Deliverables
What's actually included
Every deliverable, itemised by tier. Also itemised: the things we leave out on purpose, and the reasons. If you're comparing us against a £400-a-month package, the section at the bottom is the one to read first, it's the honest account of why the gap exists and when the cheaper option is genuinely the right call.
Last updated 31 July 2026
How we structure the work
Four layers, delivered in this order. The order matters more than the contents.
Foundation, can machines reach, parse and understand your content?
Extraction, can they lift a clean, correct answer out of it?
Evidence, is there anything in your content worth retrieving?
Authority, has the model learned your brand at all?
Most packages in this market do Foundation and Extraction thoroughly, then describe Authority in a single bullet point. That's the wrong way round. Ahrefs, across 75,000 brands, found branded mentions correlate 0.664 with AI visibility against 0.218 for backlinks. The layer everyone treats as a footnote is the one carrying the strongest signal.
AI Visibility Baseline: £1,950 one-off
Measurement
- Share of voice across ChatGPT, Google AI Overviews and Google AI Mode
- Sampling: ~50 runs per prompt per platform, 30 prompts, approximately 1,500 sampled answers per source
- Wilson 95% confidence intervals on every figure reported
- Supplementary panel: Gemini, Claude, Perplexity, Copilot
- Competitive set discovered from the answers themselves, using species-accumulation modelling rather than your assumptions about who you compete with
- Prompt-level breakdown: where you appear, where you don't, who appears instead
Technical audit
- AI crawler accessibility verified per agent, GPTBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, Bingbot
- Bing indexation check (ChatGPT retrieves through Bing; most sites have never been looked at)
- Existing schema inventory and validity
- Answer-position analysis on your key commercial pages
- Citation source analysis: which of your pages and which third-party domains the engines currently pull from
Deliverables
- Written report, findings stated with intervals
- Prioritised action plan, ordered by expected impact against effort, usable by any agency
- Full raw dataset (CSV) and complete methodology
- 60-minute walkthrough call
Turnaround: 2–3 weeks. No ongoing commitment.
Local AI Visibility: £750 setup, £595/month
Setup
- Local visibility baseline for your service area, with intervals
- Google Business Profile audit: primary and secondary categories, description, hours including holiday hours, attributes, service areas
- NAP consistency audit across directories, with corrections submitted
- Review audit across the platforms that matter in your sector, consistency of rating across platforms matters more here than raw volume
- LocalBusiness, Service and FAQPage schema on service and location pages
- Crawler access and Bing indexation
- Answer-first restructuring of your primary service pages
Monthly
- Measured share of voice for your local prompt set, reported with intervals and an explicit verdict on whether movement is distinguishable from noise
- 2 answer-shaped pages, published to your site (see the content standard below)
- Google Business Profile posts: 2 per month
- Review generation workflow and monitoring
- Directory and community presence building, submitted individually
- Monthly report
Why the numbers are small: two pages and two GBP posts is deliberately fewer than a cheaper package will quote you. The content standard below is why. We'd rather publish two pages that carry something than eight that don't.
Agent-Ready Foundation: £3,500 one-off
Retrieval
- Crawler access configured and verified per agent
- robots.txt and llms.txt handling (llms.txt included at no charge and not presented as a visibility deliverable, see exclusions)
- Server-log baseline: which AI crawlers visit, how often, which pages
Structured data
- Organization, Person, Service, Product, FAQPage, LocalBusiness, BreadcrumbList as applicable
- Validated against Google's testing tools and manually reviewed
- Entity consistency: sameAs linking across your profiles, consistent naming
Answer structure
- Answer-first restructuring of commercial pages, direct answer inside the first 40–60 words. Zyppy found 44.2% of LLM citations come from the first 30% of a page
- Heading hierarchy and terminology consistency
Commerce readiness (where applicable)
- Product feed and structured product data audit
- Pricing consistency across site, marketplaces and feeds, agents compare, and inconsistency gets you skipped
- Protocol readiness assessment: ACP, UCP, MCP
Attribution
- GA4 configuration for AI referral isolation
- Server-log analysis pipeline for AI crawler activity
- Reporting template separating impressions, CTR with and without an AI summary, and conversions
AEO Practice: £2,950/month
Everything in the Baseline, run continuously, plus:
Measurement, every cycle
- 30 prompts, ~50 runs each, ~1,500 answers per source
- Intervals on every figure, verdict stated in words
- Competitive tracking against your discovered set
- Citation source tracking: which pages earn citations, which don't
- Quarterly supplementary panel across Gemini, Claude, Perplexity, Copilot
Content, 4 pages per month
Each page must clear all of the following before it publishes:
- Answers one real question, sourced from your sales calls, Search Console data or the prompt gaps in your measurement, not from a keyword tool
- Direct answer in the first 40–60 words
- At least one piece of proprietary data, original observation or first-party evidence. No page ships that only restates what's already on the web
- Inline citations to primary sources
- Human review before publication
Citation building
- Identification of the third-party sources the engines already trust in your category
- Outreach for legitimate inclusion: industry publications, genuine directories, expert contribution, original data placement
- Verification that placements are actually being retrieved
Reporting
- Monthly report with intervals and noise verdicts
- Intervention log with dates, so causation can be assessed later rather than asserted now
- Quarterly strategy review
Authority, from £7,500/month
Everything in Practice, plus the layer that carries the strongest signal.
Entity establishment
- Naming consistency across every profile, directory and listing
- Structured entity signals and cross-platform corroboration
- Wikidata where genuine notability supports it. We will tell you if it doesn't, a deleted entry is worse than none
Digital PR
- Aimed specifically at the sources these systems retrieve from, which is not the same target list as traditional PR
- Original research and data production designed to be cited rather than to cite
Community and review
- Reddit and sector community presence, participatory rather than promotional. Perplexity weights community sources heavily
- Review platform strategy across the platforms that matter in your category
Category authority
- Thought leadership placement, speaking, podcast and industry positioning
- Direct access to me, not an account manager
What we don't include, and why
This is the section worth reading twice, because it's where most of the price difference lives.
Bulk backlink packages
You will see "60 high-DA backlinks per month" on cheaper proposals. There is no legitimate way to source sixty links a month at that price. It's private blog networks, paid placements or link farms, and it breaches Google's link spam policies.
Two further points. DA and PA are proprietary third-party metrics, not Google signals, a provider selling by DA is selling by a number Google doesn't use. And on the evidence, links are the weaker lever anyway: 0.218 correlation with AI visibility against 0.664 for branded mentions.
We build citations through legitimate placement, individually, at a rate of a handful a month. That's what earned inclusion actually looks like.
Bought brand mentions
Sometimes described as "entity authority building" and sold alongside bulk links. Same mechanism, same eventual bill.
llms.txt as a deliverable
Cyrus Shepard's 2026 synthesis scored 23 citation factors across 54 studies, patents and experiments. llms.txt came last, at 2.0 out of 10, with no credible evidence of effect in any source. Google has said it won't use it. We'll create one for free. It won't appear on an invoice.
HowTo and Speakable schema
Google withdrew HowTo rich-result support in 2023. Speakable was only ever available to news publishers in specific programmes. Both still appear on deliverables lists across this market, which tells you those lists were assembled from old blog posts rather than current practice.
High-volume content
The public record is consistent: sites that scaled AI content climbed steeply and then declined further and for longer once Google's spam systems caught up. Our caps are deliberate and we'll argue with you about them.
Guaranteed rankings, citations or share of voice
A single measurement in these systems carries a margin of error of ±9 to ±13 points. Guaranteeing an outcome inside that variance is either innumeracy or dishonesty.
Why we cost more than a £400/month package
Fair question, and it deserves a straight answer rather than a defensive one.
Because the deliverable is different. A £400 package sells inputs: number of posts, number of links, number of listings. We sell a measured change in share of voice with a confidence interval attached. Those are different products that happen to share a category name.
Because measurement at depth is expensive. Roughly 1,500 sampled answers per source per cycle is what gets a figure to ±1–2 points. A single daily check gets you ±9–13, which cannot detect the movement you're paying for. Most reporting in this market is seven coin flips a week with a trend line drawn through them.
Because the content standard is slow. Requiring at least one piece of original evidence per page means four pages a month, not twelve. Volume without information gain is the failure mode that got brands penalised.
Because we don't buy links. That removes the cheapest way to appear busy.
Because you get a thirty-year practitioner rather than a process. Which is a claim, not a proof, so read the dated public record and the ledger, including the failures, and judge it.
When the cheaper option is genuinely right
Three situations. I'd rather say so than have you find out later.
Your technical fundamentals are neglected. If your site is slow, badly structured and missing basic schema, a competent technical package at £400 will do you more good than our measurement will. Fix the foundation first, with anyone. URL accessibility is the highest-scoring citation factor in the evidence at 9.5 out of 10, and nothing above it works without it.
You want deliverables, not outcomes. If what you need is content published and listings created, and you'll judge success on activity, buy the cheaper package. You'll get more of it.
Your budget is genuinely constrained. Under about £500 a month, the honest answer is a monitoring tool, a properly finished Google Business Profile and a review process. We'll show you how in the £299 consultation and you'll capture most of the available value in local search.
The one situation where cheap is the wrong choice: if anyone's proposal includes bulk backlinks or bought mentions, the price stops being the relevant question. You're buying a risk to your existing search performance, and it usually surfaces a year later when it's expensive to unwind.
Comparing two proposals properly
Put them side by side and ask about each one:
| Ask | What a good answer sounds like |
|---|---|
| Sampling depth and margin of error | A specific number of runs per prompt and a stated interval |
| Platform coverage and method per platform | Named platforms, with different methods acknowledged |
| Link acquisition method | Individually earned placements, single digits per month, named targets |
| Content quality gate | An originality standard, not just a page count |
| A failed intervention | An actual example, with data |
| llms.txt | Free or absent, not a headline deliverable |
| Brand and entity work | Something substantive, not one bullet |
| Google's May 2026 guidance | Familiarity, and a considered position on it |
A provider offering more deliverables for less money isn't necessarily worse. But if they can't answer columns one, three and five, the extra volume is either low-value or actively risky.
FAQ
Why do you only produce four pages a month when others offer twelve?
Because every page has to carry at least one piece of proprietary data, original observation or first-party evidence, and that takes time. Content that restates what's already online gives an AI system no reason to retrieve it, and volume without information gain is the pattern that preceded some very public traffic collapses.
Do you do link building?
We build citations, individually earned placements in sources the engines already retrieve from, at a handful a month. We don't sell bulk backlink packages, because sixty links a month at commodity prices means link schemes, and because branded mentions correlate roughly three times more strongly with AI visibility than links do.
A cheaper provider offers more deliverables. Are they lying?
Probably not lying, delivering a different product. Their unit is activity; ours is measured movement with a confidence interval. Ask both of us the eight questions above and compare the answers rather than the lists.
Can I buy just the parts I need?
The one-off tiers are fixed scope and stand alone. On the ongoing tiers, measurement is not optional, without it there's no way to tell whether anything we do works, which makes the rest unaccountable.
Do you white-label or subcontract?
The work is delivered under our own name, white-label. Where specialist tooling or capacity is involved, it sits behind our process and we remain the single point of accountability for the deliverable.