An AI assistant can now discover a local business, check its reviews and availability, and book the appointment without the customer ever opening its homepage. That quietly moves the website’s main job from attracting visitors to feeding accurate, machine-readable information to the agents doing the recommending.
Sachin Puri, CEO of Newfold Digital, Bluehost’s parent company, frames the shift as a change of job, not the end of the website. “AI is actually not making websites irrelevant; it’s just changing its job,” he said.
Your website becomes the source of truth
When an assistant looks for an electrician under $300 with strong reviews and a slot tomorrow, it still has to answer the same questions a human would: service area, licence, price, ratings and availability. It may read your site directly, including schema markup, and cross-check Google Business Profile, Yelp and other reviews.
Puri calls this combined picture the business’s digital presence. The website is the one part of it the business controls, so it works as the reference point every other source should match. Consistency matters more than ever. Outdated pricing on your own page, in your schema, or in old third-party threads can follow you into an AI answer.
What 83.4 million daily crawls actually measure
Bluehost’s 90-day analysis found 83.4 million verified AI crawler requests per day across its hosted sites. But Puri warns against reading that as customer discovery. By Bluehost’s classification, roughly 90% of the requests were for model training, about 9% for search indexing, and only around 1% for live assistant fetching.
For ChatGPT and OpenAI specifically, Bluehost estimated one observable referral visit for every 5,350 verified HTML crawl requests. That is not a conversion rate—crawls include training and not every AI-influenced journey is visible—but it makes the point: machine consumption and human referral traffic are separate metrics. Keep measuring clicks, but do not mistake crawling for demand.
Track visibility, accuracy, and action
Puri’s framework gives marketers a practical score for AI search:
- Visibility: when a prospect asks a relevant question, does the business appear, get cited or recommended?
- Accuracy: are prices, services, locations and hours represented correctly?
- Action: does that exposure lead to a referral, inquiry, booking, purchase or revenue?
All three are needed. Appearing with the wrong price is useless; a correct answer that never turns into a booking does not pay. A request for a family-friendly, vegetarian-friendly restaurant with wheelchair access and a table for six tonight cannot be matched from a page that only promises “reliable service.” Track recommendations, not just mentions, because being named is not the same as being chosen.
Start with a 10-minute unbranded audit
Bluehost’s small business AI confidence study found only 13% of respondents were optimizing for search assistant visibility. That low number is an opening. Puri’s starting point is to spend 10 minutes a day acting like your customer: ask ChatGPT, Gemini or another assistant the questions a buyer would ask—without using your brand name. Check whether you appear, whether the facts are correct, which competitors show up, and what sources the assistant relies on. Then fix missing or outdated details.
The same study found 79% of business owners are aware of AI agents but only 16% have deployed one. Puri sees agents as broader than chatbots: they can answer calls, book appointments, send reminders and follow up. The time they save should go into service quality, customer relationships and asking happy customers for reviews—the signals AI assistants increasingly use as authority. Traditional funnels still work, but more of the decision now happens inside an assistant. The best move is to be understood, recommended, and chosen there.
Source: Search Engine Journal
