Free guide · Customers and products
How customers find you now
AI answers are changing how buyers search, and this chapter separates what has been measured from what is guesswork, then gives you four habits that keep your business findable whichever way search goes.
Shubham Bapna · Added · Free
The short version
Adapted from Chapter 11 of my book, Unleash Your Business With AI. Where the text mentions another chapter, it means a chapter of the book.
In July 2026 Outlook Business reported on how small businesses in India were coping with Google, ChatGPT and the worry about both. Two of the owners it spoke to are worth meeting. [1]
Vinay Singh runs Pinnacle Realty in Lucknow. He said two or three customers had recently told him they found the firm through ChatGPT. The firm has a website, pages for individual localities, a Google Business profile and listings on property portals.
Mohammad Amir runs an AC repair business with 10 employees. It has no website. He said roughly one in three of his enquiry calls come through his Google listing.
These are two owners describing their own businesses to a journalist. Nobody audited the numbers, and two anecdotes prove nothing about your trade. They are useful because they show both ends of the same change. One buyer asked an AI assistant a question and was given a firm's name. Another business, with no website at all, gets a large share of its calls from a single listing.
What the two have in common is plain, accurate, public information about the business, sitting where a machine can read it. That is the subject of this chapter. It is also a chapter built on fast-moving facts, so every figure carries its date, and the advice is chosen to stay right if the figures change.
The click that never comes
For years search worked one way. The customer typed a question, Google showed a list of links, and the customer clicked one. Now Google often writes an answer at the top of the page, and assistants such as ChatGPT, Gemini and Claude answer in a chat. The customer may get what they need without visiting anyone's website.
The clearest measurement in the research for this book is from the United States. Pew Research Center tracked the browsing of 900 US adults during March 2025, covering 68,879 Google searches. When the results page carried an AI summary, people clicked a normal result link on 8% of visits. Without a summary it was 15%. They clicked a link inside the AI summary on 1% of visits. They ended the browsing session after 26% of pages with a summary, against 16% without. In that month 18% of searches produced an AI summary. [2]
A second US measure: a SparkToro study using Similarweb data, as reported by Search Engine Journal, found that 68% of US Google searches ended without any click in January to April 2026. For every 1,000 US Google searches, 232 clicks reached the open web. The 2024 figure was 360. [3]
One vendor study points the same way. Ahrefs, which sells search software, compared 300,000 keywords between December 2023 and December 2025 and found the top-ranked page had a 58% lower click-through rate when an AI Overview was present. Its authors describe this as a correlation. [4]
Now India. Google's AI Mode launched here in June 2025, the first market after the US. Hindi was added in September 2025, and seven more Indian languages (Bengali, Kannada, Malayalam, Marathi, Tamil, Telugu and Urdu) on 8 October 2025. [5] So the same kind of answer page is in front of your customers, in their languages.
What is missing is an Indian measurement. The research for this book found no measured, India-specific figure for traffic that businesses have lost to AI answers. There is a forecast for news publishers: the Reuters Institute projects Google search traffic to publishers falling 40% over the next three years. [6] That is a projection, and it is about news sites, not shops and factories. For Indian small businesses there are only anecdotes like the two above.
So hold the US numbers loosely. The safe conclusion is modest: a customer can now learn your timings, your range and whether you deliver to their pin code without ever opening your site. Count enquiries, not visits.
That's the first part
Read the rest, free
What the measurements say about AI answers and clicks, why nobody has a proven method for getting cited, the four habits that stay right either way, and the sources. Enter your email once and every guide on this site opens.
- Fewer visitors, more serious ones
- Nobody has a proven method for getting cited
- Four habits that stay right
- Mistakes to avoid
- Do this week
- Sources
Fewer visitors, more serious ones
Two more facts keep this in proportion.
First, AI assistants still send very little traffic. A benchmark report from Conductor, covering 13,770 domains and more than 3.3 billion sessions in the US from May to September 2025, put AI referrals at 1.08% of total website traffic, with ChatGPT accounting for 87.4% of those referrals. [7] Conductor sells search-visibility software, and the article reporting the figure carries no date.
Second, that small stream is growing and seems to be getting more valuable, at least in American retail. Adobe, measuring visits to US retail sites, reported that traffic from generative AI tools rose 693.4% during the 2025 holiday season (1 November to 31 December 2025). [8] In the first quarter of 2026 it rose 393% year on year. In March 2025 AI traffic converted 38% worse than other traffic. In March 2026 it converted 42% better. [9]
A large percentage rise on a tiny base is still a small number. The shift in conversion is the part to notice. A buyer who has asked an assistant three questions and then comes to you is further along than one who clicked the first blue link.
If you sell to other businesses, this applies to you as well. LinkedIn's December 2025 report says the top sources of advice for small business professionals in India are AI tools (48%), networks (45%), search engines (44%) and friends and family (39%). [10] The respondents are LinkedIn members, a far more digital group than the average MSME owner, and the report gives no India sample size. Still, some of your retailers and purchase managers are asking an assistant before they call you.
Search is also only one door. In a Zoho survey of more than 2,700 MSME retailers, published in June 2025, 72% named social media as their primary customer touchpoint, followed by search engines (67%) and marketplaces (60%). [11] Zoho sells software to these retailers, so treat it as a vendor survey. Chapters 12 to 15 cover the other doors.
Nobody has a proven method for getting cited
You will be offered packages that promise to get your business recommended by AI. Before you pay for one, know what the evidence says.
Google's own guidance is blunt: "There are no additional requirements to appear in AI Overviews or AI Mode." [12] It says no special files or markup are needed. What it advises is ordinary: pages its systems can read, helpful original content, important information in text, structured data that matches what is visible on the page, and up-to-date Business Profile and Merchant Center details. (Structured data is the labelling your web developer adds so a machine can tell which line is the address and which is the price.)
Practitioners add one observation. In an Ahrefs study of about 75,000 brands, mentions of the brand across the web had the strongest correlation with appearing in AI Overviews (0.664), well ahead of links from other websites (0.218). [13] The study's author stresses that correlation is not causation. Brands that are talked about may simply be bigger and better known.
Then there is the gap. The research for this book found no independent, controlled test showing that any specific "AI optimisation" tactic causes more citations. Most published advice comes from firms that sell the service.
So the honest position is this. Nobody has proven a method. Anyone who guarantees that ChatGPT or Google's AI will name your business is promising something no published evidence supports. If an agency pitches you, ask three questions:
- What exactly will you change on my website and listings?
- How will we measure enquiries, not rankings or "visibility scores"?
- Which of your claims rests on a controlled test, and may I see it?
The tools will keep moving. In October 2026 Google began rolling out a business hub inside Gemini that connects to Google Business Profile, Google Ads and Workspace. [14] Features like that will come and go. The habits below do not depend on any of them.
Four habits that stay right
None of these is a trick for AI. Each one was good practice before AI answers and remains good practice if the technology changes again.
1. Accurate listings
Write one master record of the business: trading name, address, phone and WhatsApp numbers, opening hours and weekly off, areas served, what you sell, year established. One person owns it. Every public listing is corrected to match it: your Google Business profile, your website, your trade portal pages, your social profiles.
Below is the public information about my business as it appears in different places. It contains business details only, no customer data. MASTER RECORD (this is correct): [trading name, address, phone, WhatsApp number, opening hours, weekly off, areas served, what we sell, year established] LISTING 1, [e.g. Google Business profile]: [paste the text] LISTING 2, [e.g. our website contact page]: [paste the text] LISTING 3, [e.g. our trade portal page]: [paste the text] Task: 1. Make a table with three columns: Detail, Which listing differs, What it currently says. 2. List any detail in the master record that is missing from a listing. 3. Do not correct or guess anything. If a listing is unclear, write "unclear".
What you should get back: a short table of mismatches and gaps.
Check: open each listing and confirm the mismatch with your own eyes before changing anything. Make the corrections by hand.
If you sell products online, treat your product data the same way: one sheet of titles, sizes, materials and prices that every channel is fed from. Shopify, reporting on its second quarter of 2026, said that search built on its product catalogue converted about twice as well as search built on scraped data. [15] That is Shopify describing its own platform, but the logic holds anywhere. Clean data you supply beats data a machine has to guess.
2. Real answers to real questions
Google's guidance says to put important content in text. If your delivery areas, minimum order and warranty exist only inside a brochure image, a catalogue file or your salesperson's head, do not count on a machine finding them.
Take the FAQ sheet you built in Chapter 9 and publish the answers that are safe to make public, as plain text on your website.
[Paste your business note] Here are questions customers really ask us, with our approved answers from our FAQ sheet: [paste the question and answer pairs] Task: rewrite each pair for the FAQ page of our website. - Keep each question in the words a customer would type or say. - Answer in the first sentence, in plain English, then add at most two sentences of detail. - Use only the facts in my answers. Do not add claims, awards, numbers or guarantees. - Write [KEEP PRIVATE?] beside any answer that includes a price, a credit term or anything else that changes often. - If two of my answers contradict each other, list them at the end. Do not choose between them.
What you should get back: a clean set of questions and answers, with flags on anything that may not belong in public.
Check: the owner reads every answer before it is published. A wrong answer on your own website is still your promise.
Do not use AI to produce dozens of generic articles. Google's guidance asks for helpful, original content, and what you know that nobody else does is your own answers.
3. Reviews
Reviews are the part of your public record that customers write, which is why buyers read them. Ask every customer, not only the happy ones.
Write three short messages asking a customer who recently received [product or service] from [business name] to leave an honest review on [Google / our website]. - Version 1: very short, for WhatsApp. - Version 2: slightly warmer, naming the product. - Version 3: more formal, for a business customer. The same request goes to every customer. Ask for an honest review, never a five-star one. Do not offer any reward. Under 50 words each.
What you should get back: three short, neutral requests your staff can send after delivery.
Reply to every review, including the bad ones, using the complaint method in Chapter 9. Never write your own reviews and never ask AI to. India's e-commerce rules bar sellers from posing as consumers or posting deceptive reviews of their own products. [16] Chapter 23 covers this. For marketplace orders, use only the marketplace's own review-request tools.
4. Consistency
Once a month, spend half an hour as your own customer.
| Ask | Example | Write down |
|---|---|---|
| A need | "AC repair near [your area]" | Who is named, and whether you are |
| A comparison | "Best [product] wholesaler in [city]" | What is said about you, if anything |
| Your name | "[Business name] timings and address" | Any fact that is wrong |
Ask Google and two AI assistants. The answers can differ from one day to the next, so treat this as a spot check, not a ranking. When you find a wrong fact, trace it to the stale page or old listing it came from and correct it there.
Example (fictional): A pump manufacturer in Coimbatore runs this check on the first Monday of the month. An AI assistant gives the firm's old godown address and lists a model it stopped making. The owner's son finds both on a dealer directory page nobody has opened in years. He corrects the page from the master record and notes the date. He cannot know when, or whether, the assistant's answer will change, so he checks again the following month. The details are invented for illustration.
The habit also covers who else mentions you. The Ahrefs finding suggests that being named on other sites goes together with being named in Google's AI Overviews. You cannot buy that honestly. You can make sure your trade association directory, your principal's dealer locator and any local press about you carry the same name and details as your master record.
Then measure what matters. Add a "How did you find us?" column to the enquiry log from Chapter 7 and have staff ask every new enquiry. If the answer is "ChatGPT" or "Google showed me", write exactly that. In six months you will have your own numbers, which are worth more than any American study.
From my desk: In 2020 I wrote Unleash Your Business Online!, a book about getting a business online. Six years later, being online is only the starting point. The work in this chapter is making sure that what is online about you is accurate, written in plain text and the same everywhere, because your next customer may read it inside an AI answer and never visit your website.
AI assistants that complete the purchase for the customer are a separate subject. Chapter 27 sets out what is live and what is only announced.
Mistakes to avoid
- Paying for a guarantee. No one can promise that an AI answer will name your business, and the research for this book found no controlled test behind any tactic sold for it.
- Keeping your facts only in images and catalogue files. Put timings, service areas and terms in plain text as well.
- Letting old listings rot. A wrong phone number on a forgotten portal page can be repeated to your customer.
- Writing your own reviews, by hand or with AI. The rules bar it, and it spoils the one record buyers trust.
- Filling your site with generic AI articles and claims you cannot support.
- Judging by website visits. Count enquiries and where they came from.
Do this week
- Ask Google and one AI assistant five questions a customer would ask about your trade and your business. Write down every wrong or missing fact.
- Write your master record and correct your Google Business profile and website contact page to match it.
- Add a "How did you find us?" column to your enquiry log and brief your staff to ask every new enquiry.
Sources
- Outlook Business, newsletter feature on Google, ChatGPT and the AI anxiety of small businesses, 17 July 2026. https://www.outlookbusiness.com/amp/story/newsletter/the-inference/google-chatgpt-and-the-ai-anxiety-of-small-businesses
- Pew Research Center, analysis of Google users' clicks when an AI summary appears (900 US adults, 1 to 31 March 2025), 22 July 2025. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
- Search Engine Journal, report on SparkToro's study of 10 June 2026 using Similarweb data, June 2026. https://www.searchenginejournal.com/google-search-traffic-to-open-web-drops-to-23-data-shows/578724/
- Ahrefs (vendor of SEO software), study of AI Overviews and click-through rates, 4 February 2026. https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
- TechCrunch, report on Search Live in India and new AI Mode languages, 8 October 2025. https://techcrunch.com/2025/10/08/googles-search-live-comes-to-india-ai-mode-gets-more-languages
- Business Today, report on the Reuters Institute's Journalism Trends 2026, 19 January 2026. https://www.businesstoday.in/technology/news/story/google-search-traffic-set-to-drop-40-as-ai-summaries-squeeze-publishers-511426-2026-01-19
- Search Engine Land, report on Conductor's 2026 benchmarks report (article date not shown; Conductor sells search-visibility software). https://searchengineland.com/ai-1-traffic-mostly-chatgpt-464653
- Adobe, holiday shopping season release (Adobe Analytics), 7 January 2026. https://news.adobe.com/news/2026/01/adobe-holiday-shopping-season
- TechCrunch, report on Adobe's first-quarter 2026 data on AI traffic to US retailers, 16 April 2026. https://techcrunch.com/2026/04/16/ai-traffic-to-us-retailers-rose-393-in-q1-and-its-boosting-their-revenue-too/
- LinkedIn Economic Graph, "How Small Businesses Can Win in 2026" (India edition), December 2025. https://economicgraph.linkedin.com/content/dam/me/economicgraph/en-us/PDF/Work-Change-Small-Business-Report-Dec-2025-India.pdf
- Zoho, press release on "The Indian Retailer Survey" (vendor-published; online survey of more than 2,700 MSME retailers), 25 June 2025. https://zstatic.zohostatic.com/news/micro-and-small-indian-retailers-to-invest-in-ai-and-ml-to-stay-competitive-zoho-survey.html
- Google Search Central, guidance on AI features in Search, last updated 10 December 2025. https://developers.google.com/search/docs/appearance/ai-features
- Ahrefs (vendor of SEO software), study of brand mentions and AI Overviews, 26 May 2025. https://ahrefs.com/blog/ai-overview-brand-correlation/
- MediaNews4U, report on Google's Gemini tools for small businesses, 21 October 2026. https://www.medianews4u.com/google-rolls-out-dedicated-gemini-tools-for-small-businesses/
- Retail TouchPoints, report on Shopify's second-quarter 2026 results, 5 August 2026. https://www.retailtouchpoints.com/news/shopify-credits-ai-for-34-revenue-growth-in-q2-2026/620805/
- Citizen consumer and civic Action Group (CAG), blog on India's legal response to fake online reviews, 8 September 2026. https://www.cag.org.in/blogs/regulating-trust-indias-legal-response-fake-online-reviews
Want 25 more prompts like these?
The free AI Starter Kit: 25 prompts for business owners.