AI agents
What each AI agent does, when it runs, and what it proposes for your approval.
Quote Hub has an agent for each part of the quoting process. Some start automatically when the work reaches them. Others start when you select a button. Either way, an agent prepares the work and proposes the next action; a person approves it in the AI Action Inbox.
Reads the email and attachments, extracts every line, matches client, contact, site and frame agreement.
Runs suitability and scope rules, explains gaps, drafts client questions, acknowledgement or decline.
Fills the D365 opportunity: prognosis, probability, sales unit and estimated revenue.
Drafts questions for the client and turns their answers into proposed updates.
Searches SAS stock and recycle inventory and recommends stock or external supply per line.
Builds the RFQ Upload Tool form and ranks suppliers from approved lists, LTSAs and history.
Drafts the supplier RFQ invitation and messages to suppliers.
Reads a supplier's PDF or Excel quotation into the quote form.
Drafts reminders to suppliers and clients on the follow-up schedule.
Recommends a supplier per line with reasons; flags single bids and deviations.
Benchmarks the margin against previous supply and the contract, writes TBE notes.
Writes the cover letter, applies the client's deviation process, runs the pre-submission checklist.
Summarises new enquiries, the Perth agenda, overdue responses and pending approvals.
Answers questions about an enquiry or the whole pipeline, using live data.
Intake and screening
Intake Agent
Runs automatically for each new email in the enquiry mailboxes and each new file in the shared folder, and when you add a client email by hand on the Intake mailbox.
It sorts each item into a new enquiry, a reply to an existing enquiry, a supplier quotation or noise. For a new enquiry it extracts the header and every line, matches the client, contact, site and frame agreement, and creates the enquiry. A client reply is added to the enquiry's clarification thread. It makes no proposals; screening starts as soon as the enquiry exists.
Screening Agent
Runs automatically after intake, and again after you approve updates from a client's reply. Button: Run screening or Re-screen in the next-action bar, Re-run screening on the Overview tab.
It applies the suitability and scope-of-supply rules from the Rules studio, then looks for gaps and ambiguities. It writes the suitability and scope report and proposes some of: acknowledge the enquiry, decline it, no-quote out-of-scope lines, table it at the Perth meeting, send clarification questions, or proceed to quote. See Screening.
Opportunity Agent
Runs automatically after screening when the enquiry is not being declined and has no opportunity yet. Button: Draft with AI or Re-draft on the D365 opportunity card.
It estimates revenue from previous supply and purchase history, looks at the client's win rate, and proposes the D365 opportunity fields with a reason for each. See D365 opportunity.
Clarification Agent
Runs automatically when the client replies to clarification questions, and when the Perth meeting decides to clarify. Button: Draft questions with AI on the Clarifications tab, Draft more questions in the next-action bar.
It drafts questions from the open screening findings. When the client replies, it reads the answers and any attachments, and proposes the field changes with the client's words as evidence. Questions still unanswered become a follow-up proposal. See Clarifications.
Stock and sourcing
Fulfilment Agent
Runs automatically when an enquiry proceeds to quote, from a proposal or from the Perth meeting. Button: Check stock with AI in the next-action bar, Check all lines with AI on the Stock (SAS) tab.
It searches SAS stock and recycle inventory for every line and proposes stock, partial or external supply per line. See Stock check.
Sourcing Strategist
Runs automatically after you approve a fulfilment route with at least one line not fully from stock. Button: Build RFQ with AI in the next-action bar, Build with AI on the Sourcing tab.
It runs the supplier cascade, ranks the candidates, picks up to three, and writes the RFQ Upload Tool form text without naming the end client. See Sourcing.
Hub Buyer Agent
Runs automatically when an RFQ Upload Tool form is sent to the hub. Button: AI draft RFQ on the hub request.
It drafts the cover message for the online RFQ and proposes sending it to the shortlisted suppliers. The hub buyer approves. See Hub sourcing.
Quote Reader
Runs when a hub buyer uploads a quotation received by email, and when a supplier uses Auto-fill from document in the supplier portal.
For an emailed quotation it proposes recording the extracted quote. In the portal it fills the supplier's form, and the supplier checks and submits it. Also runs automatically when a submitted quote has deviations or alternatives: it compares each one with the client specification and records an internal note in the agent trace.
Follow-up Agent
Runs automatically every 15 minutes, following the follow-up rules set by an administrator.
It proposes reminders to suppliers who have not quoted, to clients who have not answered clarification questions, and to clients who have not responded to a sent quotation. It skips any RFQ, thread or quotation that already has a reminder waiting for approval. See Automation.
Evaluation, costing and quotation
Bid Evaluation Agent
Button only: Evaluate bids with AI in the next-action bar, AI recommendation on the Bid check tab.
It scores the offers, chooses the objective and weights, recommends a supplier per line with reasons, lists risks, and drafts a single-bid justification where only one supplier quoted. See Bid check.
Costing Agent
Runs automatically when a supplier selection is confirmed. Button: AI costing review in the next-action bar or on the Costing & TBE tab.
It compares margins with previous sales and the contract margin band, proposes margin changes where justified, and writes TBE notes. See Costing & TBE.
Quotation Agent
Button only: Prepare submission in the next-action bar, AI cover letter on the Quotation tab.
It writes the cover letter in line with the client's deviation process, re-renders the quotation PDF and runs a pre-submission checklist. If the quotation has passed approvals, it also proposes submitting it, with the checklist as its reasoning. See Quotation.
Assistants
- Daily Briefing prepares your priorities on the dashboard. See Daily briefing.
- Copilot answers questions and can draft a small set of proposals. See Ask Quote Hub.
How agents stay within the rules
- Propose, don't act. Anything that emails a client or supplier, changes enquiry data, or moves the work forward is a proposal. It runs only after a person approves it, and runs through the same steps as the manual action.
- Own work only. Without approval, agents record their own output: the screening report, a draft cover letter, notes in the trace. When screening or the stock check starts, the enquiry moves to that stage.
- Rules first. Screening applies the Rules studio rules before any AI reasoning, and the AI verdict can only make the result stricter. Supplier ranking starts from the cascade, and excluded suppliers are never picked.
- Confidential by default. Text for suppliers never names the end client, its sites or its projects.
- Fully traced. Every run, step and decision is on the enquiry's Agent trace tab.
- Open enquiries only. Agent buttons are hidden on won, lost and closed enquiries.
Administrators can see every agent, its runs and its cost on AI agents & usage.