Managed AI Search

Managed AI Search, with half the fee tied to the result.

Managed AI Search is an ongoing service that tracks how AI engines describe and cite your business, improves the pages and sources behind those answers, and measures the enquiries that follow. Schmitdy runs the research, content, outreach and reporting with your team in Slack, Teams or WhatsApp.

From£1,200 / month · no long lock-in

↺ Half to start. Half when either agreed monthly goal is met.

A fresh plan every 30 days · First baseline in week one

What does the free audit include?

Get your free AI Search audit →

Schmitdy and a marketing team reviewing shared search goals and results

The shift

What changes when buyers ask AI for a shortlist?

An AI answer can name a business, describe its offer and link to supporting sources before a buyer visits its website. Those are three separate opportunities to get the facts right. We check the questions that matter to your business, find missing or misleading evidence, and turn that research into work your team can approve.

Good SEO still matters. Google's guidance on AI features says existing SEO practices apply and no special AI schema is required. We improve access, useful content and supporting evidence. Engines decide what to include.

The daily loop

How does the daily research become shipped work?

We track an agreed question set across ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews. Each read records the engine, date, question and cited sources. Daily checks guide the work; completed-week comparisons help us judge a trend. Missing answers and unavailable engine data are reported as gaps.

Schmitdy and an analyst prioritising a small set of actions from fresh evidence

What the daily signal surfaces

Questions worth testingSales questions, search demand and related engine queries inform the tracked set. These are research prompts, not a feed of private buyer conversations.
Source gapsA cited guide, comparison or public discussion names a competitor but leaves your business out.
Answer-position changesWhere your brand appears in the sampled answers, compared with the same questions and engines in the previous complete week.
Share-of-voice shiftsYour slice of the category answer growing or shrinking against named competitors, per engine.
Sentiment changesThe engines start describing you more cautiously, or a competitor more warmly.
Competitor sourcesThe competitor pages linked in sampled answers. We inspect what they explain and substantiate before choosing a response.
Pages to revisitA decline in observed citations prompts a check for outdated facts, missing evidence or changed sources. The decline alone doesn't identify its cause.

How the daily-to-weekly loop runs

DailyPull the read across all five engines. Flag what moved and why it matters to pipeline.
TriageRank the day's signals by revenue impact, not by how loud they are.
Weekly commitThe week's signal decides the week's ship list. The plan bends to what the engines are actually doing.
ShipWe publish, rewrite, place, and engage against that list.
Read backVerify the published work, then compare results over the agreed period. A movement after a change is evidence to investigate, not proof that the change caused it.

The question set, countries, languages and engines are agreed before measurement. When coverage changes, we mark the change so the comparison remains clear.

Want to see what your month one would look like?Marco maps your category, competitors, and the fastest wins, free.

Free AI Search audit →

Want to test the sequence internally first? Open the full 90-day roadmap with inputs, owners and decision gates.

The four pillars

The whole operation rests on four things.

Buyer questions

A versioned set of questions built from your sales conversations, search demand and category research, prioritised by commercial relevance.

Source graph

The pages, threads, reviews, roundups, and creators the engines pull from for your category, and where the gaps are between you and the cited brands.

Verified work

Approved pages, useful community contributions and editorial outreach with a record of what happened. Publication decisions remain with the editors and communities.

In your channels, real time

You're in our Slack, Teams, or WhatsApp, and we're in yours. Not a weekly status call.

Slack Teams WhatsApp

What we actually ship

Five ways we move your position. Here's what lands each month.

AI agents handle the daily work. Expert human growth operators assess every claim, asset and placement before it ships.

Optimise what's already there

We improve the pages that answer important buyer questions: clear claims, readable passages and evidence people can check. Structured data describes the visible offer. It does not replace useful content or guarantee a citation.

On-site rewrites of the pages closest to the prompts your buyers convert on
Comparison and category pages restructured so the engine can lift a clean answer
Schema, entity, and passage fixes that make your content easier to quote

Schmitdy and an editor revising an existing business page

Create the pages the category is already asking for

Some answers you'll never win with the pages you have, because the page doesn't exist yet. We build evergreen content engineered for citations, aimed at the exact gaps the daily read surfaces. You get it as a draft pack to review before anything goes live.

Roughly 12 owned blog articles per month as a planning example; the actual volume is agreed in your scope
Comparison pages targeting the "X vs Y" prompts buyers ask before they choose
Use-case explainers for the jobs your buyers are trying to get done
Source-targeted pieces prepared for relevant category roundups and guides

Slack Teams WhatsApp

Schmitdy handing a researched article draft to a specialist reviewer

Contribute to the public communities your buyers read

We inspect the public threads and discussions cited in your category, then contribute where your team's expertise helps. We follow community rules and disclose relevant affiliations. Private groups and WhatsApp chats are collaboration spaces, not public sources we promise an engine can read.

Genuinely useful answers in the subreddits and forums the engines cite for your category
LinkedIn and group contributions where your team's expertise is the value, not the pitch
Engagement timed to the threads the daily read flags as feeding the answers

Schmitdy and a specialist preparing a helpful community answer

Reach relevant editors

We identify guides, comparisons and expert articles cited in the answers we sample for your category. We choose relevant publications from that source analysis and develop pitches that help their readers. Editors decide whether to publish.

A ranked outreach target list built from sources found in the sampled answers
Tailored pitches and follow-ups to relevant category guides and publications
A record separating pitches sent, replies received and verified publications
Researched expert articles prepared for editorial consideration; placements are subject to acceptance

Schmitdy and a colleague preparing an editorial story and its evidence

Turn buyer questions into YouTube proof

AI engines fan a buyer question out into reviews, explainers and how-to searches. We turn the strongest video gaps into useful expert videos, then package each one so buyers and engines can find and understand it.

Topics and scripts built from tracked prompts, engine fan-outs and source gaps
Titles, thumbnails, descriptions, chapters and transcripts shaped for discovery and citation
Every script, claim and final cut assessed by expert human growth operators
Useful video evidence built to answer the question, not branded filler

Schmitdy filming a useful explanation with a human specialist

Works with you in

Slack, Teams, or WhatsApp. We're in your channels, you're in ours.

No weekly status call you forget to dial into. The team runs the channel from inside the tools you already live in. You see what's shipping, you ask for changes by message, and the daily signal lands where you'll actually read it.

Slack

Teams

Schmitdy and a customer keeping decisions and delivered work in one shared conversation

What's included

Everything, in three buckets.

Baseline and tracking

An initial baseline, then completed-week visibility comparisons
Daily checks across the agreed five-engine coverage, with missing data labelled
A versioned question set with the language, market and measurement dates recorded
Source-gap analysis: where competitors get cited and you don't

Execution

New pages plus human-reviewed rewrites, around 12 owned articles a month as a worked example
On-site optimisation: schema, entities, passages, comparison and category pages
Editorial and PR outreach into the sources the engines trust
Community and forum work, only where the answer earns its place

Team and reporting

A senior team in your Slack, Teams or WhatsApp
Your Company Brain is included: a connected knowledge and reporting layer on infrastructure you own, yours to keep when the engagement ends
A fresh 30-day plan with an owner and status for each piece of work
Separate reporting for sampled visibility, citations, referred visits and qualified enquiries, using the data you connect

Measurement

What does an improvement actually mean?

We agree the baseline, two monthly goals, sources, calculations and start and end dates in writing. Each measure answers a different question.

Before the baseline, the written method sets the runs per question and engine, how repeated answers are combined, and the country, language and available personalisation controls. Each comparison reports its actual number of valid answers and any changed settings. Small samples can move by chance; unavailable controls and missing runs remain stated limitations.

Read the full AI Search measurement method, including share of voice, position, sentiment, accuracy, missing data and the difference between attribution and causation.

Measure	What we count	What it tells you
Brand visibility	Valid sampled answers naming your brand, divided by all valid sampled answers in the agreed set.	How often you appear in that sample. It isn't total market share.
Website citations	Sampled answers linking to your website, with the exact cited URLs retained.	Whether your pages are used as sources. A brand mention doesn't require a citation.
Referred visits	Recorded site sessions with an identifiable AI referrer or campaign signal.	Trackable visits, not everyone who saw an AI answer. Lost referral data leaves a gap.
Qualified enquiries	Real business enquiries meeting the agreed fit criteria, with bookings and CRM status checked where connected.	Commercial progress. We distinguish direct referral evidence from self-reported or assisted discovery.

Google includes AI-feature traffic within Search Console's Web reporting; it isn't a separate AI Overview conversion count. OpenAI documents ChatGPT referral tagging, which helps identify visits. Neither source makes visibility equivalent to revenue. Unavailable analytics or CRM data stays unmeasured.

Who it's for

Honest about fit.

This is for you if

Your buyers research with AI before they ever talk to sales
You want AI search run as an owned channel, not bolted onto an SEO retainer
You can act on what we surface: sign off content, give us a contact, move when the signal moves
You'd rather pay for the outcome than for activity

This is not for you if

You want the cheapest audit on the market and nothing after it
Nobody on your side can sign things off
You expect the whole category won in 30 days. This compounds; it doesn't teleport

Not sure if you're a fit? Ask Marco.Bring the messy version. He'll tell you the smallest scope worth starting with.

Free AI Search audit →

FAQ

The questions teams ask before they hand it over.

How does the result-based fee work?+

Each 30-day service month has two measurable goals agreed in writing, with its own baseline, method and dates. Half the fee is invoiced to start. Meeting either goal earns that month's second half. If neither is met in that month's measurement period, the result payment isn't charged. The engagement rolls month to month with 30 days' written notice.

How is Managed AI Search different from SEO?+

Managed AI Search adds repeated AI-answer sampling, citation analysis and source outreach to useful content and technical SEO work. We check how your business is described in answers as well as whether people reach your site. Keep your SEO programme: Google says the same fundamentals apply to its AI features, with no special AI markup required.

Why £1,200 to start, and how does it scale?+

£1,200/month gets a real operator on the channel: baseline, daily reads, prioritised plan, and shipped work each cycle. As we go deeper on content volume, outreach, and community, the scope and the fee scale with it. You're never paying for a bigger team than the channel needs yet.

How soon can we measure progress?+

We establish your first baseline in week one once the question set and access are ready. Work begins in the first cycle, subject to your approvals. We compare complete periods using the same coverage. Engines control recrawling and answer selection, so we don't promise a date for a citation, a ranking or an enquiry.

Can you guarantee a ChatGPT recommendation?+

No. Schmitdy controls the work it delivers, not the answers an engine generates. Results vary by question, location, language, timing and engine. The result-based fee applies to the two goals in your monthly agreement; it is not a promise of a particular AI answer or editorial placement.

What happens after I request the free audit?+

The audit form asks about your website, offer, competitors, target questions and business fit. Those details help us assess the scope and the next step. Requesting a free audit does not start the paid monthly service. Your proposed scope, access, approvals and goals are agreed before delivery begins. Managed AI Search includes your Company Brain on infrastructure you own. You keep it when the engagement ends; no three-month commitment is required.

Do we have to leave our Google SEO behind?+

No. Keep it. Clear structure, reliable sources and genuine expertise support useful search content too. We add AI-answer research to that work without asking you to abandon an existing search channel.