Evaluation criteria and lots
Setting up scoring rules and multi-lot structures for detailed, structured analysis.
Evaluation criteria and lots are optional features that add structure to your analysis. You don't need them for a basic compliance check, but they make a big difference for complex procurements where you're comparing multiple vendors.
Evaluation criteria
What they do
Criteria tell the AI exactly what to extract from each proposal. Instead of a general "check for compliance", you get specific data points that you can compare across vendors.
Two types
Pass/Fail criteria - a binary check.
- "Does the vendor have ISO 9001 certification?" - Pass or Fail
- "Is the proposed timeline within the required 12 months?" - Pass or Fail
- "Did they include three reference projects?" - Pass or Fail
Scored criteria - rated on a scale of 0-100.
- "Relevant project experience" - maybe 85/100 (strong track record)
- "Quality of proposed methodology" - maybe 60/100 (adequate but not detailed)
- "Team qualifications" - maybe 90/100 (exceeds requirements)
Setting them up
Go to your procurement page and look for the criteria section. You have two options:
Auto-extract from RFP - click the button to have the AI read your RFP and suggest criteria. It picks up on evaluation sections, scoring tables, and mandatory requirements. Review what it suggests, edit as needed, and save.
Manual setup - add criteria one by one. For each, specify the name, type (pass/fail or scored), and optionally a description of what to look for.
Set up criteria before running analysis. The AI uses them during analysis to extract specific values. If you add criteria after analysis, you'll need to re-run it (which costs additional credits).
Editing criteria
You can add, edit, or remove criteria at any time. But changes only affect future analyses - existing results stay as they were. To get updated values, re-run the analysis.
Tips for good criteria
- Be specific. "Technical capability" is vague. "Number of certified engineers on proposed team" is measurable.
- Match your RFP. If your evaluation methodology defines specific criteria with weights, replicate those here.
- Don't go overboard. 10-20 well-defined criteria gives you plenty of comparison data without making the analysis unnecessarily slow.
Lots
What they are
Lots split a procurement into separate items. Each lot has its own requirements, and vendors can bid on some or all of them.
Common examples:
- Office supplies: Lot 1 (furniture), Lot 2 (electronics), Lot 3 (stationery)
- IT services: Lot 1 (hardware maintenance), Lot 2 (software development), Lot 3 (helpdesk)
- Construction: Lot 1 (demolition), Lot 2 (structural work), Lot 3 (finishing)
Setting them up
On the procurement page, go to the lots section. For each lot, add:
- Lot number (1, 2, 3...)
- Description (what this lot covers)
- Estimated value (optional)
- Quantity (optional)
How lots affect analysis
When lots are configured:
- The AI identifies which lots each vendor is bidding on
- Findings can be tagged to specific lots
- Criterion values can be per-lot (vendor might score 90/100 on Lot 1 but 60/100 on Lot 3)
- The Lots tab shows a per-lot breakdown
When to use lots
Use lots when your RFP explicitly defines them. Don't try to create artificial lot structures for a procurement that doesn't have them - it won't help and might confuse the analysis.
Criteria + lots together
When you have both criteria and lots, the analysis becomes very structured:
- For each vendor, for each lot they bid on, for each criterion - you get a specific value
- The comparison shows this as a matrix
- You can see exactly where each vendor is strong or weak, per lot
This is the most detailed analysis the platform can produce. It takes more time and credits to run, but for large multi-lot procurements with 5+ vendors, it's worth it.