In short: people increasingly meet your brand inside an AI answer before they ever reach your website. Four changes from the last few weeks decide whether that answer is right, and whether it mentions you at all. None of them needs a big budget. All of them need someone to actually do the work.
1. Google: AI-written content must be checked by a person
Google's updated guidance (1 October 2026) says AI-assisted content has to be fact-checked by a human before it is published, and that this covers more than the article: titles, meta descriptions, structured data and image alt text too. Low-effort, mass-produced AI pages are the target.
What to do: add a human review step for every page and post. One person checks the facts and claims; one checks the title, description, schema and alt text. If you publish at volume, sample-check every batch.
2. AI summaries now appear on most brand searches
Data reported by DemandSphere in late September showed AI Overviews on around 80% of branded queries, up from about 26%. When someone searches your brand name, Google increasingly writes a summary about you before showing your site.
What to do: make your company facts identical everywhere AI reads them: website, schema, Google Business Profile, LinkedIn and directories. Same legal name, same phone, same services, same people. We found a real example of the cost: an AI answer mixing up two agencies with similar names and showing the wrong phone number.
3. Cloudflare changed how AI bots reach your site
Since mid-September, Cloudflare sorts AI bots by purpose (search, agent, training) and lets site owners set a policy for each. Blocking the wrong category can also block multi-purpose crawlers such as Googlebot and Bingbot. On 30 September it added opt-in pay-per-use controls for AI agents.
What to do: if your site runs on Cloudflare, open AI Crawl Control and check what is allowed. For marketing pages you want found, allow search and assistant bots and keep pages free to read. Pay-per-use suits publishers, not brands that want to be recommended.
4. "AEO" and "GEO" are still SEO, plus being mentioned elsewhere
Google's own AI search guidance treats answer- and generative-engine optimisation as part of SEO, not a separate trick: fast pages, clear answers and correct entity data. Off your site, third-party mentions matter most. In one 30-day test published in September, third-party "best of" lists produced about 86% of source mentions in AI answers, against 14% for the brand's own list. ChatGPT accounts for roughly three-quarters of AI referral traffic, and it leans on Bing's index, so Bing Webmaster Tools is worth the five minutes.
How Pulse helps
- Websites: fast, structured pages with schema, clean titles and a human review step before anything goes live.
- AI marketing dashboard: see AI-assistant traffic (ChatGPT, Gemini, Perplexity, Copilot) as its own channel next to search and paid, with plain-language explanations of what changed.
- Paid media: when AI answers take clicks from organic results, paid search and social carry more of the load. We find the leak, fix it, prove it and go again. Nothing goes live without your yes.
- Entity and AI-visibility setup: schema, Google Business Profile, Bing Webmaster, Cloudflare crawler settings and a monthly check of what ChatGPT, Gemini, Perplexity and Google AI Mode say about your brand.
- AR filters for Snapchat and TikTok: content people share is content AI tools see. Ahmed Sewilam designs and builds branded lenses and effects in-house.
We work with brands in the UAE, across the Gulf and worldwide, remotely and in English and Arabic.
Quick checklist
- Human review before publishing: facts, title, description, schema, alt text.
- Company facts identical on site, schema, Google Business Profile and LinkedIn.
- Cloudflare AI Crawl Control: search and assistant bots allowed.
- Bing Webmaster Tools set up and sitemap submitted.
- AI-assistant traffic tracked as its own channel in GA4.
- Monthly: ask the main AI assistants about your brand and fix what they get wrong.