AI compares vendor options against your criteria and produces a scored shortlist for your decision. Includes an evaluation framework, AI scoring matrix, and 90-day performance review loop.
A complete operational system. Built to run immediately.
Define 4-6 weighted criteria, compile a longlist of 5-10 vendors, run the AI scoring matrix, complete due diligence, and decide.
Why vendor decisions without a criteria framework produce inconsistent results and how an AI scoring matrix removes subjectivity.
Five roles assigned before the first evaluation, five required inputs, and clear handoff points at each phase.
Four phases: define criteria and compile the longlist before viewing any options, score every vendor, complete due diligence, decide.
Three primary metrics: Vendor Selection Cycle Time (under 14 days), Criteria Coverage Rate (100%), and Satisfaction Score (4.0+).
Two reusable prompt formats, a standard criteria category list, and a reference anchor set for scoring calibration.
Six common failure points with specific fixes: AI top scorer diverging from instinct, similar scores, pricing gaps, timeline pressure.
Version history, a criteria weight change log updated after each 90-day review, and a new criteria type addition record.
Takes the evaluation need statement and returns a criteria framework with 4-6 weighted criteria, 1-3 scoring anchors per criterion, and a rationale.
Takes a vendor name and available documentation and returns a standardized data profile using the criteria framework as the extraction structure.
Takes the criteria framework, weights, anchors, and all vendor profiles and returns a scored comparison table sorted by total weighted score.
Takes the shortlisted finalists and the criteria where data was thinnest, then returns 6-10 targeted questions for final conversations.
Generates a 5-step workflow for compiling the criteria framework and sending vendor questionnaires when a new evaluation is created.
Generates 8-12 reference check questions for a shortlisted finalist covering criteria verification and relationship quality.
A defined set of 4-6 evaluation criteria with weights (Must-Have 3x, Important 2x, Nice-to-Have 1x) and 1/3/5 anchors for each criterion.
A full comparison table with every longlist vendor scored on every criterion, weighted totals calculated, must-have gates applied, and the.
The complete evaluation file: criteria framework, final scored matrix with any override notes, decision summary with selected vendor and.
Six copy-paste-ready prompts covering criteria framework generation, vendor scoring, comparison matrix, decision record, declined vendor.
A configured workflow that generates the criteria framework and scoring matrix when a new evaluation is initiated.
The complete evaluation file for every vendor selection: criteria, scores, final decision, and 90-day review findings.
Most vendor selections happen on incomplete data and inconsistent criteria. Without a scoring matrix, the choice cannot be defended when the vendor underdelivers.
Without a criteria framework, each vendor gets judged on whatever stood out in the demo. The best presenter wins.
When criteria are invented after viewing options, the evaluation confirms a preference rather than tests it.
Vendors quote differently. Without a normalized format, the evaluation compares incompatible data points.
A vendor who invested time in a proposal with no outcome communication damages the relationship permanently.
Expected Outcomes
These targets come directly from the SOP. Realistic benchmarks, not aspirational claims.
<14 Days
Measures whether the evaluation process is efficient enough to produce a decision without creating a procurement bottleneck. Evaluations consistently above 14 days indicate either an oversized longlist, undefined decision authority, or a criteria framework that is being revised mid-evaluation. All three are addressable before the next cycle.
100%
Measures whether the finalist verification step is functioning as a real gate, not a formality. A must-have criterion that was never actually verified is a risk that transfers to the 90-day review. Any selection where must-have coverage was below 100% means the scoring matrix was used as the final word rather than a starting point.
100%
A vendor selection with no documentation is institutional knowledge that exists only in someone's memory. Documentation rate is not about compliance, it is about having the data to improve the criteria framework after the 90-day review, and to make the next evaluation in the same category faster and more reliable.
Evaluation habit established. More time accepted in criteria definition. Focus on completing the 90-day reviews so the framework has real performance data to improve from.
Framework improving with each cycle. Score reliability rising as criteria weights are refined from 90-day findings. Cycle time dropping as the process becomes familiar.
Intake automation active. Framework validated against multiple 90-day reviews. Decision quality consistently higher than pre-framework evaluations, with a documented record showing why.
Common Questions
QuickStart gets your first evaluation framework built in about 2 hours. That includes defining 4-6 weighted criteria, building anchor descriptions for each rating level, running AI ASSIST 1, and scoring your first vendor profile.
Define 4-6 criteria per evaluation, maximum. More than 6 reduces differentiation between vendors without improving decision quality. If the evaluation seems to require more, the need statement is not specific enough, narrow the use case first.
Include pricing as a criterion but treat the initial scores as provisional. Send a pricing request to all longlist vendors at the same time the evaluation opens, before running the scoring matrix. Mark pricing scores as data-thin in the matrix.
Not directly. A renewal evaluated in isolation is not a competitive evaluation, it is a rubber stamp. For renewals, pull the original evaluation record, update the scores based on actual performance since selection, and include at least one.
The 6 prompts cover criteria framework generation, vendor scoring from data inputs, comparison matrix creation, decision record documentation, declined vendor notification drafting, and a criteria calibration review.
