By Sara Khan · WeProms Digital · October 24, 2026
Every agency pitch landing in a halal takeaway owner’s inbox this season sells a version of the same promise: that a monthly subscription or a retainer will place your restaurant inside ChatGPT’s answers, and that the customers of the near future arrive that way. The evidence deserves a harder look than the pitch deck invites. Across twelve months of data from 100 websites of varying sizes — 145 million sessions in total — visits that arrived from ChatGPT amounted to 0.62 percent of the volume sent by Google’s organic search results, in an analysis published by Search Engine Journal’s Marie Haynes. Zero point six two. The pattern repeats among the halal takeaways, caterers and grocers whose marketing we review in Slough, Bradford and East London: the phones still ring from Google, Maps and the delivery apps long before an AI assistant sends anyone at all.
The 0.62 percent figure the sales deck never quotes
The dataset behind that number is not small. Marie Haynes sampled 100 properties over a full year and found that adding every other assistant — Gemini, Copilot, Perplexity — barely moved the figure above ChatGPT’s contribution alone, as her Search Engine Journal analysis reports. Usage, meanwhile, is genuinely large and genuinely growing: 54 percent of UK adults now use AI tools such as ChatGPT, Copilot or Gemini, according to Ofcom’s Adults’ Media Use and Attitudes 2026 research, as reported by IBC. Britain has adopted the chatbots; the chatbots have not adopted the habit of sending Britain to small business websites.
Where the habit actually persists, the numbers are blunt. Google carried 91.39 percent of UK search traffic in September 2026, according to Statcounter’s UK market-share data. A Slough family deciding where to order karahi on a Friday night searches Google, asks their Just Eat or Deliveroo app, or messages the family WhatsApp group; almost none of that journey passes through a chatbot. For a business whose catchment is three miles of the Slough trading estate and the streets behind the high street, that is the whole market.
One caveat belongs here, because national datasets get quoted as if they settle local questions. Haynes’ 100 sites are websites of varying sizes with no industry breakdown — not takeaways; a browsing study says nothing about whether your specific customers ask ChatGPT for halal food near Slough. Measure locally before believing either the pitch or the rebuttal.
Serious buyers hide inside small numbers
The honest version of the pro-AI case rests on quality, not quantity, and it deserves to be stated fairly. In the same dataset, Google organic visitors converted at 3.02 percent while ChatGPT visitors converted at 13.8 percent — more than four times the conversion rate, the share of visitors who complete an action you care about, such as placing an order. Volume swamps intent, though: Google organic still delivered 27 times more total conversions than every assistant combined, because 160 visitors at 3 percent beats one visitor at 13. The mosque noticeboard sends fewer readers than the high street, but nearly everyone who reads it arrives with their coat already on and their order in mind.
“In August of 2025, in conjunction with the release of GPT-5, the exponential growth of ChatGPT sessions ended.” — Marie Haynes, Search Engine Journal
The trend line matters as much as the snapshot. ChatGPT has been sending fewer and fewer visitors to websites since that August 2025 turning point, according to the same analysis, which cuts against the idea that patient subscribers are riding an incoming wave. The buyers who do arrive from an assistant are excellent buyers; they researched, they compared, and they arrive late in their decision with high intent. Treat them as a bonus layer on a plan that works without them, never as the base of it.
There is a measurement caveat that cuts the other way, and fairness demands it. A takeaway’s version of conversion is a phone call, a WhatsApp message or a walk-in, not a website visit, so any comparison built on site analytics undercounts discovery that began in a chat and ended on the phone. The family that asked an assistant for the best halal takeaway near Slough, then rang the number from Google Maps, shows up in nobody’s dataset as an AI win. Nobody can measure that journey from the outside, which is precisely why the fifty-bookings count in the closing section beats any vendor’s dashboard.
The subscription that measures a door few walk through
The tools being sold to small businesses deserve their own arithmetic. HubSpot’s standalone AEO tracker costs 50 dollars a month — about 40 pounds, or 600 dollars a year — while Ahrefs bundles AI visibility into its SEO suites and Profound steers buyers toward enterprise contracts, as HubSpot’s comparison lays out. What the tools mostly count is mentions, occasions where an answer names your business, and citations, occasions where it links to your website as a source — a distinction HubSpot’s tool guide draws clearly. A mention without a citation is a rumor with good manners; it changes nothing until someone acts on it.
The research most often cited in these pitches, a Pew Research Center study of US adults’ browsing, found people clicked a traditional result link in 8 percent of visits showing a Google AI summary, against 15 percent without one — the arrival of the zero-click search, a search that ends on the results page without a visit to any website. For a national brand, fewer clicks make being named inside the answer matter enormously. For a takeaway whose question is rarely asked of an assistant in the first place, the same study is evidence for a different conclusion than the one in the pitch.
There is also an uncomfortable parallel in who buys the tools. In Haynes’ sample, 45 of the 100 websites tracked no key events — analytics’ term for the actions you have decided count as progress, such as a booking or a call. Most takeaway owners stand in the same place: unable to say what share of last week’s orders came from Maps, from Just Eat, from the laminated menu by the till. The subscription tells you what ChatGPT says about you; it will not put a chicken karahi on table six.
What a Slough takeaway should fix before paying anyone
The boring door carries the trade, so the boring fixes come first. A booking sheet with one added column — how did you hear about us — produces better data in a month than any dashboard, and it costs a pen. A Google Business Profile with correct hours, a current menu, halal details and answered reviews does more for a Friday-night search than a year of mention tracking, a subject we examined directly in whether Google rankings still matter for a Leeds mithai shop. And the AI layer is not uniformly irrelevant for every UK Muslim business: when Manchester searches for the best abaya shops, Google’s AI now answers first, and fashion retail behaves differently from a three-mile takeaway.
The delivery apps belong in the same audit. Just Eat and Deliveroo take their commission, but they carry the Friday-night intent that chatbots currently send nowhere, and their review scores do more persuading than any mention inside an AI answer. Replying to every review — the two-line thank-yous and the calm answers to the angry ones — is unpaid work that compounds, which is the exact quality the subscriptions claim to sell.
A distinction also protects your wallet. Paying for visibility inside ChatGPT’s answers is a different purchase from buying ChatGPT’s own advertising, whose first-test economics we have already broken down for halal food brands; salespeople deliberately blur the two. Here is the falsifiable claim this article stands behind: between now and Eid al-Fitr next spring, a Slough halal takeaway will collect more incremental bookings from a corrected Google listing and a booking sheet that asks one question than from a year of any AI visibility subscription priced in pounds today. Count your own bookings and test it.
The principle worth keeping
Spend where the customers already arrive. Keep your hours, menu and halal details accurate enough that any engine — search or assistant — can quote them without error. Treat the AI layer as compounding interest on a shop that already converts, not as a replacement for the door that feeds it. The decision rule fits in one line: count the source of your next fifty bookings, and if five or more mention an assistant, an answer-engine budget is justified; if none do, the same 40 pounds a month is better spent on photography, listing accuracy and faster replies. If you do monitor assistants, watch the unbranded questions — best halal takeaway near Slough — rather than your own name, because a brand that answers the question it was never asked is invisible either way.
Read next: Do Google Rankings Still Matter for My Mithai Shop in Leeds? and Are ChatGPT Ads Worth It for My Halal Food Brand? What a First Test Costs.
WeProms Digital approaches this question the way this article does — with arithmetic before enthusiasm. As an agency that has worked with halal restaurants, caterers and grocers across the UK and the Gulf, we run answer engine optimization for family businesses for owners who have decided the AI layer is worth doing properly, and our SEO team for UK takeaways and halal retailers handles the Google and Maps work that still carries most of the trade. Tell us your booking count and we will tell you honestly which side of the fifty-bookings rule you sit on: hello@weproms.com, WhatsApp +92 300 0133399, or weproms.com/contact-us.
Sources & References
- Search Engine Journal — GEO Is A Distraction: Here’s Where The Real Opportunity Is — October 2026
- IBC — AI use rivals social media activity in UK adults, Ofcom finds — 2026
- Statcounter — Search Engine Market Share United Kingdom 2026 — September 2026
- Pew Research Center — Google users are less likely to click on links when an AI summary appears in the results — July 2025
- HubSpot Marketing Blog — Ahrefs Brand Radar vs. Profound: Which fits your team? — October 2026
- HubSpot Marketing Blog — 8 LLM visibility tools to track your brand across answer engines — October 2026
- HubSpot Marketing Blog — Which AI prompts should you track? A practical guide — October 2026
Additional reading from industry feeds:
- Search Engine Journal — GEO Is A Distraction: Here’s Where The Real Opportunity Is
- HubSpot Marketing Blog — Ahrefs Brand Radar vs. Profound: Which fits your team?
- HubSpot Marketing Blog — 8 LLM visibility tools to track your brand across answer engines
- HubSpot Marketing Blog — Which AI prompts should you track? A practical guide
