Why Is “Contact Us for Pricing” a Dead End for AI Agents?

How do agents use pricing?

• Shortlisting. Asked for "project management tools under $15 a user", an agent filters by price. Without one, you are filtered out or listed with a caveat.
• Comparison tables. Agents build side-by-side comparisons. An empty price cell next to competitors' numbers reads as a warning.
• Budget checks. Asked whether an option fits a stated budget, an agent needs at least a range.
• Next steps. An agent can start a free trial, add a product to a basket or request a quote through a form built for machines. It cannot sit through a discovery call.

When the price is missing, agents often reach for whatever figure they can find: a review site's estimate, an old article, a reseller's price. How that produces wrong answers is covered in why Google AI Overviews hallucinate prices, competitors or business details.

What should you publish if prices are negotiated?

You do not have to publish your final negotiated prices. Publishing the structure lets agents and buyers estimate, and it keeps you in the comparison. How to write SaaS pricing pages so assistants describe them correctly is covered in how SaaS pricing pages should be written so AI describes them correctly.

Can AI agents understand negotiated B2B pricing and account-specific terms?

Only if they are given access to them. An agent searching the public web sees only what you publish, so it can know your list prices and discount rules, not a customer's contract. Account-specific pricing becomes available to agents in two ways:

• Through the buyer's systems. Procurement platforms already hold contract prices, often through punchout catalogues, which link a buyer's procurement system to a supplier's catalogue using the cXML standard. An agent working inside the buyer's procurement system can use those negotiated prices.
• Through your authenticated channels. A customer portal, price list API or a connector built on the Model Context Protocol, the open standard Anthropic introduced in 2024 for connecting AI agents to external systems. These can let a customer's agent retrieve that customer's own prices and terms securely.

For prospects, the public structure is what matters. For existing customers, machine-readable access to their own terms is what keeps reorders and renewals smooth when agents do the buying.

What about "request a quote" forms?

Keep them for complex deals, but make the path clear for agents and the people using them:

• Show what the quote depends on, so buyers can supply it up front.
• Keep the form short, with clear labels on every field and no steps that only work with a mouse.
• Say how quickly a quote arrives, and aim for instant estimates where possible.
• Offer an alternative: a published starting price or a calculator for buyers who only need a ballpark.

How agents read and use forms is covered in how brands can make their websites readable to AI agents.

Which businesses find this hardest, and what can they publish?

• Enterprise software: publish a starting price, the main pricing unit and typical ranges by company size.
• Professional services: publish fixed fees for common packages, day rates or hourly ranges, and what drives the cost of bespoke work.
• Manufacturing and wholesale: publish price bands by quantity, minimum order quantities and lead times, and offer instant quotes for standard items.
• Construction and home services: publish typical ranges for common jobs by size, and what a survey or site visit adds.
• Healthcare and regulated services: publish self-pay prices where rules allow, and explain how insurance or funding affects the cost.

Each of these has a reason not to publish a single number. None has a reason to publish nothing.

How should sales teams adapt?

When prices are public, the first call changes. Buyers, and the agents working for them, arrive knowing the range and asking about fit, terms and the details of their case. Train the sales team to start from the published structure and discuss the specifics. Record which questions still come up on calls, and add the answers to the pricing page. Over time, the page should answer everything except the negotiation itself.

How do you check the effect?

Ask ChatGPT, Gemini, Perplexity and Copilot questions buyers ask: "How much does [product] cost?", "[category] tools under [budget]", "[you] vs [competitor] pricing". Record whether answers state your price correctly, give a range, say "not published", or quote a third-party estimate. After publishing starting prices and ranges, recheck. Track whether you now appear in budget-filtered answers, and whether sales calls start with fewer basic pricing questions.

What might an illustrative case look like?

Consider an illustrative facilities management software vendor whose pricing page said only "Every organisation is different. Contact us." AI answers to "facilities management software pricing" gave competitors' prices and described the vendor as "enterprise pricing, likely expensive". A few cited a three-year-old review estimate that was far too high. The vendor published a starting price per building, a typical range for mid-sized estates, and its discount rules for multi-year contracts, and added a calculator. Within two months, answers quoted its starting price and range. The vendor also began appearing in answers to budget questions it had previously been left out of.

What do people ask about pricing and AI agents?

Will publishing prices weaken our negotiating position?

Publishing a starting price and structure rarely does. Buyers already see competitors' prices, and a range sets expectations early.

What if our prices change often?

Publish the structure and update the figures when they change. Stale prices cause more trouble than ranges.

Do agents actually buy B2B products today?

Some tasks, such as reordering and renewals, are increasingly automated inside procurement systems. Research and shortlisting by AI is already common.

Should prices be in structured data too?

For products sold online, yes, matching the visible price. For services and negotiated offers, a clear text statement of the starting price and range is what agents and assistants read.