---
title: "Work with the Tender Assistant"
description: "Chat with the Tender Specialist about a tender, start it from the New Tender wizard, answer its questions, and understand what it can read and change on your behalf and who can see your sessions."
url: https://helpcenter.complytrain.net/en/articles/241-work-with-the-tender-assistant
collection: "Submission and learning"
updated: 2026-09-22
language: en
---

# Work with the Tender Assistant

Chat with the Tender Specialist about a tender, start it from the New Tender wizard, answer its questions, and understand what it can read and change on your behalf and who can see your sessions.

Chat with the Tender Specialist about a tender, start it from the New Tender wizard, answer its questions, and understand what it can read and change on your behalf and who can see your sessions.

The Tender Assistant is an AI assistant, the Tender Specialist, that works inside one tender. It can read the tender and your organisation's material, answer questions about them, and do work for you: record requirements and deadlines from the RFP, outline and draft response sections, run assessments and prepare the submission package. Everything it adds appears on the tender's tabs, where you can check, edit or delete it.

## Open the assistant

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

1.  Open the tender.
    
2.  Choose **Chat with AI** in the header, or the **Chat** tab.
    

The page **AI Assistant** opens, with the tender's name in the path above it. Your most recent session for this tender opens; if you have none, a new one starts.

The assistant in the document editor is a different assistant, focused on the document you are editing; see [write, review and approve sections](https://help.complytrain.net/en/tender/write-review-sections/).

## Who can use it

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

Anyone who can open the tender can use the assistant: you need **View Tenders** (`tender:view`) and a place on the tender's team, or the Admin role.

The assistant acts as you. It can read what you can read, and it can change only what your permissions allow. With **View Tenders** alone you can ask questions, but the assistant cannot change anything. When you ask for a change you are not allowed to make, the assistant tells you which permission is missing.

## Keep your sessions private

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

A session is one conversation. Each session belongs to the person who started it: other team members do not see your sessions in their list and cannot open them. Anything the assistant changes on the tender, though, is visible to everyone who can open the tender.

The **Conversations** list on the left shows your sessions for this tender, newest first, with the date and number of messages.

-   Choose **New** to start a fresh session, for example for a new topic. The assistant keeps each session's history separate.
    
-   Use **Search conversations...** to find an earlier session by title.
    
-   Choose a session to reopen it and continue where you left off.
    
-   Choose **Delete conversation** on a session to delete it. Changes the assistant already made to the tender stay.
    

## Start from the New Tender wizard

Product location: `/tender/workspaces/new`. Open Tenders and choose New Tender.

When you start a tender with **Describe to AI** or **Upload a tender document**, the wizard opens the assistant in a new session and sends an opening message for you.

<table style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Wizard choice</p></th><th colspan="1" rowspan="1"><p>What the assistant does</p></th></tr><tr><td colspan="1" rowspan="1"><p><strong>Describe to AI</strong></p></td><td colspan="1" rowspan="1"><p>Asks you about the opportunity: the issuing organisation, key dates, the estimated value and any formal requirements. It then sets up the tender's details, draft requirements, deadlines and formalities from your answers.</p></td></tr><tr><td colspan="1" rowspan="1"><p><strong>Upload a tender document</strong></p></td><td colspan="1" rowspan="1"><p>Reads the uploaded documents, records what it can find (the issuing organisation, reference, deadlines, requirements, formalities and expected output documents), sets a sensible tender name, and asks you about anything the documents do not answer.</p></td></tr></tbody></table>

Review what it recorded on **Overview**, **Requirements**, **Deadlines**, **Formality** and **Output documents**, and correct anything it got wrong. See [create a tender](https://help.complytrain.net/en/tender/create-tender-workspace/).

## Send messages and follow the work

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

Type in the message box at the bottom and press Enter to send. Press Shift+Enter for a new line.

-   The reply appears as the assistant writes it. Each step it takes, such as reading the tender or adding a requirement, appears as a card in the conversation. Use the details button in the header to show or hide the technical output of those steps.
    
-   **Stop generating** stops the current reply.
    
-   **Copy message** on a message copies its text, for example to paste into an email.
    
-   **Regenerate response** on one of the assistant's replies sends your message again for a fresh answer.
    
-   You can send a message while the assistant is still working. **AI is currently working** asks how to deliver it: **Send Now** lets the assistant see it immediately, for example to steer it, and **Wait & Send** sends it when the current work is finished. **Cancel** keeps the message unsent.
    
-   For longer jobs, the assistant continues on its own for several rounds, and a status line shows which round it is on.
    
-   If it pauses at its limit for one turn, the panel **Analysis paused** explains why. Choose **Continue analysis** to let it carry on where it left off.
    

The assistant names its sources, such as the RFP section or document it used. When the material does not cover something, it says so rather than inventing an answer. Check facts, figures and commitments before they go into a bid.

## Answer the assistant's questions

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

When the assistant needs a decision from you, a **Decision required** card appears, marked **Waiting for your response**.

1.  Choose one or more of the options. An option marked **Recommended** is the assistant's suggestion.
    
2.  Where the card offers **or write your own answer**, you can type your own answer instead of, or as well as, choosing an option.
    
3.  Choose **Send answer**. The card shows **Answer sent**, and the assistant continues.
    

Choose **Skip** or **Dismiss** to let the assistant continue without your answer. Unanswered questions are still there when you reopen the session.

## What it can read

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

<table style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Source</p></th><th colspan="1" rowspan="1"><p>Examples</p></th></tr><tr><td colspan="1" rowspan="1"><p>The tender</p></td><td colspan="1" rowspan="1"><p>Details, status, requirements with dispositions, deadlines, formalities, output documents and their guidance, section progress, requirement coverage, the compliance matrix, clarifications and submission readiness.</p></td></tr><tr><td colspan="1" rowspan="1"><p>The tender's documents</p></td><td colspan="1" rowspan="1"><p>The RFP and other documents in the tender's Document Vault, once they are indexed.</p></td></tr><tr><td colspan="1" rowspan="1"><p>Analysis results</p></td><td colspan="1" rowspan="1"><p>RFP analysis and AI assessments started on the tender.</p></td></tr><tr><td colspan="1" rowspan="1"><p>Your organisation's material</p></td><td colspan="1" rowspan="1"><p>Organisation facts, documents in vaults you can read, and reference knowledge bases.</p></td></tr><tr><td colspan="1" rowspan="1"><p>Reusable content</p></td><td colspan="1" rowspan="1"><p>Approved answers in the <a target="_blank" rel="noopener noreferrer nofollow" class="linktype-link link" data-linktype="link" href="https://help.complytrain.net/en/tender/reuse-answers-and-history/">Answer Library</a>.</p></td></tr><tr><td colspan="1" rowspan="1"><p>Experience</p></td><td colspan="1" rowspan="1"><p>Win/loss figures, tenders you have won or lost, and your <a target="_blank" rel="noopener noreferrer nofollow" class="linktype-link link" data-linktype="link" href="https://help.complytrain.net/en/tender/submission-history/">Submission History</a>.</p></td></tr></tbody></table>

The assistant reads only what you are allowed to see. It does not read documents in vaults you cannot open, and it does not read other people's sessions.

## What it can change

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

<table style="min-width: 75px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Area</p></th><th colspan="1" rowspan="1"><p>What you can ask for</p></th><th colspan="1" rowspan="1"><p>Permission needed</p></th></tr><tr><td colspan="1" rowspan="1"><p>Tender details</p></td><td colspan="1" rowspan="1"><p>Set the name, description, issuing organisation, reference, tender type, estimated value and submission deadline; set up a tender from a description.</p></td><td colspan="1" rowspan="1"><p><strong>Create &amp; Work on Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>RFP</p></td><td colspan="1" rowspan="1"><p>Analyse an uploaded RFP and add the requirements, deadlines, formalities and expected output documents it finds.</p></td><td colspan="1" rowspan="1"><p><strong>Create &amp; Work on Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Tender items</p></td><td colspan="1" rowspan="1"><p>Add requirements, deadlines, formalities and output documents; write an output document's purpose and guidance; record whether you can meet a requirement, with reasoning; add and resolve clarification questions.</p></td><td colspan="1" rowspan="1"><p><strong>Create &amp; Work on Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Analysis</p></td><td colspan="1" rowspan="1"><p>Start readiness, improvement and gap assessments; apply the suggestions you accepted; find existing evidence for a requirement; cross-reference everything you know on a topic.</p></td><td colspan="1" rowspan="1"><p><strong>Create &amp; Work on Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Response writing</p></td><td colspan="1" rowspan="1"><p>Build a document's section outline; draft a section; insert an approved library answer; move a section through review; link requirements to sections; generate the compliance matrix document.</p></td><td colspan="1" rowspan="1"><p><strong>Create &amp; Work on Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Outcome</p></td><td colspan="1" rowspan="1"><p>Record the win/loss debrief.</p></td><td colspan="1" rowspan="1"><p><strong>Create &amp; Work on Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Commitments</p></td><td colspan="1" rowspan="1"><p>Record the bid decision, only once you have agreed it; generate the submission package.</p></td><td colspan="1" rowspan="1"><p><strong>Submit Tenders</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Library</p></td><td colspan="1" rowspan="1"><p>Promote a response document's approved sections into the Answer Library.</p></td><td colspan="1" rowspan="1"><p><strong>Manage Answer Library &amp; Past Submissions</strong></p></td></tr><tr><td colspan="1" rowspan="1"><p>Team</p></td><td colspan="1" rowspan="1"><p>Assign a section to a team member.</p></td><td colspan="1" rowspan="1"><p><strong>Manage Tender Team</strong></p></td></tr></tbody></table>

The assistant never submits a tender or records its outcome. A person with **Submit Tenders** does that; see [validate, package and submit](https://help.complytrain.net/en/tender/prepare-submission/). Once a tender is submitted, the assistant can change only the debrief and clarifications, and on an archived tender it can only read.

## Messages you may see

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

<table style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Message</p></th><th colspan="1" rowspan="1"><p>What to do</p></th></tr><tr><td colspan="1" rowspan="1"><p><strong>Could not start a chat. Please try again.</strong></p></td><td colspan="1" rowspan="1"><p>Try again. If it keeps happening, check that you can still open the tender.</p></td></tr><tr><td colspan="1" rowspan="1"><p><strong>Could not load messages.</strong></p></td><td colspan="1" rowspan="1"><p>Reload the page.</p></td></tr><tr><td colspan="1" rowspan="1"><p><strong>Failed to send message. Please try again.</strong></p></td><td colspan="1" rowspan="1"><p>Send the message again.</p></td></tr><tr><td colspan="1" rowspan="1"><p><strong>The response was cut off because it exceeded the size limit. Try a more specific question.</strong></p></td><td colspan="1" rowspan="1"><p>Ask for less at once, for example one output document instead of all of them.</p></td></tr><tr><td colspan="1" rowspan="1"><p><strong>This chat session is no longer active. Start a new chat to continue.</strong></p></td><td colspan="1" rowspan="1"><p>Choose <strong>New</strong>.</p></td></tr><tr><td colspan="1" rowspan="1"><p><strong>You are not signed in. Please sign in again.</strong></p></td><td colspan="1" rowspan="1"><p>Sign in again and reopen the tender.</p></td></tr></tbody></table>

If the assistant is not available for your organisation at all, ask your administrator: it is set up together with the Tenders module.

## Example: start from a description

Product location: `/tender/workspaces/{workspaceId}/chat` (fallback: `/tender/workspaces`). Open the tender and choose Chat with AI or the Chat tab.

A bid manager hears about a university's upcoming call for a student-wellbeing app.

1.  In **New Tender**, the bid manager chooses **Describe to AI**, names the tender "University wellbeing app" and continues. The assistant opens and asks about the opportunity.
    
2.  The bid manager answers: the university's name, a submission deadline of 30 June at noon, an estimated value of 250,000 EUR, a requirement for a WCAG 2.2 AA accessibility statement, and a 20-page limit.
    
3.  The assistant asks, in a **Decision required** card, whether the university's data must stay in the EU. The bid manager chooses **Yes** and chooses **Send answer**.
    
4.  The assistant records the details, the deadline, three draft requirements, a **Page limit** formality and a **Technical** output document, and names each step as it goes.
    
5.  The bid manager checks the **Requirements** tab, raises one requirement's priority to **High**, and asks the assistant: "Outline the technical proposal and find library answers for accessibility and data protection."
