Can we keep our current SEO agency?

Yes, and roughly a third of our programmes run that way. They keep rankings, content and links; we take the retrieval half — agent access, passage structure, entity work, citation measurement. We'll share our findings with them directly rather than through you. It only breaks down if their technical recommendations contradict ours, and in that case we'd rather have the argument in month one than month six.

Do you guarantee we'll be cited?

No, and be careful with anyone who does. These systems are generated, non-deterministic and change without notice. What we commit to is the work that measurably shifts the odds — eligibility, structure, entity clarity, corroboration — and to reporting citation share as a sampled rate with a confidence interval rather than a screenshot of one lucky chat.

How do you measure something that changes every time you ask?

By sampling rather than asserting. Every prompt runs at least thirty times per model with temperature, locale and persona held constant, and we report a rate with a margin of error — ±4.1% at n=30. Re-runs use the identical prompt set, so month-over-month deltas are genuinely comparable instead of vibes.

What if our traffic is fine right now?

Then this is cheap insurance rather than an emergency, and we'll say so. The pattern we see repeatedly is rankings holding while click-through quietly erodes for six to nine months before anyone connects it to answer surfaces. The teardown is free precisely so you can find out which situation you're in without committing to a retainer.

What do you need from us?

Access to your site, Search Console, GA4 and ideally server or CDN logs, plus one person who can approve content and roughly two hours a month for the review call. We write, build and publish everything else. If your dev team owns deployment, we hand over tested changes as tickets with acceptance criteria rather than fighting for access.

AI SEO services

Your SEO retainer is covering half of search.

Ten blue links are now one surface among six. We run classic SEO and retrieval optimization as a single programme — so you rank, get retrieved, and get named in the answer. We'll show you which half you're missing before you sign anything.

Get my free visibility teardown See the coverage gap

7 working days · No retainer required to receive it

answer surfaces your buyers now use to shortlist: 6answer surfaces your buyers now use to shortlist
median click loss on the top result when an AI answer appears: -47%median click loss on the top result when an AI answer appears
surfaces a classic SEO retainer is actually built for: 1 of 6surfaces a classic SEO retainer is actually built for

Where your buyers actually askedlast 30 days

Google classic resultscovered
Google AI Overviewsnot cited
ChatGPTnot named
Gemininot cited
Perplexitynot cited
Copilotnot named

Covered by a classic SEO retainer: 1 of 6Illustrative

The coverage gap

Compare what you're paying for against what search now is

Fourteen workstreams. Most retainers cover the first six well and stop. Toggle to see the difference — then ask your current agency which column they're on.

Technical crawlabilityIndexation, site speed, canonical hygiene, Core Web Vitals.

Keyword and intent mappingCommercial query research mapped to pages and funnel stage.

On-page optimizationTitles, headings, internal linking, page-level relevance.

Content productionBriefs, drafts, publication cadence against a content calendar.

Link acquisitionEarned links and digital PR to build domain authority.

Rank and traffic reportingPosition tracking, sessions, conversions in GA4.

Agent crawler accessGPTBot, ClaudeBot, PerplexityBot and Google-Extended verified in server logs — not assumed.

Passage-level structuringAnswers rewritten to stand alone in the first 40 words, the unit retrieval actually lifts.

Entity and knowledge graphOrganization, sameAs and knowledge-panel alignment so models can resolve who you are.

Third-party corroborationIndependent sources repeating your claims — the signal retrieval weights most heavily.

Citation share measurementYour prompt set re-run monthly across six models, sampled not screenshotted.

Extractability engineeringServer-side rendering, HTML tables, transcripts — content agents can actually read.

Freshness and date signalsdateModified and refresh queues for the queries that carry temporal intent.

Competitor slot displacementIdentifying who holds the citation you want and what would move them out of it.

Six of fourteen workstreams covered. Your rankings are being maintained while five of the six surfaces your buyers use go unmeasured and unworked.

Close the gap

The programme

Six workstreams. One number they all report to.

See a sample month

01
Visibility teardown
Your classic rankings and your presence across all six answer surfaces, measured on the same commercial query set so the two are finally comparable.
Your baseline, in writing
02
Technical and access repair
Rendering, indexation and agent-crawler access fixed together. Half the brands we audit are blocking retrieval bots without knowing it.
Verified in server logs
03
Content built for both
Pages that rank and passages that get retrieved. Same asset, two jobs — answer-first structure without sacrificing depth.
Published and tracked
04
Entity and authority
Schema, knowledge-panel alignment and the third-party corroboration that decides whether a model is confident enough to name you.
Validated on publish
05
Measurement across surfaces
One monthly report covering rankings, sessions, citation share and the competitor holding each slot. No two dashboards that disagree.
One page, one call
06
Revenue attribution
AI-referred sessions tagged, tracked and tied to pipeline in GA4, so you can see what the programme is actually worth.
Reported in leads

Proof

Three programmes, one shared pattern

B2B SaaS
Lumen
Rankings had been stable for two years while organic revenue slid 28%. The teardown found them cited in 12% of category prompts. Rendering fix plus 22 passage rewrites closed most of it.
+41%Organic pipeline6 months
Healthcare
Verdant
Regulated category where models default to conservative sources. We built author credentials, medical review sign-off and structured FAQs until they became a safe citation.
+63%Organic enquiries5 months
Industrial
Atlas+Co
Ranked page one, invisible to every retrieval agent — an over-tuned WAF was 403ing four crawlers. Found in week one; citations appeared within a crawl cycle.
+318%Cited queries4 months

The first 90 days

You'll know if it's working by day 45

Days 1–14
Teardown and baseline
Query set built, rankings and all six answer surfaces measured, crawl and render issues found. You get the report and the priority list.
Report delivered
Days 15–35
Foundations
Technical repair, agent access, schema and entity alignment — everything that decides whether you're eligible at all.
Foundations live
Days 36–65
Content and structure
Money pages restructured answer-first and published, with corroboration outreach running in parallel.
First movement visible
Days 66–90
Compound and defend
Second content wave, competitor slot displacement, and a monitoring cadence you keep whether or not you stay.
Both metrics reported

Investment

Pick the size of the problem

No setup fees, no twelve-month lock-in. Month-to-month after the first ninety days.

Start here
Visibility Teardown
Teams that suspect the gap is real and want the number before committing.
Free7 working days
- Up to 25 commercial queries
- Classic rankings baseline
- All six answer surfaces measured
- Agent crawler access check
- Prioritised fix list
- 30-minute findings call
Request the teardown
Core programmeMost chosen
AI SEO Programme
Businesses that need both halves of search working against one revenue number.
from $7,500per month · 90-day term
- Everything in the Teardown
- Technical and agent access repair
- Six answer-first pages per month
- Schema and entity implementation
- Digital PR and corroboration
- Monthly rankings + citation report
Talk about the programme
Multi-market
Category Dominance
Brands defending a category across several markets, languages or business units.
Customscoped to markets
- Everything in the AI SEO Programme
- Multi-market and multilingual query sets
- Competitor slot displacement
- Publisher and trade press programme
- Dedicated strategist and content lead
- Quarterly board-level reporting
Scope an engagement

Before you ask

The questions that decide it

Answer

Is AI SEO just SEO with a new name?

Partly, and we'd rather say so than invent a discipline. The foundations overlap — crawlability, structure, authority. What's genuinely different is the unit of optimization: classic SEO optimizes a page for a position, retrieval optimization optimizes a passage for extraction by a system that will never show your page. Different unit, different diagnostics, different fixes, and frequently different pages. Any agency selling you 'AI SEO' that can't name that distinction is selling you a rebrand.

Ask us something harder

See both halves of your search visibility. Free.

Send your domain and up to 25 commercial queries. You get your classic rankings and your presence across all six answer surfaces on the same query set, a prioritised fix list, and a 30-minute call to walk through it. No retainer required to receive it.

Get my free visibility teardown Email the team

7 working days · Founded 2014 · 40+ specialists in-house

Works hardest alongside

All AI search services

Your SEO retainer is covering half of search.

Compare what you're paying for against what search now is

Six workstreams. One number they all report to.

Visibility teardown

Technical and access repair

Content built for both

Entity and authority

Measurement across surfaces

Revenue attribution

Three programmes, one shared pattern

Lumen

Verdant

Atlas+Co

You'll know if it's working by day 45

Teardown and baseline

Foundations

Content and structure

Compound and defend