bidder and keeps reviewer disagreements visible side by side.

    Decisions here:4. Conflict check

  5. Accountable person

    The evaluation committee owns every score and the award decision; the swarm prepares evidence and never ranks bidders.

    Decisions here:5. Person decides

Checked before anything is accepted

  • Each assessment quotes the bid passage and page it rests on
  • Answers are checked against pricing and legal annexes for contradictions within the bid
  • The same requirement is read with the same checklist for every bidder
  • Requirements with no located answer are listed, never assumed met

What comes back

  • Compliance matrix per bidder with quoted evidence
  • Mandatory requirements that are missing, partial or contradicted
  • Clarification questions to send to each bidder
  • Points where reviewer briefs disagree about the same answer

What to measure

  • Mandatory gaps found by the committee that the swarm missed
  • Evaluator hours per bid
  • Consistency of assessments for the same requirement across bidders

Names of measures only. No result is claimed for this template.

Templates open in the workspace chat with the ask filled in. Nothing runs until you send it.

Get early accessSign in to use

More in Procurement and supply chain

Reconcile agreed terms across supplier contracts and amendments

For: Procurement lead or category manager preparing renewals or a supplier performance review

The price, delivery date and penalty that apply to each supplier are spread over a master agreement, amendments and emails.

Pattern: Map, verify, reduceRuns today5 decisionsDesigned for 4 to 150 agents