Copilot · Bids, strategy and analysts
Reads long documents and gives you a summary you can check line by line.
The research assistant reads tender documents, reports and internal files, and writes summaries with a reference to the page each point came from. Your team checks the summary and decides what to act on.
Research assistant
Summarise the municipal smart meter tender and both corrigenda. Do we meet eligibility?
Bid due 6 Nov 2026, 3 pm (moved from 28 Oct by corrigendum 2). EMD INR 12 lakh. Scope: supply, install and maintain 40,000 meters over 5 years.
Eligibility: average turnover of INR 25 crore over FY 2023 to FY 2026 (corrigendum 1 lowered it from INR 40 crore, p. 3). Two similar projects of 10,000 meters each (p. 47).
I cannot confirm whether you meet these; I do not have your audited financials or project list. Please check with the bids team.
Summary saved with 18 page references for review.
- Type
- Copilot
- Team
- Bids, strategy and analysts
- Connects to
- 6 systems, including Google Drive
- Limits
- 5 set, plus approvals
How it is done today
The work before the agent
A government tender arrives as four PDFs and two corrigenda, three hundred pages in all. A bid manager spends two days pulling out eligibility rules, deadlines, the EMD amount and the scope, then checks whether the corrigenda changed any of it. Small details, like a revised turnover threshold, are easy to miss.
What the research assistant takes on
Structured summaries
It pulls out the points your team always needs, such as eligibility, key dates, EMD, scope and payment terms, into a fixed format with page references.
Change tracking
When a corrigendum or new version arrives, it lists what changed against the earlier document so nothing is read twice or missed.
Questions across files
Analysts can ask questions across a folder of reports or tenders, and each answer cites the file and page it relied on.
Checklists and notes
It drafts a compliance checklist or a briefing note from the summary, using the report summary generator template, for the team to edit.
How the work moves
A typical piece of work, stage by stage
- Step 1Question setBy an analyst
- Step 2Sources gatheredYour files and approved sites
- Step 3Read and comparedClaims linked to sources
- Step 4Draft briefWith open questions
- Step 5Analyst reviewsBefore it is shared
Where a person decides
The bid manager or analyst checks the summary against the cited pages and decides whether to bid, what to recommend or what to quote. The summary is a starting point for that review.
Limits it works within
- Every point links to the page or paragraph it came from
- Works only from documents you give it, unless you allow web sources
- Says plainly when it cannot find something in the documents
- Folder access follows your existing sharing permissions
- Confidential bid files can stay in a private cloud or on premise setup
Connects to
Works with the systems you already use
Through APIs, exports or a custom connector our team builds. The exact list depends on your versions and setup.
- Google Drive
- SharePoint
- OneDrive
- Gmail
- Notion
- Microsoft Teams
Questions
About the research assistant
Can it read scanned PDFs?
Yes, through text extraction, though poor scans and complex tables can be misread. That is why every point carries a page reference you can check.
Does it search the internet?
Only if you switch that on. By default it works from the documents you upload or connect, which keeps answers traceable.
Which AI model does it use?
We choose per project from leading providers, such as OpenAI, Anthropic Claude, Google Gemini or open models, based on your data rules and the documents involved.
Keep going
Other agents and reading
Copilot
Sales copilot
Gets a rep ready for the next meeting from the records you already keep.
Learn moreCopilot
Legal copilot
Reads a contract against your own playbook and points out what needs a lawyer's eye.
Learn moreGuide
Retrieval augmented generation (RAG) for company documents, in plain English
RAG lets an AI model answer from your own documents. The system finds the most relevant passages first, gives only those to the model, and asks for an answer that cites them, so people can check the source.
Learn moreSee it on your work
Bring one slow, repetitive process. We will show you an agent doing it.
A working session with someone who builds these every week.
Or get a ballpark proposal in two minutesWhat happens next
- Step 1A short call about one process and the systems involved
- Step 2A demo of an agent working on a version of your task
- Step 3A clear view of fit, approvals and what it would take