An AI search strategy playbook

0:00 The AI search strategy playbook

You've now covered how AI search works, how to write and structure content for it, how to control crawler access, how to measure visibility, and how to separate evidence from hype. In this final lesson of the core course, we'll pull everything into a practical ninety-day playbook you can run for a client or your own business, with five pillars, a month-by-month plan, adaptations by business type, and a way to explain it all to a client in plain language.

0:35 Five pillars

Here's the key idea: AI search optimisation works best as an extension of a strong SEO, content and brand programme, not a separate bag of tricks. There are five pillars. Access: bots and search engines can fetch and index your key content, and your policy is deliberate. Clarity: pages answer questions directly with specific, sourced facts. Entity: facts about your brand, people, products and locations are consistent everywhere. Corroboration: you're present and accurately represented in the sources answers draw on. Measurement: referrals, first-party AI reports, a prompt panel, accuracy tracking and honest reporting.

1:15 Days 1–30: foundations

Days one to thirty: foundations and baseline. Write the AI access policy and fix accidental blocks. Audit access on your top twenty pages: raw HTML checks, Google and Bing indexing, and logs for AI bots and status codes. Create the entity fact sheet and fix the About page, Organization markup and top five profiles. Set up measurement: the GA4 AI channel, baselines from the Search Console AI report and Bing's AI Performance report, and a thirty to fifty prompt panel with three runs each. Record your decision on Google's Search generative AI control. And grab quick wins by rewriting the intros of your ten most important pages answer-first.

2:02 Days 31–60: content + corroboration

Days thirty-one to sixty: content and corroboration. Create or improve pages for your high-value prompt gaps: comparisons, how-to-choose guides, pricing model pages, specific use cases and local pages, with tables, clear headings and sources. Plan one piece of original research or data. Run the AI source gap analysis, start a compliant review programme on the platforms answers cite, brief the authors of key comparison articles, and participate transparently in communities. And correct the misinformation sources you found in the baseline.

2:37 Days 61–90: amplify + review

Days sixty-one to ninety: amplify and review. Launch the research or expert commentary to the publications cited in your category. Add transcripts to videos and podcasts, and descriptive text for charts. Prioritise Arabic or Urdu versions of key pages if your markets need them. Then run the panel a second time and report: mention rate, citation rate, share of voice, accuracy, AI referrals and conversions, AI impressions, Copilot citations and branded search. Finish with a retrospective: which actions correlated with changes, and what will you continue, stop and start?

3:16 Adapt by business type

Adapt the emphasis by business type. Local services: Business Profile accuracy, reviews, citations and local pages. SaaS and B2B: comparison and alternatives content, review platforms, integration pages and documentation, perhaps with llms.txt for docs as an experiment. E-commerce: product data accuracy, feeds, reviews, buying guides and accessible product pages. Publishers: access policy and licensing decisions, original reporting, authorship, and measuring substitution versus referral. Agencies and consultancies: thought leadership, case studies, expert entity signals, and directory and review presence.

3:50 Example 1: Birmingham barbershop

Worked example one, simple. A two-chair barbershop in Birmingham doesn't need a ninety-day programme with five workstreams. Its version fits on an index card. Month one: accurate Business Profile and website hours, prices in HTML, and allow the search bots. Month two: ask happy customers for honest reviews. Month three: check two assistants for best barber near the Jewellery Quarter, and see who's mentioned. The pillars still apply. The scale shrinks to fit.

4:22 Example 2: Dubai fragrance brand (illustrative)

Worked example two, with illustrative details. A Dubai home fragrance brand runs the full playbook. In month one, the log report reveals OpenAI and Perplexity search bots getting four-oh-threes from a CDN rule, which is fixed. Product specs move into server-rendered HTML, JSON-LD is generated from live data, and the Merchant Center feed updates more often during promotions. In month two, fair oud versus bakhoor and how to choose a diffuser guides go live, reviews are invited from verified buyers, and gift-guide editors are briefed. In month three, Search Console shows AI impressions on the new guides and Bing shows the first Copilot citations. The report shows leading indicators and caveats, with no promised rankings.

5:12 The client narrative

How do you explain this to a client? Keep it simple and honest. AI assistants and AI search features increasingly answer questions directly. They mostly draw on content they can find and trust on the web. We'll make sure they can access your content, that it answers questions clearly and accurately, that the facts about your business are consistent everywhere, and that trusted third parties talk about you. We'll measure mentions, citations and traffic, with the caveat that these systems vary and change. No one can guarantee placements. We can make you a strong, credible candidate.

5:54 Watch me do it: a 90-day plan

Watch me do it. I'll fill the one-page ninety-day plan for a real client in about twenty minutes. The client is a Leeds physiotherapy clinic. Row one, access. Days one to thirty: check robots.txt and the CDN, run the fetch test on the five service pages, and confirm Bing indexing. Owner: the web developer. Row two, clarity. Days one to thirty: rewrite the top five service page intros answer-first, with prices and areas covered. Days thirty-one to sixty: add a how to choose a physio guide. Owner: the clinic manager with me editing. Row three, entity. Fact sheet and Organization markup in month one; Business Profile and directory corrections in month two. Row four, corroboration. Source gap analysis in month one; review invitations and a local news pitch in month two; a community talk in month three. Row five, measurement. GA4 AI channel, Search Console AI report, Bing AI Performance, and a fifteen-prompt panel in month one; the quarterly report in month three. Then I sanity-check capacity. It's too much for month two, so I move the guide to month three. I share the page with the owner and book a check-in at day thirty.

7:19 Governance and mistakes

Governance ties it together. Keep a written AI crawler policy and review it every six months. Never use hidden text, cloaking, fake reviews, sock puppets or undisclosed paid placements. Keep comparison claims fair and substantiated. And respect privacy when handling data for original research or log analysis. Common mistakes: running AI search as a disconnected side project, chasing tactics before fixing access and clarity, and overpromising to clients.

7:49 Recap and try this now

Recap. Five pillars, ninety days, adapted to the business, explained honestly. Try this now. Copy the one-page ninety-day template from the lesson text and fill in one task per pillar per month for your own site or a client, with an owner for each row. And if you'll be producing content at scale, your next step is the course Programmatic SEO and AI Content at Scale, which covers templates, data and guardrails that keep scaled content on the right side of the spam policies.