Overview
Core concepts and terminology used throughout the platform.
Before you start using the platform, it helps to understand a few terms. None of this is complicated - just want to make sure we're speaking the same language.
Procurement
A procurement is your project workspace. Think of it as a folder that holds everything related to one purchasing process: the requirements, the vendor proposals, the analysis results, the comparison data.
You give it a name (like "Office Supplies 2026" or "IT Infrastructure Upgrade"), set the language, and optionally add a deadline for your own reference.
RFP documents
RFP stands for Request for Proposal. These are your requirements - the documents that describe what you're looking for. Technical specifications, contract terms, evaluation rules, appendices - upload them all.
The system parses them (converts them from PDF/Word into text the AI can read) and uses them as the baseline for every analysis.
Bids and bid documents
A bid represents one vendor's submission. You create a bid entry ("Company X"), then upload their proposal documents - the main response, supporting materials, CVs, financial statements, whatever they sent.
You can have as many bids per procurement as you need.
Analysis
When you click "Analyze", the system reads the RFP and the bid documents, compares them, and produces a report. The analysis runs in the background - you can watch it progress in real time or come back to it later.
Each analysis goes through four phases:
- Preparation - reading the RFP, understanding what's required
- Specialist review - pulling out specific data points from the proposal
- Main analysis - comparing requirements against what the vendor offered, flagging gaps
- Verification - a separate AI double-checking the most serious findings
Findings
Findings are the individual issues the AI flags. Each one has:
- A severity level: Critical, Major, Minor, or Strength
- A description of what's wrong (or right)
- Evidence - exact quotes from the documents
- A provenance badge showing which agent found it (Lead Analyst, a specific criterion specialist, or a lot specialist)
Here's what the severity levels mean in practice:
| Severity | What it means | What to do |
|---|---|---|
| Critical | The proposal fails a mandatory requirement. Likely grounds for rejection. | Address immediately or reject the bid. |
| Major | A significant gap or unclear response. Not a deal-breaker on its own, but risky. | Request clarification from the vendor. |
| Minor | A small omission or formatting issue. Unlikely to affect the outcome. | Note it for the vendor to fix. |
| Strength | Something the vendor did well. | Factor into your evaluation positively. |
Strategic briefing
The strategic briefing is the executive summary at the top of each analysis. It gives you a recommendation and a plain-language overview:
- Strong Submit - the proposal looks solid, few or no issues
- Submit with Improvements - decent but has gaps that should be fixed
- Significant Revision Needed - serious problems, risky to submit as-is
- Do Not Submit - critical compliance failures
Evaluation criteria
These are optional scoring rules you can set up before running analysis. Two types:
- Pass/Fail - does the vendor meet a specific requirement? (e.g., "Has ISO 9001 certification")
- Scored - how well does the vendor perform on a scale of 0-100? (e.g., "Relevant project experience")
When you set up criteria, the AI extracts values from each proposal. This is what makes bid comparison possible - you get apples-to-apples scores across all vendors.
You can have the AI suggest criteria by reading your RFP, then edit the list yourself. This saves time on complex procurements.
Lots
Some procurements are split into lots - separate items that vendors can bid on independently. For example, an office supplies procurement might have Lot 1 (furniture), Lot 2 (electronics), and Lot 3 (stationery). Vendor A might bid on Lots 1 and 3, Vendor B on all three.
If your procurement has lots, you can define them in the system. The analysis will then track which vendor participates in which lot.
Credits
Credits are how you pay for analysis. You buy them upfront (minimum 5 at a time), and each analysis uses a certain number depending on the size and complexity of the documents.
Your credit balance is always visible in the top bar. Click Buy to purchase more. More details on the pricing page.
Composing RFPs and bids
The platform doesn't just analyze documents — it can also draft them. There are two composition flows:
- Composing an RFP — for buyers. The agent drafts the RFP document set from your description and any reference materials.
- Composing a bid — for vendors. After a short setup interview where you approve a plan, the agent drafts your bid sections and runs them through a validation pass that checks for filler text, unsupported claims, and missing coverage.
Both flows let you choose between Autopilot (runs straight through), Guided (pauses after each section so you can review), and Interview (asks more questions during composition).
RFP quality review
Before publishing your RFP, you can run an AI quality review on it. The system reads your requirement documents and flags issues across eight categories: clarity, completeness, consistency, fairness, criteria quality, timeline, legal, and budget.
Each issue gets a severity level (Critical, Major, Minor, or Strength) and an overall recommendation: Ready to Publish, Minor Improvements Needed, Significant Revision Needed, or Not Ready for Publication.
Think of it as a pre-flight check for your procurement documents. It costs credits, but catching a vague requirement before vendors start bidding is much cheaper than dealing with the fallout later.
Preparation checklist
After uploading your RFP, you can generate a preparation checklist — an AI-generated list of everything you need to prepare before submitting a bid. Items are organized by category and tagged with priority levels: required, recommended, or optional.
You can mark items as done and track your progress. Generating the checklist is free.
Market research
Before writing an RFP, you may want to understand what's available on the market. The Market Research feature lets you describe what you need, optionally upload reference documents (existing specs, standards), and an AI agent searches the web to compile structured findings.
Each finding is categorized (product, supplier, regulation, risk, recommendation) and backed by clickable source URLs with extracted evidence. Confidence scores (HIGH, MEDIUM, LOW) tell you how reliable each claim is. The agent also produces an editable research report and, when appropriate, a draft technical specification.
Market research runs independently of procurements — use it to inform your RFP writing or to validate specifications before publishing.
Captured entities
During analysis, the AI picks up names of people and companies mentioned in the documents. For Latvian procurements, these are automatically matched against the official business register - so you can see who the beneficial owners are, check for conflicts of interest, and do basic due diligence without leaving the platform.