TL;DR
For one sample business, published studies put the value of better AI visibility anywhere from ₹36,000 to ₹11.3 lakh a month.
The biggest swing is how many buyer questions get an AI answer at all. The weakest evidence is for clicks.
Almost all the data is from the US, and most of it comes from companies that sell search or analytics tools.
The value of being named in AI answers can be worked out as a chain: searches, AI answers, answers that name you, visits, customers, rupees. We reviewed every 2025 and 2026 study we could find on the four rates in that chain. For one sample business, the published figures support anything from ₹36,000 to ₹11,26,400 a month of extra value. That is a 31-fold range.
The biggest swing comes from one question: how many buyer questions get an AI answer at all. The weakest evidence is for clicks, where no study measures exactly what a brand needs to know. Almost all the data is from the United States, and most of it comes from companies that sell search, analytics or commerce tools.
Key findings
- AI answers now appear on most buying searches. Pew Research Center found an AI summary on 18% of Google searches in March 2025. Similarweb put it at 43% in 2026. Semrush found AI Overviews on 88% of 400 buying-type keywords in August 2026.
- Few people click from an AI answer. The published figures sit between 1% and 2%. Brands named in Google's AI Overviews get about twice the clicks of brands that are not.
- Many look you up later instead. One study, from Similarweb, links being recommended by ChatGPT with 2.5 times more visits within a week, mostly through searches for the brand name.
- AI visitors seem to buy more often. The largest recent data sets, from Adobe and Shopify, find they convert 60% to 80% more often than other visitors. The only academic study, on older data, found the opposite.
- No published study measures any of these steps in India.
The chain from a question to a sale
A brand cannot see how often an AI answer names it in the way it sees Google rankings. No AI company shares that number. So the value has to be built up step by step, from numbers a brand has and rates that others have measured.
| Step | Where the number comes from |
|---|---|
| 1. People search your category's questions | Your own keyword data |
| 2. Some of those searches get an AI answer | Published studies, reviewed below |
| 3. Some of those answers name you | Your own tracking of AI answers |
| 4. Some people click through to you | Published studies, reviewed below |
| 5. Others search for your name later | Published studies, reviewed below |
| 6. Some of these people become customers | Your own conversion rate, adjusted by published studies |
Steps 1 and 3 and the value of a customer belong to each business. Steps 2, 4, 5 and 6 are rates that apply across businesses, and those are what this paper reviews.
Step 2: How many searches get an AI answer
| Source | Who they are | Finding | Period and scope |
|---|---|---|---|
| Pew Research Center | Independent research centre | An AI summary on 18% of Google searches | March 2025. Real browsing of 900 US adults, 68,879 searches |
| Similarweb, reported by TechCrunch | Sells web traffic data | AI Overviews on 43% of Google searches, up from 15% a year earlier | 2026. Panel data. Country and device not stated |
| Semrush | Sells search marketing tools | AI Overviews on 88% of 400 tracked buying-type keywords, and 97% when weighted by search volume | August 2026. US desktop. Banking, software and online shopping |
The share has risen fast, and it depends on the kind of question. Questions like "best X for Y", "alternatives to X" and "X pricing" get AI answers far more often than searches in general. Those are the questions buyers ask, so the higher figures are closer to what matters for a brand. The 88% figure, though, comes from one vendor and 400 chosen keywords.
None of these counts questions people ask ChatGPT, Gemini or Perplexity directly, without a Google search. So the true share of buyer questions that end in an AI answer is higher than any of them.
Step 4: How many people click when you are named
| Source | Who they are | Finding | Period and scope |
|---|---|---|---|
| Pew Research Center | Independent research centre | People clicked a link inside the AI summary on 1% of visits. They clicked a normal result on 8% of visits with a summary, against 15% without one | March 2025. 900 US adults |
| Seer Interactive | Marketing agency | Brands named in an AI Overview had a click-through rate of 2.07%, against 0.94% for brands not named and 3.35% when no AI Overview appeared | 2025 average. 53 brands, 5.5 million queries, information-seeking searches |
| Ahrefs | Sells search marketing tools | The top-ranked page got 58% fewer clicks when an AI Overview appeared, up from 34.5% in its earlier study | December 2023 against December 2025. 300,000 keywords |
No study measures exactly what a brand needs here: of the people who see an AI answer that names the brand, how many go on to visit it. The nearest figures sit between 1% and 2%. Two points hold across the studies. AI answers cut clicks to websites overall. Within that smaller pool, being named roughly doubles a brand's share of the clicks.
Step 5: How many people search for you later
| Source | Who they are | Finding | Period and scope |
|---|---|---|---|
| Similarweb, reported by Search Engine Journal | Sells web traffic data | Brands recommended by ChatGPT were 2.5 times more likely to get a visit within 7 days than brands not recommended. 55.9% of that traffic came through searches for the brand name | June 2026. US desktop users. Finance, travel and beauty |
This is a single study from a vendor, and it shows a link, not a cause. It also does not give the number a calculation needs: how many later searches follow each click. Any figure for this step is a translation of the study, not a measurement. That is why our low case below leaves it out.
The pattern is still worth taking seriously. A person who reads an AI answer and searches the brand name a day later shows up in analytics as brand search or a direct visit, not as AI. This is the main reason AI looks small on dashboards.
Step 6: Do AI visitors buy more often
| Source | Who they are | Finding | Period and scope |
|---|---|---|---|
| Adobe | Sells analytics software | AI visits to retail sites converted 60% more often than other visits, for the 11th month in a row | 2026. More than 1 trillion visits to US retail sites |
| Shopify, reported by CMSWire | Runs an online store platform | AI visitors who reached a product page converted about 80% better than organic search visitors, and about twice as well in categories where shoppers compare products | April to June 2026. Shopify stores |
| Similarweb, reported by Digiday | Sells web traffic data | Visits from ChatGPT converted at 11.4%, against 5.3% for organic search | September 2025 report |
| Kaiser and Schulze, reported by Digiday | University of Hamburg and Frankfurt School of Finance and Management. Working paper, not yet peer reviewed | Visits from ChatGPT converted about 13% less often than organic search | 12 months of data to mid-2025. 973 online stores |
| Ahrefs | Sells search marketing tools | AI search brought 0.5% of its visitors and 12.1% of its sign-ups, which it puts at 23 times the rate of organic search | 30 days, June 2025. Its own website only |
The largest recent data sets agree that AI visitors buy more often, by about 1.5 to 2 times. The "23 times" figure often quoted comes from one company's own website. The only academic study found the opposite, though on data that ends before the more recent studies begin.
There is also a simple explanation that has nothing to do with AI. People who click through from an AI answer have often done their comparison already. They may have been ready to buy anyway.
Three cases for one sample business
To see what the evidence adds up to, we ran the chain for a sample business: 40,000 monthly searches for its category's questions, AI visibility rising from 5 to 30 out of 100, 2% of visitors becoming customers, and ₹1,00,000 per customer. We changed only the four published rates.
| Low | Middle | High | |
|---|---|---|---|
| Searches that get an AI answer | 18% (Pew) | 43% (Similarweb) | 88% (Semrush, buying keywords) |
| People who click when you are named | 1% (Pew) | 2% (Seer Interactive) | 2% (Seer Interactive) |
| Later searches for each click | Left out | 1 | 1 |
| Visitors who become customers | 2% | 2% | 3.2% (2% plus Adobe's 60%) |
| New customers a month | 0.4 | 3.4 | 11.3 |
| Extra value a month | ₹36,000 | ₹3,44,000 | ₹11,26,400 |
The high case is 31 times the low case. Of that gap, the share of searches with an AI answer accounts for the most, about 5 times on its own. Counting later searches doubles the result, and so does moving the click rate from 1% to 2%. The conversion lift adds the last 1.6 times.
Where our calculator sits
Branly's ROI calculator starts at 35% for searches with an AI answer, 5% for clicks and 1 later search for each click, with your normal conversion rate. For the sample business that gives ₹7,00,000 a month, between the middle and high cases.
It gets there by a different route from the studies. Its 35% is below the figures for buying questions, and its 5% click rate is above every published figure. The two partly cancel out. A reader who wants to stay strictly within published figures can set the AI answer share to 43% and the click rate to 2%, which gives ₹3,44,000.
What the evidence does not tell us
- Anything about India. We found no published study that measures any of these rates in India. The one Indian figure we found is that AI chatbots sent about 0.09% of website referrals in India in August 2026, the same as a year earlier. That is Statcounter data, as reported by upGrowth, a growth agency that sells AI visibility audits.
- Mobile, apps and business software. Most studies cover US desktop browsing and online retail.
- Being named without a link. AI answers often name a brand without linking to it. No study counts how often that happens or what it leads to.
- Cause and effect. The studies on later searches and on conversion show links between AI exposure and behaviour. None shows that the AI answer caused the visit or the sale.
- Independence. Of the nine organisations behind these figures, one is an independent research centre and one is an academic team. The other seven sell search, analytics, marketing or commerce services. Branly sells AI visibility tracking, so we have an interest in this question too.
What a brand can do with this
Use the published rates as a starting point, then replace them with your own as soon as you can. Three numbers are within any brand's reach:
- How many of your buyer questions get an AI answer, and how often those answers name you. Check this every week, because answers change.
- Visits from ChatGPT, Gemini and Perplexity in your analytics, and how often those visitors buy compared with others.
- Searches for your brand name over time, set against how often AI names you.
Start the record early. Nobody can look up later what an AI said about your brand last month. Our article What AI search sends you, and what it hides sets out three ways to estimate the effect once you have that record.
How we did this
We looked for studies published in 2025 and 2026 that measure any of the four rates: searches with an AI answer, clicks when a brand is named, later brand searches, and conversion of AI visitors. For each one we recorded who published it, what they sell, the period, the country and the sample. We used the figures as published and did not adjust them. Where a study was reported by a news site and we could not read the original, we cite the report and say so. The three cases use the same sample business as our calculator and change only the four published rates. Research was done in October 2026.
Sources
- Pew Research Center, July 2025: Google users are less likely to click on links when an AI summary appears in the results
- TechCrunch, July 2026: Google's AI search is rapidly becoming the default (Similarweb data)
- PPC Land, October 2026: Google AI Overviews now hit 88% of tracked commercial keywords (Semrush data)
- Seer Interactive, April 2026: AIO impact on Google CTR, 2026 update
- Shopifreaks, May 2026: AI Overviews reduce top-ranking pages' click-through rates by 58% (Ahrefs data)
- Search Engine Journal, June 2026: AI-recommended brands saw 2.5x more site visits (Similarweb)
- Digital Commerce 360, August 2026: Adobe data on AI referral traffic to US retail sites
- CMSWire, August 2026: Shopify data on AI search and organic conversion
- Digiday, October 2025: E-commerce sites see low sales from ChatGPT traffic (Kaiser and Schulze study, and Similarweb data)
- Ahrefs, June 2025: Does AI search traffic convert better than traditional search?
- upGrowth: AI traffic share report 2026 (Statcounter data for India)
Date
October 28, 2026
