---
title: "Lead to Quote: AI Workflow Blueprint | QCS"
description: "An AI workflow blueprint on ERPNext: capture requests, collect missing details and apply pricing, with people approving every quotation."
url: https://quarkcs.com/agentic/lead-to-quote
---
Workflow 1 of 5 · Sales

# From first enquiry to approved quote. *With your team in control.*

Capture requests from WhatsApp, email or your sales team. Collect missing details, apply your pricing rules and prepare the quote, with separate approvals for the quote and the outgoing response.

[Discuss your lead-to-quote pilot](https://quarkcs.com/contact?interest=pilot&workflow=lead-to-quote) [See the architecture behind it](https://quarkcs.com/agentic/architecture)

Fixed scope and success measures, agreed before any build.

The workflow in three moves

1. **Capture & qualify** Ask for missing information automatically.
2. **Build & approve the quote** Apply rules, audit with AI and resolve review comments.
3. **Approve the response & send** Check the message, recipient and approved attachment.

Two separate approval queues: **A · the quotation** and **B · the outgoing response**.

Related QCS delivery **Quotation workflows for Location Solutions**

Our published ERP case covers product bundles, costing and quotation workflows for a UAE fleet technology company. The AI stages on this page are a proposed pilot on top of that kind of foundation.

[Read the delivery story →](https://quarkcs.com/blog/success-stories/shaping-the-future-of-quotations-how-tech-driven-efficiency-is-revolutionizing-sales)

The workflow, stage by stage

## Lead to Quote. *From the first message to the recorded result.*

Proposed pilot design: the fields, channels, rules and approval roles are agreed for your business before anything is built.

**9 stages.** The first is open; the rest expand as you go, or all at once.

1. 01 **Receive the request from any channel** WhatsApp · Email · Manually logged lead Capture

   Create or update the lead from the incoming request. Attach the original message, documents and channel reference; link subsequent messages to the same enquiry.

   Record the customer, contact details and preferred reply channel. Check for an existing lead so the same request is not worked twice.

   **Ready for qualification**

   One lead, its source conversation and the information supplied so far.

   Next: qualify the lead →
2. 02 **Qualify the lead and request missing information** Automatic clarification on the permitted channel AI + checklist

   AI extracts the supplied information and checks the required-field checklist for this type of enquiry: for example, item or service, quantity, specifications, delivery location and required date.

   If data is missing, automatically send a focused request by email or WhatsApp using the permitted channel and agreed clarification templates. Ask only for what is still missing. Read the reply, update the lead and qualify it again.

   Count customer replies during qualification; the initial enquiry and outgoing clarification requests do not count as replies.

   **Complete data → pricing**

   Proceed to stage 4 as soon as the required information is complete.

   **Still missing → ask or hand over**

   After reply 1 or 2, request the remaining information. If reply 3 is still incomplete, stop the automatic loop.

   See the handover rule →
3. 03 **Hand over when more than two replies are insufficient** Conditional · The third customer reply still leaves required data missing Sales owner

   Stop further automated clarification and put the lead in the named sales owner's queue. Show the conversation, questions already asked, reply count, confirmed fields and outstanding gaps.

   The person contacts the customer, resolves ambiguity or holds the lead. Qualification resumes only when the owner releases it with the required data.

   **Human handover with context**

   No repeated automatic questioning after the threshold is crossed. A missing reply is not counted as a customer reply.

   When qualified: apply the rules →
4. 04 **Apply pricing and other static rules** Business calculations and configured validations Rules

   Apply the relevant price list, quantity breaks, permitted discounts, taxes, currency, margin thresholds and standard terms through the business system's configured rules. Check the required stock or delivery constraints.

   AI does not invent prices or override failed rules. Unmatched items and exceptions are made visible for review.

   **A calculated quotation draft**

   Line items, amounts, terms and rule results are recorded against this version of the quote.

   Next: AI audit and enhancement →
5. 05 **Audit the draft with AI and enhance it** Completeness, consistency and customer-facing clarity AI review

   Compare the draft with the customer's request. Look for omitted requirements, inconsistent quantities or units, unclear descriptions, conflicting delivery assumptions and missing supporting information.

   Improve wording and presentation within the approved scope. Record proposed commercial changes separately; changes to items, amounts or terms return through the static rules. Unresolved issues remain visible to the approver.

   **A reviewed draft with findings**

   The reviewer sees the quote, rule results, AI findings and suggested improvements alongside the original request.

   Next: quotation approval queue →
6. 06 **Add the quotation to approval queue A** Approve the commercial document before composing its response Quote approver

   Present the exact quote version, qualified lead, pricing and rule checks, AI findings and any unresolved exceptions. The approver can approve, request changes, or reject and hold. A failed mandatory rule blocks approval until corrected or resolved through an authorised exception.

   Approval applies to that quote version. Editing the commercial content makes its prior approval invalid and requires review again.

   **Approve → stage 8**

   Use the approved quote to draft the response.

   **Request changes → stage 7**

   Send specific review comments to the revision loop.

   **Reject or hold**

   Keep the enquiry with its owner; do not send a quote.

   Next: Use AI to make reviewer-requested changes →
7. 07 **Use AI to make reviewer-requested changes** Conditional · Changes requested in quotation review AI revision

   Turn the reviewer's instructions into edits to the quotation draft. Preserve the comments and show what changed from the previous version. Ask the reviewer when an instruction is ambiguous.

   Reapply the pricing and static rules, repeat the AI audit, then return the revised quote to queue A. AI revision does not approve its own work.

   **Revise → rules → audit → approve again**

   Stages 4 → 5 → 6 repeat for the new quotation version.

   Next: Draft the response and add it to approval queue B →
8. 08 **Draft the response and add it to approval queue B** Only after the quotation is approved in queue A Response approver

   AI drafts the email or WhatsApp response from the approved quote: the cover message, recipient, channel and the correct quote attachment. It must reflect the approved amounts and terms.

   A person reviews this outgoing package in a separate queue. Message-only edits return to queue B after revision. Changes to the quotation's commercial content return through stages 4 to 6 before a new response is prepared.

   **Two approvals, two decisions**

   A approves the quote. B approves what is sent, to whom and through which channel.

   **Response approved → stage 9**

   The message and attachment are ready for dispatch.

   Next: send the approved quote →
9. 09 **Send the quote and record the delivery result** Approved quotation + approved outgoing response Dispatch

   Confirm that both approvals still match the current quotation, message, recipient and attachment. Send through the approved email or WhatsApp channel.

   Record the message reference, quote version, recipient, time and delivery status on the lead. If sending fails or its result is uncertain, retain that state and reconcile it before retrying to avoid duplicate sends.

   **Sent, failed or awaiting confirmation**

   The owner can see what was sent and what the channel reported. Failed delivery is not marked as successful.

#### Where staff time is recovered

Capturing enquiries, requesting missing details, assembling quotes, checking consistency, applying reviewer edits and drafting customer responses.

#### What we measure in the pilot

Hands-on minutes per enquiry, clarification rounds and handovers, time in each approval queue, revision cycles and time to an approved quote being sent.

#### Does this match your process?

Bring one real case and your current approval roles.

[Bring one case →](https://quarkcs.com/contact?interest=pilot&workflow=lead-to-quote)

Is it worth doing?

## Put your own numbers in. *See the hours.*

The pilot is measured against this baseline. We report the share of these hours it actually removes, with quality and rework alongside.

Measured in the pilot Correct completion Hands-on minutes Waiting time Rework Cost per outcome

The other workflows

## Same pattern, *different department.*

[All five workflows →](https://quarkcs.com/agentic#workflows)

[Finance · Accounts payable **Supplier Invoice to Payment** Invoices read on arrival, three-way matched by your tolerances, posted by the ERP, paid in an approved run with remittance sent back on the same channel.](https://quarkcs.com/agentic/invoice-to-payment)
[Customer service **Complaint to Resolution** Complaints attached to the order on arrival, evidence gathered, resolution proposed by policy, approved when it matters, posted by the ERP and confirmed on the customer's channel.](https://quarkcs.com/agentic/complaint-to-resolution)
[After-sales · Workshop **Service Booking to Invoice** Bookings qualified from the message, slots and estimates proposed by your rules, approved by the customer, and the job card, parts and invoice handled in the ERP.](https://quarkcs.com/agentic/service-to-invoice)
[Finance · Receivables **Collections** Reminders by age and amount on the customer's channel, promises to pay recorded from replies, payments matched on arrival, and credit holds approved by a person.](https://quarkcs.com/agentic/collections)

Bring one real case and your current approval roles

## Start with your lead to quote process.

We agree the scope, the controls and the evidence a pilot must produce. See [how a pilot runs](https://quarkcs.com/agentic#pilot).

[Discuss your lead-to-quote pilot](https://quarkcs.com/contact?interest=pilot&workflow=lead-to-quote) [Estimate its monthly running cost](https://quarkcs.com/automation-cost-estimator#e=eyJ2IjoxLCJkIjpbInNhbGVzIl0sIndmIjp7ImxlYWQtdG8tcXVvdGUiOnsibyI6MSwidiI6MzAwfX19)
