Complex sourcing events rarely produce standardised supplier responses. Suppliers submit proposals, pricing schedules, and supporting documents in different formats, structures, and terminology — making manual evaluation slow, inconsistent, and easy to get wrong. Comparing bids across dozens of attachments and question responses is one of the most time-consuming parts of running a sourcing event.
AI Bid Comparison brings an AI assistant directly into your comparison workspace to solve this. It analyses selected supplier submissions and supporting documents, then normalises the content into a structured, apples-to-apples comparison — giving you a conversational way to explore bid data, identify gaps and inconsistencies, and stay in full control of the evaluation.
What it is
AI Bid Comparison is an AI-powered assistant that appears on the Comparison tab of a request. When the feature is enabled for your company, a floating AI button appears on the page. Clicking it opens a full-screen panel — an embedded AI assistant — alongside your existing comparison view, and summarises and analyses the responses your suppliers have submitted, so you spend less time digging through documents and more time making decisions.
How it works
Log in to DeepStream and navigate to the Request module.
Switch to the Sent folder and open the relevant request.
Navigate to the Comparison tab.
Click the 'Bid comparison' button (mid-right side of the page). It opens in a full-screen view without leaving your comparison page.
Optionally, use the Content to include dropdown to select which supplier responses and supporting documents you want to compare.
Click Compare & Summarise. The assistant reviews the selected documents and question responses and prepares a structured summary as a starting point for your analysis.
Ask follow-up questions. Use the messaging panel to type focused, supplier-specific questions about the report - for example, asking it to clarify a gap or dig deeper into a particular supplier's pricing.
Once you've added follow-up questions, click 'Update summary' to regenerate the comparison with that additional context folded in, rather than starting a new comparison from scratch.
Exit the assistant at any time and return to your comparison as normal.
Core capabilities
AI Bid Comparison analyses supplier proposals, questionnaires, pricing files, and supporting documents to generate an evidence-based comparison. It can:
Compare content across documents and across suppliers
Normalise terminology, formats, and commercial structures
Detect missing, boilerplate, or non-responsive answers
Identify contradictions, exclusions, assumptions, and unsupported claims
Separate supplier-stated facts from missing information and AI interpretation
Flag non-comparable pricing and the buyer inputs needed to resolve it
Generate supplier-specific clarification questions
Answer natural-language questions about the completed comparison, and refresh the report as new questions are answered
What data the assistant can see
We built this feature with data protection front of mind. When the panel opens, only the following is sent from the request to the AI:
Supplier document attachments — files uploaded by suppliers as part of their bid response (for example, certificates or proposals).
Supplier question answers — text, multiple-choice, and other question responses suppliers submitted.
Only data you're already permitted to see is forwarded, and only for the request you're currently viewing — the assistant has no ability to reach outside that data. The following are always excluded:
Locked exchanges — responses hidden behind stage gates, sealed-bid rules, or auction locks (for example, sealed bids or auction bids that haven't yet been opened).
Obsolete exchanges — any exchange that has been removed or superseded.
Disclaimers
AI Bid Comparison is a decision-support tool, not a substitute for buyer judgement:
It is not a simple side-by-side document summary.
It does not prove that a supplier claim is true.
It does not treat a populated response field as a complete or valid answer automatically.
It does not make non-comparable pricing comparable by inventing assumptions.
It does not remove buyer or supplier accountability.
Best practices
Compare only the documents and responses relevant to the evaluation.
Check that the comparison clearly separates facts, missing evidence, and AI interpretation.
Don't draw commercial conclusions where pricing models aren't directly comparable.
Use the generated clarification questions to resolve material gaps before making evaluation decisions, and use Update Summary to keep the report current as you do.
The value
Save time — skip the manual read-through of every attachment and answer; get a head start with an AI-generated summary.
Consistency — reduce the risk of missing details buried in long documents or dense question sets.
Confidence — sensitive, sealed, obsolete, or stage-gated bid content stays protected and is never exposed to the assistant