Engineering manufacturer: faster quotes and a CRM built around distributors
An illustrative $25M engineering manufacturer relies on two senior estimators for every quote, keeps distributor relationships in inboxes and spreadsheets, and pays for CRM seats nobody uses. The system captures RFQs in a structured way, uses AI to turn them into quote-ready briefs, prices routine work with encoded rules, and puts accounts, quotes and margin in one CRM connected to the ERP.
This is a worked example: a realistic scenario we've built to show how we'd approach the problem. The organisation is fictional, and every number is a stated assumption or a calculation from those assumptions. It is not a client case study and makes no claim about results.
The situation
Every custom quote passes through two senior estimators, who carry the pricing logic in their heads and in a large spreadsheet. RFQs arrive in sales reps' inboxes, with the estimators, at a general address and through distributors, often without material grades, quantities, tolerances or finish, so the first reply is usually a question.
Distributor relationships live in reps' inboxes and a shared spreadsheet of pricing tiers. The business pays for CRM seats, but quotes, orders and margins sit in the ERP, so the CRM shows a pipeline nobody trusts and the reps have stopped updating it.
The do-nothing baseline carries real risk. One estimator plans to retire within a few years, turnaround stretches whenever either is on leave, and as RFQ volume grows the business either declines work or quotes loosely and gives away margin.
Constraints we'd work within
- The ERP is the system of record for jobs, costing, inventory and invoicing, and it stays.
- Estimators keep the final say on price. No quote reaches a customer without a person approving it.
- Customer drawings and specifications are confidential, some under NDA, so they can't go into public AI tools.
- Distributor pricing tiers are commercially sensitive and need role-based access.
- Reps are technical and often on the road. Anything new must work on a phone and take less effort than email.
What the Diagnostic would find
Findings by lens, with the commercial effect of each. See how the Diagnostic works.
| Lens | Observation | Commercial effect |
|---|---|---|
| Operations & process | RFQs arrive through several routes and many are incomplete. There's no log of what has arrived, what's waiting on the customer and what's waiting on an estimator. | Turnaround is set by the busiest inbox, and slow quotes lose work to faster competitors, particularly on repeat parts where price and speed decide the order. |
| Workforce capability | Pricing logic for material, machine time, setup, finishing and customer tier sits with two people. Junior staff can't price even a repeat part without them. | Key-person risk sits directly on the revenue line, and the most experienced people spend time on the simplest quotes. |
| Systems & technology | The CRM was set up for transactional sales and holds no quotes, orders or distributor tiers. Most seats show little or no activity. | Licence fees are paid with no return, and there's still no shared view of key accounts or distributors. |
| Data & reporting | Finance sees margin per job in the ERP after the work is done. Sales can't see it when quoting, and win rates by sector, account or distributor aren't tracked. | Discounts are given without margin history, and there's no evidence for which distributors or sectors deserve more attention. |
| Strategy & commercial model | Distributors bring in a significant share of orders, yet they chase quote status by phone and email and get different answers depending on who they ask. Nobody owns the channel's performance. | Distributors put their effort behind suppliers who are easier to deal with, and the business can't see which partners deserve more support. |
| AI readiness & opportunity | RFQ emails and drawings follow recognisable patterns: title blocks, material callouts, quantities and required dates. | AI can turn a thread and its attachments into a structured brief for a person to check, so estimators start from a complete picture instead of a long email chain. |
The moves
Every recommendation resolves to one of nine moves: build, buy, improve, integrate, automate, remove, change, train or leave alone. Building is only one of them.
- Automate
RFQ intake. One address and a web form for customers and distributors that asks for material, quantity, tolerances, finish and delivery up front, with AI extraction of the details into a quote-ready brief and a same-day list of anything missing.
- Build
A quoting tool that encodes the estimators' pricing rules for material, machine time, setup, finishing, freight and margin by customer tier, so routine and repeat work is priced consistently and checked quickly.
- Build
A custom CRM shaped around how this business sells: key accounts, distributors, tiers and territories, RFQs, quotes, orders from the ERP, visit notes and follow-ups, usable on a phone.
- Integrate
The ERP and accounting data. Customers, items and cost rates are read from the ERP, accepted quotes are written back as jobs, and actual job costs return for quoted versus achieved margin.
- Buy
Business-grade AI models and document-reading services under terms that exclude training on the company's data. There's no case for building or hosting a model.
- Leave alone
The ERP itself. It works for jobs, costing, inventory and invoicing, the team knows it, and replacing it would add cost and risk without fixing the quoting problem.
- Change
Who quotes what. A trained quoting coordinator prices routine and repeat work in the tool, estimators review anything above a value or below a margin threshold, and they own every complex job. We train the coordinator and the reps as each part goes live.
- Remove
The unused CRM subscription at its next renewal, once data has moved across, and the shared pricing spreadsheet once the rules live in the quoting tool.
The system, end to end
Each RFQ is captured once, turned into a structured brief that a person checks, priced in the quoting tool and logged against the account or distributor in the CRM. Accepted quotes become jobs in the ERP, and actual costs flow back so the dashboard compares quoted and achieved margin.
- RFQ address, web form and distributor uploads
- AI extraction into a quote-ready brief
- Person checks the brief
- Quoting tool with pricing rules
- Estimator approval where required
- CRM record for the account or distributor
- ERP job, costing and invoice
- Pipeline and margin dashboard
Structured RFQ intake
One RFQ address and a web form that asks for the details estimators need and accepts drawings and models, so incomplete requests are caught at the door.
RFQ extraction assistant
Reads the email and attached drawings, pulls title-block data, materials, quantities and dates into a brief, and lists what's missing. Every field links back to its source so a person can check it quickly.
Quoting tool
Pricing rules for material, machine time, setup, finishing, freight and margin by customer tier. Estimator overrides are recorded with a reason, so the rules improve over time.
Distributor and key-account CRM
Accounts, sites, contacts, distributor tiers and territories, RFQs, quotes, orders and follow-ups, with role-based access to pricing and a mobile view for reps.
ERP integration
Customers, items and cost rates read from the ERP through supported methods, accepted quotes written back as jobs, and actual costs read back after completion.
Pipeline and margin dashboard
Open RFQs, quote turnaround, win rate, and quoted versus achieved margin by sector, account, distributor and estimator.
Delivery stages
Indicative timings.
Diagnostic and quote audit
Review a sample of past RFQs and quotes with the estimators, document the pricing logic, confirm ERP data access, and baseline turnaround, win rate and CRM use.
3 to 4 weeks (indicative)
Intake and extraction pilot
The RFQ form and address go live. The extraction assistant runs on real RFQs alongside the current process, and its accuracy is checked field by field.
4 to 6 weeks (indicative)
Quoting tool
Rules for routine and repeat parts first, tested against historical jobs until results sit within an agreed tolerance of estimator pricing, then used live with review.
6 to 10 weeks (indicative)
CRM and ERP integration
Accounts and distributors migrated from spreadsheets and the old CRM, ERP read and write-back connected, and the old CRM retired at renewal.
8 to 12 weeks, overlapping the quoting tool (indicative)
Margin reporting and refinement
Dashboard live, with a monthly review of rules, overrides and achieved margin with the estimators.
Ongoing (indicative)
The arithmetic
Assumptions first, then the calculations they produce. Change an assumption and the result changes with it, which is the point.
| Assumption | Value |
|---|---|
| RFQs received per month | 120 (assumption) |
| RFQs for routine or repeat parts | 50% (assumption) |
| Reading, clarifying and re-keying time per RFQ | 45 minutes today, 15 minutes with structured intake and AI extraction (assumption) |
| Estimator pricing time per routine or repeat RFQ | 60 minutes today, 15 minutes of review once the tool prices it (assumption) |
| Senior estimators | 2 (assumption) |
| CRM seats paid for | 12, at $130 a seat a month (assumption) |
| Average value of a quoted job | $15,000 (assumption) |
| Quote win rate | 25% today, 27% with faster, more consistent quotes (assumption, modelled scenario) |
Intake and clarification time
120 RFQs × (45 − 15) min = 3,600 min ÷ 60
60 hours a month released across estimators and sales (modelled)
Estimator time on routine quotes
120 × 50% = 60 RFQs × (60 − 15) min = 2,700 min = 45 hours a month ÷ 2 estimators = 22.5 hours each ÷ 4.33 weeks a month (52 ÷ 12)
About 5 hours a week each, moved to complex and higher-margin work (modelled)
CRM licence spend avoided
12 seats × $130 × 12 months
$18,720 a year, gross. To be set against the custom CRM's hosting and support, which are scoped separately (estimated)
Order value from a modelled win-rate change
120 × (27% − 25%) = 2.4 more orders a month × $15,000 = $36,000 a month × 12 months
About $432,000 a year in order value. Revenue, not profit, and a scenario rather than a forecast (modelled)
What we'd measure
- Median RFQ-to-quote turnaround, routine and complex
- Share of RFQs complete on arrival, with no clarification needed
- Win rate by sector, account and distributor
- Quoted versus achieved margin per job
- Estimator hours on complex versus routine quotes
- Weekly active CRM users and follow-ups completed on time
Risks and how we'd manage them
Pricing rules that drift from reality
We build the rules with the estimators from past jobs, test them against historical quotes before going live, record every override with a reason, and review the rules monthly against achieved margin.
Extraction errors on drawings
Drawings vary in quality and format. The assistant shows each value against its source and flags low-confidence fields, a person confirms every brief, and complex assemblies stay with the estimators.
ERP data quality and access
Item, cost and customer data may be inconsistent. We audit it in the Diagnostic, fix the fields the quoting tool depends on, and connect only through methods the ERP supports.
Adoption and confidentiality
Reps abandon CRMs that slow them down, so we design with them, log emails automatically and keep data entry to a minimum. Customer drawings are processed only under business AI terms that exclude training, with access controls and NDA jobs flagged.
The takeaway
The constraint isn't the estimators' skill. It's that every quote, simple or complex, queues for the same two people. Capture RFQs properly, encode the routine pricing and give the CRM a reason to exist by putting quotes and margin in it, and senior time goes back to the work that protects margin.
Services in this example
- Custom CRM developmentA CRM built around your pipeline, accounts, quoting and service history, with AI that drafts follow-ups and does the data entry. Owned by you, not rented.
- Business process automationAutomate intake, quoting, follow-up, onboarding and reporting across the systems you already use. Measured in hours released, with people on exceptions.
- System integrationsConnect your CRM, Xero, job system and email so nobody re-keys data. APIs, webhooks and sync rules, plus monitoring that alerts you to failures first.
- AI assistants & knowledge systemsRole-specific AI assistants that answer from your own documents, cite the source and respect permissions. Built on Microsoft 365, Google or a custom stack.
- Reporting & dashboardsOne source of truth for finance, sales, marketing and operations: agreed definitions, automated data pipelines and AI commentary that explains the numbers.
Recognise your business in this?
A 30-minute call is enough to tell whether the same moves would pay off for you.