Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Drowning in prompts? How to build an AEO progra...

Drowning in prompts? How to build an AEO program around the data that matters

AI search teams now have access to more data than ever: prompts, citations, competitors, mentions, sources, visibility scores, and more. But more data does not automatically mean better decisions.

This session will show how to build a practical AEO program around the data that actually matters, so your team knows what to track, what to ignore, and what to work on next.

Attendees will learn how to:
- Choose the right prompts and customer journeys to track
- Find where competitors are winning and why
- Separate useful signals from vanity metrics
- Prioritize the opportunities most likely to improve visibility
- Turn insights into an execution plan the team can actually run
- Use automation where it helps, without replacing human judgment

The goal is simple: less noise, better priorities, and a clearer path from AEO data to action.

Avatar for Samanyou Garg

Samanyou Garg

October 08, 2026

More Decks by Samanyou Garg

Other Decks in Marketing & SEO

Transcript

  1. Meet Pacewell Road and trail shoes UK market Road model:

    £120 Fictional brand · illustrative programme
  2. Five steps. Repeat weekly. Prioritize Track ↑ Act Your team

    chooses the work Repeat weekly Verify Measure 4
  3. T HE AE O P ROG R AM ME 01

    Track Organise the prompts so the data tells you where to work 5
  4. T R AC K The dashboard is the starting point

    Which part of the business needs work? Writesonic · example dashboard 6
  5. T R AC K Topics organise your prompt coverage 7

    ROA D RUNNI NG TRA IL R UNNING Casual runners under £150? Trail shoes under £150? Short everyday runs? Grip on wet paths? Wide-fit options? Occasional trail running?
  6. T R AC K A 50% average hides a 20%

    problem Trail running 80 Road running 20 0 20 Overall visibility 40 60 Illustrative: road running 6/30 answers; trail running 24/30. Equal-sized samples. 8 50% 80 100
  7. T R AC K Cover the buying journey 9 QUESTION

    TYPE TOFU MOFU BOFU Category How do I choose running shoes? Good road-running shoes under £150? Which shortlisted shoe best suits casual road runs? Competitor / comparison Which brands make running shoes? Alternatives to Nike roadrunning shoes? Nike Pegasus or Adidas Supernova for casual runs? Branded What kind of shoes does Pacewell make? Which Pacewell shoes suit road running? Is Pacewell Road worth buying? What are the returns terms?
  8. T R AC K Where customer questions come from SEARCH

    CONSOLE KEYWORD PLANNER / SEO SALES CONVERSATIONS Queries that already bring people to you Relevant categories and buying terms Questions and objections before purchase SUPPORT CHANNELS REDDIT / INDUSTRY FORUMS CUSTOMER REVIEWS Fit, price, product and service questions Comparisons and problems in customer language What people value, dislike or misunderstand 10
  9. T R AC K Turn the search query into a

    buyer question road-running shoes under £150 ↓ Real prompt, inferred from a search need · ChatGPT, 7 October 2026 11
  10. T R AC K For a sales goal, prioritise MOFU

    and BOFU MOFU · SHORTLIST “Best running shoes under £150?” Get included among relevant options. BOFU · CHOOSE “Pacewell or Nike for casual road runs?” Help the buyer compare and choose. TOFU · LEARN LATER “How do I choose running shoes?” Expand into broader educational questions. 12
  11. T R AC K New wording or a new buying

    need? SAME BUYING NEED “What are the best road-running shoes under £150?” “What are the top road-running shoes under £150?” A DIFFERENT NEED “What are good road-running shoes under £150 for wide feet?” 13
  12. T R AC K Non-branded: shortlist. Branded: evaluate. NON-BRANDED ·

    GET SHORTLISTED BRANDED · HELP THE BUYER EVALUATE 14
  13. T R AC K A stable core, plus seasonal questions

    EVER GREEN SEASONA L / LA UNCH Road-running shoes under £150 Wet winter roads New product model Keep for ongoing comparison Track separately, with dates 15
  14. T R AC K Start with a manageable benchmark 40–50

    30 Daily Prompts per priority Days with a stable core Collection; weekly topic set decisions Expand by business need, journey stage and market. 16
  15. T HE AE O P ROG R AM ME 02

    Prioritize Turn a topic gap into a specific task 17
  16. PR I O R I T I Z E Road

    running: find the competitor gaps PROMPT COMPETITOR APPEARS PACEWELL APPEARS Road shoes for casual runners under £150? Yes, repeatedly Rarely Road shoes for short everyday runs? Yes, repeatedly Rarely Topic → missing prompts → cited pages Pacewell worked example 18
  17. PR I O R I T I Z E Group

    the exact pages across those answers EXAMPLE PAGE CITED IN WHAT IT SAYS RunGuide: Best road shoes under £150 runguide.example/road-shoes-under-150 7 / 10 answers Lists FleetRun; Pacewell absent FleetRun Start product page fleetrun.example/start 6 / 10 answers Explains casual road use and fit RunTalk: Shoes for casual road runs runtalk.example/casual-road-shoes 3 / 10 answers Discusses FleetRun and sizing Illustrative pages and counts · .example addresses are fictional 19
  18. PR I O R I T I Z E ·

    T H I R D - PA R T Y O U TR E A CH A missing listing becomes an outreach task RUNGUIDE THE NEXT ACTION Road shoes under £150 Check Pacewell fits the criteria. FleetRun included Ask the editor to evaluate it. Pacewell absent Pacewell worked example · Inclusion is not guaranteed. 20
  19. PR I O R I T I Z E ·

    O W N PR O D U C T PA GE Which buying questions are unanswered? VI SIBLE MIS SING Product image Suitable runner? Colours and £120 price Road or trail use? Ordering information Fit guidance? Returns terms? Pacewell worked example 21
  20. PR I O R I T I Z E ·

    F AC T C H EC K Factuality: find the wrong information £190 → £120 Find the outdated page or feed. Request the correction, then verify it is live. Illustrative price correction · Owned, affiliate and third-party pages all matter. 22
  21. PR I O R I T I Z E ·

    N E GATI V E S E N T I M E N T Negative sentiment: find the reason “Pacewell returns are difficult.” OLD INFORMATION Correct the policy or description REAL CUSTOMER PROBLEM Fix the service behind the complaint Illustrative brand-perception investigation 23
  22. PR I O R I T I Z E ·

    N E GATI V E S E N T I M E N T A branded question can surface concerns Real ChatGPT answer excerpt · 7 October 2026 · Model output, not verified product advice. 24
  23. PR I O R I T I Z E Different

    platforms cite different sources 3.8% of cited domains shared by all four platforms ChatGPT · Gemini · Perplexity · Google AI Overviews Writesonic study · 70,879 four-platform citation comparisons 25
  24. PR I O R I T I Z E ICE:

    choose the work to do first I · IMPACT How much could it help customers or sales? C · CONFIDENCE What makes this a justified task? E · EASE How quickly can your team deliver it? ICE: Impact, Confidence, Ease · Sean Ellis 26
  25. PR I O R I T I Z E A

    short worklist, ranked with reasons TASK I C E MEAN Correct £190 on the affiliate page 9 9 8 8.7 Add missing answers to Pacewell’s page 8 8 9 8.3 Ask RunGuide to evaluate the product 8 6 4 6.0 Illustrative ratings · Higher is better · Mean = (I + C + E) ÷ 3 27
  26. T HE AE O P ROG R AM ME 03

    Act What exactly will we change, and who will do it? 28
  27. ACT One page update, with a clear owner PACEWELL ROAD

    PAGE ALREADY THERE Product image, £120 price and ordering information ADD Casual road runners · paved roads Standard width + size guide · 30-day unworn returns OWNER / DUE Content lead · Friday Product team confirms the details Illustrative page update 29
  28. ACT One owner per task Content and product facts Content

    / product marketing Third-party inclusion PR / partnerships Access problems Web / SEO Results and reporting One person checking the numbers and updating the team One or two people can cover several roles in a small team. 30
  29. ACT Agents prepare. People approve. Agents Find competitor gaps. Draft

    page updates. Humans Choose priorities. Check facts. Approve what gets published. 31
  30. T HE AE O P ROG R AM ME 04

    Verify Check that the agreed correction is live and correct 32
  31. V ER I F Y Verify the change on the

    published page 33 YOUR PAGE THIR D-PA RTY PAG E Missing questions answered? Correction actually live? Product facts correct? Correct product and link? Content readable? Publicly visible?
  32. V ER I F Y Fetched → cited → brand

    included FETCHED OAI-SearchBot: search-crawler fetch ChatGPT-User: user-triggered fetch CITED Is this exact page cited more often? BRAND INCLUDED Brand mentioned or recommended in answers citing that page Separate observations; no stage guarantees the next · OpenAI bot documentation 34
  33. V ER I F Y Work completed ≠ impact measured

    35 WOR K COMPLETED IMPA CT MEASURED Page updated Page citations New page published Visibility and answer accuracy Correction live Business outcomes
  34. T HE AE O P ROG R AM ME 05

    Measure Baseline changes · Work completed · Business outcomes 36
  35. M EA S UR E Compare the same prompts before

    and after COMPARE Same prompts, market, platform and period length CHECK Brand visibility, page citations and correct facts Keep branded and seasonal prompt sets separate. 37
  36. M EA S UR E Connect an AI visit to

    the order Source → visit → order / lead → revenue Loves Data published GA4 example · Not Pacewell results 38
  37. M EA S UR E Ask about discovery and the

    decision Where did you first hear about Pacewell? What helped you decide to buy? If AI: which tool, and what did you ask? (Optional) Example questions · Save with the order or lead; avoid counting a sale twice. 39
  38. MEASURE AI can help them choose your brand Knows Pacewell

    ↓ Checks reviews in ChatGPT ↓ Buys later through Google
  39. M EA S UR E Three parts to the monthly

    report 1 · BASELINE Road-running visibility: 20% to 30% 2 · WORK Product page updated. Affiliate price corrected. 3 · BUSINESS 40 AI-referral orders · £4,800 revenue 12 AI-assistance reports (may overlap) · £2,000 cost NEXT MONTH Finish road-running gaps, review impact, follow up the editor. Expand to trail running when the team has capacity. Illustrative report. These figures alone do not establish incremental sales or ROI. 41
  40. E XPAN D Crawl. Walk. Run. CRAWL WALK RUN One

    or two topics A reliable weekly More categories and routine markets More buyer needs Checked automation A task through to review 42
  41. CL OSE The AEO programme, on one page STEP WHAT

    YOUR TEAM DOES Track Organise prompts by topic and buying journey Prioritize Find gaps and rank tasks with ICE Act Assign the specific change, owner and due date Verify Check the agreed change is live and correct Measure Report baseline changes, completed work and business outcomes Repeat weekly. Use the results to choose the next task. 43
  42. CL OSE Your next Monday 1 Choose one important topic

    2 Check its prompts and save the baseline 3 Find one repeated competitor gap 4 Assign one task and a review date 44
  43. BOOTH 17 Get the slides Meet the Coconut Man Free

    fresh coconut water. Bring your AEO questions. speakerdeck.com/samanyougarg
  44. T R AC K Brand visibility and page citations BR

    AND V ISI BILI TY PAGE CITATIONS “Good running shoes under £150?” “How do I choose running shoes?” Is your product among the Is your guide cited as a source? options? A page can be cited without the product being recommended. Both can happen together. 46
  45. PR I O R I T I Z E Check

    the cited page before creating a task CHATGPT ANSWER THE PAGE IT CITES Price agrees. No price correction needed. ChatGPT and SportsShoes.com · Captured 7 October 2026 47