ChatGPT's new AI price estimates could underbid you first

A customer now gets a price estimate from ChatGPT before they ever reach your quote form, and that number sticks whether it's right or wrong.

OpenAI's GPT-6 update started rolling out on October 7 to Plus, Pro, Business, and Enterprise users, then expanded to Free and Go accounts on October 8. The headline change is what OpenAI calls an intelligent UI: GPT-6 can now answer with charts, tappable controls, or a small built-in tool instead of plain text. OpenAI's own examples include a savings calculator and a bill-splitting tool, generated on the spot inside the chat window, no separate app or site required.

What GPT-6's intelligent UI actually does

The model decides case by case whether a question needs a quick calculator, a side-by-side comparison, or just a sentence. Responses also start appearing while GPT-6 is still thinking, so for search-heavy questions it begins answering about 44 percent sooner than the prior model, by OpenAI's own measure. None of this requires the business being asked about to do anything. The tool builds itself from whatever pricing, sizing, or cost assumptions the model already holds.

Why AI price estimates become the anchor price

For a service business, the risk isn't that ChatGPT gets the number wildly wrong. It's that a customer now forms an opinion before you ever hear from them. Ask GPT-6 what a kitchen remodel, a move across town, or a fence install runs, and it hands back an interactive range in seconds, built from whatever public averages it has seen. That range becomes the customer's starting point. If your real quote lands higher, and it often will once you account for their actual square footage or site conditions, you're now negotiating down from their number instead of presenting your own.

What to do about it this week

1. 1

   Find your anchor

   Ask GPT-6 the same cost question a prospect would. Note the range it gives back, since that's what you're now quoting against.

2. 2

   Publish the real range

   Put an honest, specific price range in plain language on your own site. Clear public pricing is what eventually corrects a generic AI estimate, not a complaint to OpenAI.

3. 3

   Make your real quote faster than theirs

   If ChatGPT answers in seconds, your own quote tool or callback needs to land the same day. A slow follow-up lets the generic number sit unchallenged.

Where this cuts the other way

The same feature can work for you. A business with clear, structured pricing content on its own site gives GPT-6 something accurate to draw from instead of a generic average, and a fast, specific follow-up still beats an anonymous chatbot estimate every time. The businesses that lose here aren't the ones AI mentions. They're the ones who never checked what AI was already saying about their price before a customer did. If you want a second look at what your own quote workflow is up against, that's worth comparing notes on.