Between the relaunch of vridhiidigital.com in April 2026 and 5 September 2026, this website produced zero clients. Every one of the roughly 30 leads we worked in those five months came warm, through referrals and the founder's own network. Then on 7 September a prospective interior design client called and said they had found us through ChatGPT, and on 8 September a second prospect called and said they had found us through a Google search. Two inbound calls in two days after five months of none. This report sets down the before, the two afters, and the gap between them.

A note on dates. This report was due on Wednesday 9 September, the run that should have written it did not happen, and it was written on 11 September. It carries both dates.

What did the site actually do for five months?

It reassured people who had already been told about us. The relaunch went live in April and the daily blog began on 28 April. In all that time, not one client came in cold through the website, the WhatsApp button or an AI answer. The roughly 30 leads we worked arrived because someone we had worked with mentioned our name, or because a founder we already knew called back.

That does not mean the site was idle. A referred prospect looks you up before they reply, and what they find decides whether the referral survives the first meeting. We used this figure in the CA firm playbook to argue that a professional services website's first job is proof for referred prospects. The site did that job. It did not do the other one, and we said so in the ledger every week rather than round it up.

What changed on 7 and 8 September?

Monday 7 September. A prospective interior design client called. Asked how they found us, they said ChatGPT. It is the first enquiry a prospect has ever attributed to an AI assistant, and the first inbound enquiry of any kind since April.

Tuesday 8 September. A second prospect called and said they had found us through a Google search. The second inbound enquiry in two days, and the first attributed to Google.

Here is what we do not know. For neither caller do we yet have the query they typed, whether the engine showed a link or only our name, which page the answer drew on, or whether the call came through the WhatsApp button, the contact form or a direct dial. Both are logged as a fact and a date, and neither caller is identified.

Did the 2 September citation cause the 7 September call?

We cannot say. On 2 September our Hyderabad small business agency shortlist surfaced in an AI answer for its target query, and it has returned on every check since, on 4, 7 and 11 September. We reported it in three comparison pages, 23 days, one citation and wrote that no enquiry could be attributed to it. Five days later a ChatGPT attributed call arrived. The timing is suggestive, the content is not: the cited page is a list of Hyderabad marketing agencies, and the caller is an interior design prospect, whom ChatGPT may just as easily have sent from our July interior designer marketing post, our Sainikpuri pages or the homepage. Until the caller's query is known, the two facts sit beside each other and we decline to draw the arrow.

There is a second reason for caution, and it is the most useful thing here for anyone tracking AI search. A Similarweb study of real user journeys found that people who received an AI recommendation were 2.5 times more likely to visit that brand's website within seven days, and that over 56 percent of them arrived by typing the brand name into a search engine, not by clicking a link in the answer (source: AuthorityTech summary of Similarweb research, 6 August 2026). So the caller who says Google may have met us first in an AI answer, and the caller who says ChatGPT may have reached the site as a branded search our analytics filed under Google. The caller's own account is the best attribution available, and it is imperfect both ways.

What does two calls prove?

Less than a chart would suggest, and we are not drawing one. Two enquiries are not two clients, and neither has signed as this is written. Five months of zero followed by two calls in two days is a change in the pattern. It is not a trend, because a trend needs a third point, and it is not a business result, because that is a signed engagement with a confirmed route in.

What it does do is move the ledger. Our tracker has counted visibility since 12 June: after the 11 September check the tally stands at 9 Vridhii returns across 92 non brand query checks. Until this week every one of those returns was a visibility result with nothing on the enquiry side to set against it. Now there is a second column, with two entries, and the two can start to be read together.

What are we doing about it?

Three things. First, confirming the details with both callers on the next conversation: the query, the surface, the likely page, the route in. Those answers go into the ledger and the next report, whichever way they point. Second, counting enquiries by source every month from now on, so the next before and after is measured rather than remembered. Third, leaving the cited page alone. It has not been edited since 10 August and has returned four times; changing it now would make the next reading meaningless.

Key takeaway: Zero inbound clients from the April 2026 relaunch to 5 September, against about 30 warm referred leads. Then a ChatGPT attributed call on 7 September and a Google attributed call on 8 September. Two calls are not two clients, neither route in is confirmed, and the 2 September citation cannot yet be linked to either. Publish the before and the after, and keep the gap visible until it closes.

Counting enquiries by source is one chapter of the Marketing Operating System we run for every client, where numbers like these get reviewed weekly. The search side of that work, from the map pack to the AI answer layer, is our SEO and AI search service, and the method behind the citation tally is written up in how to check whether AI search cites your business. If you would rather have a team handle this, Vridhii Digital builds growth systems for founders who care about results. Message us on WhatsApp and tell us what you run.

Frequently asked questions

You got cited by AI search on 2 September and a ChatGPT lead on 7 September. Is that cause and effect?

We cannot say so. The cited page was a Hyderabad agency shortlist and the caller was an interior design prospect, so the engine may have drawn on a different page entirely. Until the query and the page are confirmed, the two facts sit side by side, not in a chain.

Five months with zero inbound clients sounds bad. Why publish it?

Because a falling or flat number with an explanation is worth more than a rising one without. The site did a different job in that period, reassuring about 30 referred prospects, and saying so plainly is the only way the two calls can be read honestly.

How do I know whether an enquiry came from AI search?

Ask on the first call, then check the analytics. Most AI influenced visitors arrive by typing the brand name into Google days later, so the referral report will not show it. The caller's own account is the only attribution you get for now.

Two calls is a small number. What would make it a real result?

A signed client whose route in is confirmed, then a second one. Until then we count enquiries by source monthly and publish the count. Two calls in two days after five months of none is a change in the pattern, not yet a business result.