A quote looks like a document. That is why so many quoting projects start with templates, PDF output, and a nicer screen for sales.
The document is the easy part. The hard part is the answer inside it.
A credible B2B quote is the moment several operational facts have to become one authorized commercial statement: who the customer is, what they are allowed to buy, at what contract or negotiated price, in which unit, in which configuration, with what is actually available, at what quantity break, with what freight, at what margin, under whose approval, against what credit, with what substitution if the specified item cannot ship.
If a salesperson still looks those facts up in the ERP, a spreadsheet, an email thread, and a colleague who "knows the account," the company does not have a quoting system. It has a typing exercise attached to tribal knowledge.
A new CPQ interface does not change that. It often relocates the same reconstruction into a more expensive form.
The Quote Is an Authorization, Not a Layout
Consumer checkout can show a list price because the commercial rule is simple. B2B commerce is not a consumer storefront with a login. The price on a quote is frequently account-specific, contract-bound, quantity-sensitive, and exception-prone.
The person assembling the quote is not writing prose. They are deciding, or guessing:
- which customer record is the buying account
- whether this SKU is on the customer's catalog
- which UOM the warehouse will ship and which UOM the contract is written in
- whether the configuration is valid
- whether inventory is available to promise, or only visible in one location
- whether a substitute is commercially allowed
- whether the discount is inside policy or needs finance
- whether credit will hold after the order is booked
When those answers live in people's heads and in the ERP, two failures show up immediately. The quote is slow, because the facts have to be collected. The quote is fragile, because two people can collect them differently.
Buyers feel both as delay and as revisions. Internally it shows up as margin leakage, credit surprises, and orders that cannot be fulfilled as quoted.
Why Generic CPQ Still Becomes a Spreadsheet
Off-the-shelf CPQ is often bought to "standardize quoting." It works when the commercial model is already simple enough to configure: a short catalog, stable list prices, few exceptions, a single approval path.
It struggles when the real rules are the ones a mid-market manufacturer or distributor actually runs. Customer-specific price files. Pack-to-each conversion. Kits whose components are priced one way and sold another. Freight that depends on destination and weight. A margin floor that only certain roles may cross. A contract that expires. A customer part number that maps to three internal SKUs.
In those businesses, the CPQ project has a familiar shape. The tool is implemented. Sales still exports from ERP to check the price. Operations still confirms availability in another screen. Finance still approves in email because the threshold logic never matched policy. The CPQ becomes a document generator sitting beside the systems that actually know the truth.
That is not a vendor-brand problem. It is a design problem. The software was asked to produce a quote. The business needed its commercial rules to become executable across the systems that already own them.
Make the Rules Run, Then Generate the Document
Quoting automation works when the rules that already govern the business can be applied without a person assembling them.
That means identity first: the quote is bound to an account, a contract, and a selling unit, not to a guest cart. It means price is retrieved from the system of record at the moment of the quote, not copied from last quarter's file. It means configuration cannot produce a combination the plant cannot build or the warehouse cannot ship. It means availability is the same ATP the order will later consume. It means approval is a routed decision with a recorded authority, not a CC line.
The document, the portal view, and the email are outputs of that execution. They should not be the process.
Inquiry to invoice is the cycle that follows: request, quote, order, fulfillment, invoice. This article stops at the authorized commercial answer. That one starts when the answer has to travel through the rest of the revenue path without being rebuilt.
If a buyer or an agent is going to request a price, they inherit the same requirement. What agentic commerce means for manufacturers and distributors is not that machines should invent a discount. It is that machines can only quote what the business has made machine-readable and authorized.
What AI May Do — and Must Not Do
AI can help a quoting team see context. It can surface that similar deals closed at a given discount band. It can flag an exception that looks unusual. It can draft a narrative around a configuration a sales engineer already validated. It can help a person prepare for a negotiation.
It must not invent the price. It must not invent commercial terms. It must not grant itself approval authority.
Those are policy objects. They belong to contract, ERP, credit, and named human roles. A model that "finds a competitive number" without being bound to those objects is not assistance. It is a new source of leakage, dressed as speed.
The Arizon position is the same one that applies to other operational automation: use deterministic rules for deterministic commercial logic. Use intelligence where judgment is actually required — which deals to review, which exceptions are dangerous, which substitutions to offer a person. Do not let a generated sentence become a commitment the warehouse and finance cannot honor.
The Test Is Whether Anyone Still Has to Reconstruct the Answer
A quoting program has worked when a routine request produces an authorized price without a scavenger hunt, and when an exceptional request arrives to the right person with the account, inventory, margin, and policy already attached.
If sales still opens the ERP to "make sure," the interface did not fail. The rules never became executable.
Arizon Digital's Agentic Commerce Readiness Index treats pricing, quoting, and negotiation as a distinct capability for that reason. Agents, portals, and salespeople all need the same thing: a price the business has already agreed it is allowed to give. The assessment is not a CPQ. The work is connecting identity, contract, configuration, availability, and authority so the quote is an execution of policy — not a reconstruction of it.
The question for leadership is narrow. When a buyer asks what it will cost, does the company already know, in systems, or does a person have to go find out?