AI CPQ for manufacturing that starts from your industry, not a blank slate.
Generic CPQ hands you an empty rules engine and six months of implementation. ServiceCPQ starts from a pre-packaged knowledge graph for your industry — its products, options, rules, terminology and vocabulary already modelled — so you improve on a working configuration instead of building from scratch. Conversational, knowledge-graph-driven, and adaptive to the way you sell: configure-, make-, measure- or engineer-to-order.
Forge AI is the engine behind your AI CPQ.
Forge AI — the knowledge-graph AI engine inside ServiceCPQ — reads a plain-language request, matches your products, and builds a complete quote across configure-, make-, measure- and engineer-to-order.

Describe it in plain language.Tell Forge AI what the customer needs — “an 8-lane bowling alley near the food court, 5000 × 2000 ft” — and it organises the RFQ and starts matching products.

Forge AI reads and classifies it. It reads the RFQ, classifies every line as CTO, MTO or ETO, checks each choice against the knowledge graph, and prices what it can — one guided feed instead of a form.
One request. One feed. One decision at a time.
Forge AI turns every request into a guided feed. Each step appears when it’s needed and collapses to a single line once it’s answered — so nobody faces a forty-field form.
The request is read in the customer's own words and shown back as facts you can correct — application, site, duty, quantity.
Standard lines are matched to your catalogue; lines that need sizing or engineering are separated out, each marked CTO, MTO or ETO.
The right product is chosen, then its options, group by group. Required items accept in one click; optional ones stay out of the way.
Every choice is validated against the knowledge graph as you make it. Conflicts appear the moment they happen, with the reason.
Only the lines that need an engineer go to one, with the closest past jobs attached. The rest of the quote keeps moving.
Standard, built-to-order and engineered lines come back together as one quote.
A blank slate is not a head start.
Most CPQ tools are empty rules engines. You spend months teaching them your products before they return a single quote — and re-teaching them every time the catalogue changes.
Every rule is hand-coded up front and re-coded every time your products change. You maintain the tool instead of selling with it.
You wait half a year to even see how the product behaves for your business — configuring a blank slate before a single real quote goes out.
Generic tools don't know your terminology, options or the relationships between them, so every vocabulary and rule has to be taught from zero.
A flat option list can't model made-to-measure or engineered-to-spec products — the exact places where quoting is hardest and margin is won or lost.
You start from a working model of your industry.
ServiceCPQ ships with a pre-packaged knowledge graph for your sector — so the configurator already understands your products, rules and language on day one, and you refine it instead of building it.
Ships with your sector's products, options, compatibility rules, terminology and vocabulary already modelled — so you start from a working configuration and improve it, not build from scratch.
Variants are generated from the knowledge base and its rules — not hand-built one at a time — so the catalogue stays consistent as the product line grows.
The graph adjusts to the mode you're actually in — configure-, make-, measure- or engineer-to-order — and applies the right workflow and validation automatically.
The industry graph carries the vocabulary of your sector, so your sales team, your dealers and your buyers all speak the same language in the quote.

Start from a working model: upload a product sheet, spec PDF or ERP export and the AI extracts feature groups, options, prices and EXCLUDES/REQUIRES compatibility rules into the knowledge graph — instead of hand-coding every rule.
One graph adapts to your manufacturing mode.
The same knowledge graph reasons across configure-, make-, measure- and engineer-to-order — and routes each quote to the right workflow automatically.
A defined range with selectable options. Rules resolve valid combinations into a locked, priceable BOM — no engineering involvement for standard variants.
A base platform with non-standard variants. Flagged items route to engineering review; the quote is conditional on sign-off.
Dimensioned, measured products — custom windows, cabinets, fenestration and building products configured to exact size, then matched to the right maker.
Every project designed to spec. The closest past jobs attach as a starting point, only the lines that need an engineer go to one, and the rest of the quote keeps moving.

Dealer workspace. Describe what the customer needs or upload their RFQ; each line is matched and configured, with dealer net and margin in view throughout.

Engineer-to-order. Measured and specialist items routed to an expert — site survey, tolerances and compliance confirmed before pricing, then released back into the quote.
Everyone works the same quote, in their own language.

A live feed of incoming requests. Configure line by line with checks, lead time and price always in view.
Describe what the customer needs, or upload their RFQ. Low-confidence matches are flagged, and dealer net and margin stay visible throughout.
Plain language, no part numbers. Say what you need, answer one question at a time, and track the order after you accept.
Only the lines that couldn't be quoted as standard — each with the reason it arrived, similar past jobs, a feasibility decision and a hand-back.
Conversational input, AI matching, graph reasoning.
The Forge AI intelligence that lets a rep, a dealer or a retailer build a correct, priced configuration in minutes — even for measured and engineered products.
Describe what the customer needs in plain language. The AI asks only the questions that matter and returns a valid, priced configuration — no rigid forms.
Matches each request against your own historical quotes with similarity scores, so repeat and near-repeat work is quoted fast — and gets sharper as your history grows.
For catalogue companies and retailers selling manufacturers' custom products — a DIY retailer selling made-to-measure windows or cabinets — the engine maps each measured requirement to the right manufacturer, product and supplier automatically.
Products, options, rules, pricing and the installed base as one connected graph, so every quote is consistent, explainable and correct across channels.
See what you’re quoting.
A 2D and 3D view sits beside every quote, drawn from the configuration itself. Change a dimension, add an attachment or fit a guard, and the view updates with it — what the customer sees is what’s being quoted.

Customer self-service.Describe the goal — “a fleet for an open-cast coal operation” — and the AI maps it to the right configuration.

One request, parallel tracks, one quote. A live quote-request feed, the Forge AI assistant that reads a project or RFQ, and purpose-built workspaces for sales, dealer, customer and expert — all on one platform.
Keep it. Forge AI works in front of it.
Your CPQ stays the system of record for prices, rules and orders. Forge AI handles the requests it was never built for — the RFQ that doesn’t fit the catalogue, the dealer who needs an answer today, the engineered line that takes your team a week to price. Nothing is ripped out. Start with one product line and one dealer group.
A complete solution quote — as if you’re sitting in front of the customer.
ServiceCPQ doesn’t stop at the base product. It builds the whole solution — accessories, add-ons, attachments and after-sales — with margin in view as it comes together, the discount kept within policy, and room to negotiate line by line, the way a great salesperson would across the desk.
AI-native CPQ vs a generic rules engine.
The difference isn't a longer feature list — it's where you start. A pre-packaged industry knowledge graph, versus an empty tool you configure for six months.
| Capability | Generic / rules-based CPQ | ServiceCPQ |
|---|---|---|
| Starting point | Empty rules engine, 6-month build | Pre-packaged industry knowledge graph |
| Industry terminology & vocabulary | Taught from scratch | Built into the industry graph |
| Variant creation | Hand-coded rules per variant | Generated from the knowledge base |
| Configuration modes | One mode, hand-coded | CTO · MTO · MTM · ETO on one graph |
| Conversational input | Rigid forms and picklists | Natural-language, AI-guided |
| AI quote & supplier match | Manual lookup | Similarity-scored against your history + supplier match for measured products |
| Solution quote with margin in view | Base product, manual pricing | Accessories and after-sales in one quote, margin in view, discount within policy |
Frequently asked questions
Generic CPQ hands you an empty rules engine and a six-month implementation to configure it. ServiceCPQ starts from a pre-packaged knowledge graph for your industry — products, options, rules, terminology and vocabulary already modelled — so you refine a working configuration instead of building from a blank slate. It's conversational, knowledge-graph-driven, and adapts to how you actually sell.
It's a connected model of your products, options, compatibility and pricing rules, industry terminology and installed base. Instead of hand-coding and re-coding rules, ServiceCPQ's AI reasons over the graph — which is why it handles CTO, MTO, MTM and ETO on one engine and stays consistent as your products change.
Yes. Make-to-measure covers dimensioned products configured to exact size — custom windows, cabinets and fenestration. Engineer-to-order attaches the closest past jobs as a starting point, routes only the lines that need an engineer to one, and keeps the rest of the quote moving while they're in review.
Yes. AI supplier match is built for exactly that — a DIY retailer or catalogue company selling made-to-measure windows, cabinets or any product customised to measurement. The engine maps each measured requirement to the right manufacturer, product and supplier automatically, so the retailer quotes a custom product without owning the factory.
It's designed to go live in weeks rather than months. Because you start from a pre-packaged industry knowledge graph rather than a blank slate, the first working configurator reflects your product family early — then you refine the configuration on top of a model that already works for your sector.
No. Forge AI can work in front of your existing CPQ, which stays the system of record for prices, rules and orders. Most customers start with the requests their current configurator handles worst — non-standard RFQs and dealer quoting — on one product line and one dealer group.
Only those lines are handed off. The engineer sees why the line couldn't be quoted as standard, what's already decided, and the closest past jobs to start from. The standard lines can be quoted while the engineered ones are in review.
All three, plus engineering. Each has a workspace built for what they need to do — a sales advisor feed, a dealer workspace with net and margin in view, a plain-language customer portal, and an engineering queue — and all of them work on the same quote.
Every option is checked against the knowledge graph as it's chosen. Combinations that conflict are flagged the moment they happen, with the rule behind them, so an invalid configuration never reaches the customer.