Home/AI Agents/Evaluation Designer

Pre-award

Evaluation Designer

Drafts criteria you can actually evidence.

Draft — agency owns the rubricCited to your documents
Criteria against the submissionExample
Technical approach30% \x{00B7} evidencedRelevant experience25% \x{00B7} evidencedKey personnel20% \x{00B7} evidencedInnovation10% \x{00B7} not evidenced
By handCriteria you cannot evidence

A rubric asks evaluators to score "experience" with nothing in the submission that establishes it, and the award is argued about later.

With the agentCriteria you can defend

Every criterion checked against what bidders are actually required to submit, with the unevidenced ones flagged before publication.

How it works

Drafts evaluation criteria, weights and a scoring rubric that map to the scope, and checks each criterion is measurable from what bidders submit.

Criteria that cannot be evidenced from the submission are the most common protest opening.

01

It reads the scope first

Criteria that do not map to something the scope asks for are criteria nobody can defend at award.

02

It drafts criteria and weights

A rubric with weights that add up, and scoring levels that describe what each score means rather than leaving it to instinct.

03

It checks each one is evidenced

For every criterion it asks the same question: what in the required submittals would an evaluator read to score this?

04

It flags what cannot be scored

Where nothing in the submission evidences a criterion, it says so, so you can add a submittal or drop the criterion.

What it checksStep 01

Mapped to the scope

Each criterion traced to a scope sectionUnmapped criteria flaggedScope gaps surfaced

Weights that total

0

percent allocated

Across five criteria, with levels described.

Criterion against submittal

✓✓✓✓✓✓—✓✓✓✓✓—✓✓✓

Cannot be scored as written

InnovationLocal presenceEverything else is evidenced

Ask it anything

Every answer quotes its page.

Answers come only from the documents on the run, and each one carries the citation you would have gone looking for.

Evaluation Designer

Not as written. Nothing in the required submittals asks a bidder to describe innovation, so evaluators would be scoring an impression. Either add a submittal that calls for it, or drop the criterion.

What it will not do

It drafts and tests the rubric against what bidders must submit. The weights, the criteria and the final rubric are the agency\x{2019}s, and so is every score given later.

Set the weightsScore a proposalApprove the rubricDecide the method

Pre-award

The rest of this stage

Each agent is scoped to one task, and they work on the same record.

Solicitation ReviewerChecks the solicitation before it goes out. Scope WriterDrafts the scope of work. Requirements ArchitectProposes the requirements that fit the buy. Cost & Price AnalystLays out what the record says about price. Market Research AnalystKnows your buying history. Justification WriterDrafts the written determination. Cooperative FinderChecks if you need to bid at all.

See all 23 agents →

Run it on a document you already have.

Tell us how your office works today and we will show you this agent on your own file.

Book a discovery call →