GEO Pilot
GEO done for you in 30 days. $999, fixed.
You get your website measured, repaired and rewritten for AI search by us, not handed a report to fix yourself. Works on WordPress, WooCommerce, Shopify and custom code.
One time price. No retainer. We reply within one working day.
Baseline and audit
Technical fixes
Ten pages rewritten
Retest and report
It is built for businesses that want the work done instead of another dashboard. It also works as a safe first step before any larger commitment: you see what we do, what moves and what does not, then you decide.
We do not guarantee rankings, citations or traffic. AI answers change by user, place and day. What we guarantee is the scope below, delivered in the time stated.
What is included
Week by week, exactly what we do
The scope is fixed and written down before you pay.
Baseline and audit
We build a prompt set with you, up to 20 real buyer questions, and run it through five AI engines. We save each answer with the date, note which brands and sources appear and where you are missing.
In the same week we check which AI and search crawlers can reach your site, review your structured data and compare your entity facts across your website and public profiles.
- Prompt set approved by you
- Five engine baseline, saved with dates
- Crawler access and rendering check
- Technical and entity audit
Technical fixes, implemented
We make the fixes ourselves: crawler rules and firewall settings where we have access, sitemap and canonical clean up, structured data that matches visible content and an llms.txt file where it adds value.
Every change is logged. On a live site we work on a staging copy or in small reversible steps, and we confirm each fix with a fresh fetch.
- Robots and crawler access
- Sitemap and canonicals
- Schema that matches the page
- Change log you keep
{ "@type": "LocalBusiness", "name": "Your brand", "areaServed": "Dubai", "sameAs": [ "linkedin.com/..." ] }
- ValidMatches visible page text
- ValidSame name on profiles
- CheckOpening hours missing
Ten pages rewritten to be quoted
We pick up to ten priority pages with you: service pages, product or category pages, an about page and the pages your buyers ask about. Each one gets a direct answer near the top, headings that match real questions, tables where they help and FAQs.
Everything is written to be accurate, sourced where needed and edited by a person. You approve before it goes live.
- Answer first structure
- FAQs from real buyer questions
- Entity facts consistent with your profiles
- Your approval before publishing
- Direct answerTwo sentences with the range
- What moves the priceScope, markets, competition
- Comparison tableRetainer vs project vs tools
- Proof and limitsDated facts, honest caveats
- FAQSix real questions
Retest, report, hand over
We run the same prompts on the same engines again and show you before and after, including what did not move and why we think so. Engines change, so we report what we saw and when.
You get a plain language report and a 90 day roadmap: what to do next, in order, whether you do it yourself, hire us or hire someone else.
- Same prompts, same method
- Before and after report
- 90 day roadmap
- All documents are yours
- DoneAnswer blocks added to 6 service pages
- DoneAI crawlers allowed, sitemap cleaned
- NextThird party profiles to match entity facts
Scope
What is in the Pilot and what is not
We would rather you know now than find out in week three.
| Included in $999 | Not included | |
|---|---|---|
| Website | One domain, one language | Extra domains or languages |
| Pages | Up to ten priority pages rewritten | Full site rewrites |
| Testing | Up to 20 prompts, five AI engines, before and after | Ongoing monthly tracking |
| Technical | Crawler access, sitemap, canonicals, schema, llms.txt where useful | A new website build or migration |
| Authority | Entity facts and profile consistency review | Digital PR and outreach |
| Results | The work, the report and the roadmap | Guaranteed rankings, citations or traffic |
Larger sites, extra markets or ongoing work are quoted separately after the Pilot.
Any platform
We make the changes where your site lives
WordPress
Admin access or a staging copy. We handle robots, schema, sitemaps and content.
WordPress GEO
Shopify
Theme level changes, robots rules and structured data, with a duplicated theme for safety.
Shopify GEO
WooCommerce
Product schema, category copy and cache rules that suit stores.
WooCommerce GEOCustom code
We work through a Git repository and pull requests your developers can review.
Custom site GEOWhat we need from you
Four things, and then we get to work
A short list keeps the 30 days honest. If any item is late, the timeline moves by the same number of days and we tell you.
- Access: admin login, a staging site or a Git repository, shared securely
- A named person who can approve changes within two working days
- Your top buyer questions and your main competitors, so the prompt set is real
- Accurate facts about your business: name, services, locations, contact details
We never ask for your hosting or payment passwords, and we remove our access when the Pilot ends.
What we need from you, and what we never ask for
- ✓Admin login, staging site or Git access, shared securely
- ✓A named approver who answers within two working days
- ✓Your top buyer questions and main competitors
- ✓Accurate facts about your business
- Your hosting or registrar login
- Any payment details
How we judge it
What counts as a good result
AI answers are not stable, so one lucky screenshot proves nothing. We judge the Pilot on three things: the technical fixes are live and confirmed, the rewritten pages are published and accurate, and the same prompt set shows what changed across five engines, including the answers that did not change.
Citations in AI answers can take weeks or months to appear, and some will not. That is why the report separates what we controlled (the work) from what the engines decided (the answers). You will see both.
Week 1 in depth
What the baseline week actually looks like
The first week decides whether the other three are well aimed, so we do not rush it. It starts with the prompt set. You send us the questions your buyers ask, your main competitors and the pages you think matter. We turn that into up to 20 prompts written the way a real person types them into an AI tool, not the way a marketer writes a keyword. A prompt such as "best managed IT support company for a small office in Dubai" tells us far more than the bare phrase "IT support". You approve the list before we run it.
Then we run every prompt on five AI engines and save each answer with the date, the engine and the exact wording we used. We note whether your brand is named, whether it is linked as a source, which competitors appear, which third party sites the engine leaned on, and whether any statement about you is wrong. Wrong statements matter as much as missing ones, because an engine that describes you badly is worse than one that says nothing.
In parallel we check the technical side. We test which AI and search crawlers can fetch your pages, whether a firewall or CDN rule blocks them, whether the main text appears without running scripts, whether your robots rules say what you think they say, and whether your sitemap and canonical tags agree with each other. You can try a quick version yourself with the AI crawler access checker. We also compare your entity facts, meaning your name, services, locations and contact details, across your own site and your public profiles, because mismatches make engines unsure which version to trust.
The week ends with a short findings note. It lists what we found, ranks the fixes by effect and effort, and names the ten pages we propose to rewrite. You can change that list. Nothing in week two starts until you have read it.
- Prompt set written in buyer language and approved by you
- Every answer saved with engine, date and prompt text
- Wrong or outdated statements about you logged separately
- A ranked fix list before any change is made
- Prompt setUp to 20 buyer questions that you approve
- Five enginesEvery answer saved with the date and wording
- Crawler accessRobots, firewall and rendering checks
- Entity factsName, services and locations compared across profiles
- Findings noteRanked fixes and the ten pages we propose
Weeks 2 and 3 in depth
How the fixes and rewrites are handled day to day
Fixes first, writing second, because a perfect page that a crawler cannot read helps no one.
In week two we work through the ranked list in order. Each change is small, logged and checked with a fresh fetch afterwards. On a live site we prefer a staging copy. Where there is no staging, we make one reversible change at a time and keep the previous version ready to restore. You get the change log at the end, written so your own developer or agency can read it.
In week three we write. For each of the ten pages we agree one main question the page should answer, then build the page around it: a direct answer in the first lines, headings that mirror how people ask, short sections that each stand alone, a table where a comparison helps, and FAQs taken from real buyer wording. Structured data is added only where it matches the visible text. A person edits every page. Where a claim needs a source, we ask you for it or we cite a public one and name it. You approve each page before it is published.
- One change at a time, each one confirmed
- Staging copy or reversible steps on live sites
- Answer first pages edited by a person
- Your approval before each page goes live
Record what AI engines say today.
Open access and repair the foundation.
Make priority pages answer ready.
Test again and report.
The calendar
A realistic 30 day timeline
- 1
Days 1 to 3: access and prompts
You send access, a named approver and your buyer questions. We confirm the prompt set and the ten page shortlist in writing.
- 2
Days 4 to 7: baseline and audit
Five engines, up to 20 prompts, saved with dates. Crawler, rendering, schema and entity checks. Findings note delivered.
- 3
Days 8 to 14: technical fixes
Robots and crawler rules, sitemap, canonicals, structured data and llms.txt where it adds value. Change log kept.
- 4
Days 15 to 24: page rewrites
Up to ten pages drafted, reviewed by you and published in batches so problems are caught early.
- 5
Days 25 to 30: retest and handover
Same prompts, same engines, same method. Before and after report, 90 day roadmap and all files handed over.
How the 30 days are spent
- Access and prompts 10%Days 1 to 3
- Baseline and audit 13%Days 4 to 7
- Technical fixes 23%Days 8 to 14
- Page rewrites 34%Days 15 to 24
- Retest and handover 20%Days 25 to 30
What we need from you
The inputs that keep the 30 days honest
The Pilot is fast because the scope is small, and it stays fast only if the inputs arrive on time. The list on this page is short, but each item has a reason behind it, so here is what good looks like.
Access. The best option is a staging site or a limited editor or developer account. For a code based site, read and write access to a Git repository lets us send changes as pull requests that your team can review. For WordPress and WooCommerce, an administrator account created only for us is simplest. For Shopify, a staff account with permission to edit themes and content is usually enough. We never need your hosting login, your registrar login or any payment details. You can revoke our access on day 30 and we will remind you to.
An approver. One person who can say yes or no within two working days. Projects slow down most when drafts wait in an inbox. If your decisions need a committee, tell us in advance and we will plan the review dates around it.
Real questions. The prompt set is only as useful as the questions in it. Ask your sales team what prospects ask on the first call. Ask support what customers misunderstand. Look at the questions that come through your contact form. Those are the prompts that matter.
True facts. We need your correct legal and trading names, services, locations, opening hours, certifications you can prove and any claims you cannot make. If a page says something you are not allowed to say, such as a medical or financial claim, we need to know before we write.
- Staging site, limited account or Git access shared securely
- One named approver with a two working day turnaround
- Questions from sales calls, support and your contact form
- Accurate business facts and a list of claims to avoid
If something is late, the timeline moves by the same number of days and we tell you. It never silently shrinks the work.
Platform differences
What changes from one platform to the next
| Where we work | Typical constraints | How we stay safe | |
|---|---|---|---|
| WordPress | Admin dashboard, theme files, SEO plugin settings | Plugins can output duplicate or conflicting schema; caching can hide changes | Staging copy, one plugin setting at a time, cache cleared and page refetched |
| WooCommerce | Product and category templates, schema output, cache rules | Dynamic cart pages must not be cached; variants can create duplicate URLs | Test on a few products first, then widen |
| Shopify | Theme code, content pages, robots template, structured data | Some system pages cannot be edited; theme updates can overwrite edits | Duplicate theme before editing, changes documented for the next update |
| Custom code or headless | Git repository, build pipeline, server or edge config | Content may render only in the browser; deployments follow your release process | Pull requests, preview builds, your team merges |
The scope does not change by platform. Only the method of making and checking the change does. See WordPress, Shopify, WooCommerce and custom site pages for detail.
WordPress
Admin dashboard and theme files
- Staging copy first
- One plugin setting at a time
- Cache cleared, page refetched
WooCommerce
Product and category templates
- Test on a few products first
- Dynamic cart pages not cached
- Variant URLs checked
Shopify
Theme code and content pages
- Duplicate the theme first
- System pages left alone
- Changes documented
Custom code
Git repository and build
- Pull requests
- Preview builds
- Your team merges
The deliverable
What a finished report contains
The report is written for a business owner first and a developer second. It is plain language, with the detail kept in appendices so the front pages stay readable in ten minutes.
It opens with a one page summary: what we did, what changed on the site, and what changed in the answers. Next comes the before and after table for each prompt and engine, with the dates of both runs, so you can see exactly where your brand appeared, where it did not and where the answer was unchanged. We add a short note on method, because comparing two runs only means something if both were run the same way.
After that we list every technical change with the page or setting affected, the reason and how we confirmed it. Then we list the ten rewritten pages, what each one now answers and what we recommend you watch. A separate section covers what we found but did not fix because it was out of scope, such as a language version, a migration, or third party profiles that need outreach.
The last section is the 90 day roadmap. It is ordered, with owners next to each action, so you can hand it to your team, do it yourself, ask us for a quote or give it to another provider. We state plainly which items we expect to matter most and which are uncertain. Every file is yours to keep.
- One page summary in plain language
- Before and after for each prompt and engine, with dates
- Full change log and rewritten page list
- A 90 day roadmap with owners
Fit
Who the Pilot fits, and who it does not
It fits you if
You run a business with a real website and real buyers who ask AI tools questions before they call. You want the fixes made, not just listed. You have one main market and one language. You can name an approver and share access safely. You would like proof on a small scale before a larger commitment.
It also fits larger firms as a first step
A bigger company can use the Pilot to test one business line or one section of the site. The learnings and the roadmap then guide a wider programme, which we would quote separately.
It does not fit if
You need a brand new website, a migration, several languages or several domains inside one price. You want a promise of citations, rankings or traffic by a date. You cannot share any form of access to your site. You expect ongoing monthly reporting from a one time project.
Think twice if
Your business cannot publish accurate claims, for example because of unresolved regulatory questions, or your site has almost no pages worth rewriting yet. In those cases we will say so after a short conversation, before you pay.
The problem
Objections we hear, answered plainly
"We have an SEO agency already."
Good SEO and good GEO share a base: crawlable pages, clear content, accurate data. We can work next to your agency, keep every change documented and avoid touching anything they own without telling them. The overlap between GEO and SEO is large, and the differences are specific.
"Ten pages seems small."
It is small on purpose. Most sites have a few pages that carry most of the buying questions. Ten well chosen pages tell us whether the method works for you and give your team a pattern to copy. The roadmap covers the rest.
"AI results change daily, so how can you measure anything?"
You are right that they move. That is why we save every answer with its date, run the same prompts the same way twice and report the stable and the unstable parts separately. We do not claim more than two runs can show. See how to measure GEO performance.
"What if nothing changes in the answers?"
It is possible. Engines decide, not us. You would still own the fixed crawler access, the corrected structured data, ten better pages and an honest report on why little moved, plus a roadmap. We do not hide a flat result.
"Is it safe to give a vendor access to my site?"
Only give what is needed. We prefer staging and limited accounts, log every change and ask you to remove access when we finish. We never ask for hosting, registrar or payment passwords.
Platform behaviour
Why we test five engines and not one
ChatGPT, Google AI Overviews and AI Mode, Perplexity, Gemini, Claude and Microsoft Copilot each retrieve and present information differently, and each publishes different controls for its crawlers. As of October 2026 the details keep shifting, so we check current official documentation from Google Search Central, OpenAI, Anthropic, Perplexity and Microsoft before we change anything crawler related.
In practice this means a fix that helps one engine can be neutral for another. A Google focused change such as indexing and snippet eligibility matters most for AI Overviews and AI Mode. Bing indexing matters for Copilot. Source led engines such as Perplexity reward pages that state a clear fact and cite their own sources. We therefore baseline all five and report each engine on its own, rather than one blended number that hides the differences.
If you want the detail per engine, start with ChatGPT SEO, Google AI Overviews optimization and Perplexity SEO. The Pilot applies the shared parts of all of them.
FAQ
GEO Pilot questions
Is $999 really the full price?
Yes. The Pilot is a fixed $999, one time, for the scope above. Third party costs such as a paid plugin you choose to buy are yours and are agreed before we spend anything.
Do you guarantee we will be cited by ChatGPT?
No. No one can honestly guarantee that. We do the work, show before and after, and tell you plainly what changed.
What if my site is large?
The Pilot covers one domain and up to ten priority pages. A larger site is fine, we simply pick the pages that matter most and put the rest in your roadmap.
Can you work on Shopify, WooCommerce and custom builds?
Yes. We work on WordPress, WooCommerce, Shopify and custom code including headless sites, through whichever access you can safely give us.
Will you change my live site without asking?
No. You approve changes. We use a staging copy or small reversible steps, and we keep a change log.
What do I own at the end?
Everything: the content, the code changes, the audit and the report. If you stop after 30 days, nothing renews.
Do you work with companies in India?
We serve clients in the UAE, USA and United Kingdom. We are not taking on India based clients at this time.
How do I start?
Send your website and the five questions your buyers ask using the form below. We reply within one working day with a short plan and the next step.
How much of my time does the Pilot take?
Plan for about an hour in the first week to review the prompt set and the findings note, then short reviews for each batch of pages. The biggest time cost is usually getting access set up and getting drafts approved quickly.
Can I choose which pages are rewritten?
Yes. We suggest ten based on the baseline, and you can swap any of them. We will tell you if a swap is unlikely to help, but the final list is yours.
What happens if I do not approve a page?
It stays as it is. We revise the draft once or twice with your comments. If a claim cannot be supported or you are not comfortable with it, we remove it rather than publish it.
Will you add structured data to every page?
Only where it matches what the page actually says. Markup that describes things a visitor cannot see is a risk, not a help. Our guide to schema for AI search explains the reasoning.
Do I need an llms.txt file?
Sometimes. It is optional and no engine promises to use it, so we add one only when it helps your site, for example a large documentation set. Read whether you need an llms.txt file first.
Can I see what a Pilot looks like before I buy?
Yes. Read what 30 days of implementation looks like, and run the free AI visibility checker to see your starting point.
Keep reading
Guides on this topic
Why your business is not showing in ChatGPT, and how to diagnose it
If ChatGPT never names your business, one of seven causes is usually behind it. Use this diagnostic to find which one, in the order that saves the most time.
Read moreAI search traffic in Google Analytics: how to track ChatGPT and AI referrals in GA4
GA4 can show some visits from ChatGPT and other AI tools, but not all. This guide explains how channels and referrers work, how to set up a report, and where the data stops.
Read moreAI visibility metrics: citation share, share of voice and mention rate defined
Mention rate, citation rate, citation share and AI share of voice are often used loosely. This guide defines each one, shows how to compute it from a fixed prompt set, and explains the limits.
Read moreStart your GEO Pilot
Tell us your website and the five questions your buyers ask. We reply within one working day with a clear next step.
- Fixed $999, one time
- We implement the work for you
- Any platform, delivered in 30 days