Building Your First AI Visibility Package

Cornerstone Guide · 21 minutes · Introduces the AI Visibility Package Blueprint

Summary

The first service you sell is the most consequential one you will design. It shapes the clients you attract, the work you learn to deliver, and the language the market uses to describe you. This guide sets out the LaunchAnAEO methodology for designing that first offer and introduces The AI Visibility Package Blueprint — five layers that must line up before a service is credible enough to be bought at a non-commodity price.

Introduction

Most new agencies build their first package the wrong way round. They begin with a list of activities they know how to do, group those activities into "Starter", "Growth" and "Pro", pick prices that feel brave but not offensive, and put the result on a page. The result is usually competent. It is almost always forgettable.

Building a first package is not a template exercise. It is the moment a new agency decides what kind of business it intends to be. Get it right and the rest of the first year is a series of refinements. Get it wrong and the next twelve months are spent quietly apologising for an offer that neither the founder nor the buyer quite believes in.

This guide is written for entrepreneurs, freelancers, consultants and early-stage agencies designing their first AI Visibility service. It is deliberately long — a first package is worth a slow reading.

By the end you will have the vocabulary of the AI Visibility Package Blueprint and a clear method for building an offer that the market reads as considered rather than improvised.

A Package Is Not a Rate Card

Before anything else, the distinction that decides everything.

A rate card lists what an agency can do and at what price. It is a menu. It answers "what could we buy from you?".

A package is a design. It answers "what will change for us if we hire you?". It refuses more work than it accepts, describes a specific transformation, and reads as though the agency has already thought carefully about businesses like the buyer's.

Rate cards compete on price. Packages compete on clarity. The AI Visibility market is early enough that clarity is by far the more valuable of the two, and it is available to any agency willing to think for a few uncomfortable weeks before writing anything down.

The buyer of a first AI Visibility engagement is rarely an expert in the discipline. They are frequently a founder or a marketing lead who has read a handful of articles, watched the discipline appear in conversations, and formed a working intuition that "we should probably be doing something about AI". They are not shopping for prompts, audits or optimisation tasks. They are shopping for confidence that someone competent has thought about this on their behalf.

Everything that follows assumes that reader.

The Framework — The AI Visibility Package Blueprint

The Blueprint is the model LaunchAnAEO uses to turn a market position into a defensible, buyable service. It is deliberately small: five layers, in order, resolved before anything is written on a website.

The Blueprint is a diagnostic, not a template. Two agencies working through it can produce entirely different offers and both be right. What it will not tolerate is a package where the layers disagree with each other — a positioning that speaks to founders paired with a cadence designed for procurement teams, or an outcome that speaks of strategic representation paired with deliverables that describe keyword research. When the layers disagree, the market reads the offer as improvised, and prices it accordingly.

Each of the five layers has its own section below. Read them in order, and treat the price at the end as a summary of the previous four decisions rather than as an independent one.

Layer 1 — Positioning

Positioning is who the offer is for, what business context they are in, and the change the offer is designed to produce for that specific group. It is the layer new agencies most often skip, because it is the layer that feels most like turning work away.

Good positioning for a first package tends to share three characteristics. It names a buyer, not a job title. It describes a business context, not a demographic. And it takes a position on the change the work will produce, rather than on the activities that produce it.

A first-package positioning that reads "We help ambitious B2B service firms be found, understood and cited by AI answer engines when their next best client asks for a recommendation" is doing several difficult things at once. It refuses e-commerce brands. It refuses tactical buyers who just want an audit. It commits to a specific decision moment — the point at which a prospect is asking another system for a recommendation — and it stakes ownership of what happens in that moment.

Compare that to "We help businesses improve their AI Visibility". The second sentence is not wrong. It is simply available to every other agency in the category, which is another way of saying it is worth nothing.

The temptation is to keep positioning wide "so that we do not miss opportunities". In practice, wide positioning misses more opportunities than it captures. Buyers select agencies that appear to have been designed for them. A wide package appears to have been designed for no one.

Layer 2 — Promised Outcome

Once positioning has decided who the offer is for, the promised outcome decides what will demonstrably change for them. This is not the deliverables. Deliverables belong two layers down. The outcome is the sentence a client could use, six months in, to explain to their board what the engagement produced.

Three tests distinguish a promised outcome from a description of activity.

• The observable test. Could a reasonable person, looking at the client's business six months from now, tell whether the outcome had occurred? If the only evidence is a report the agency itself produced, the outcome is really an activity.
• The transferable test. Would the outcome still matter if AI Visibility were called something else next year? Outcomes anchored to a business change survive vocabulary shifts; outcomes anchored to a technique do not.
• The honesty test. Is the outcome something the agency controls, or something it merely influences? Confusing the two produces the guarantees that destroy trust in the second quarter.

Well-formed promised outcomes for a first package tend to be modest, specific and time-bound. "Within ninety days, your brand appears accurately and favourably in AI-generated answers to the five decision-shaping questions your buyers actually ask, and you have a maintained system for keeping it that way." That is a sentence a client can carry into a board meeting. It is also a sentence the agency can be held to, which is the point.

Layer 3 — Deliverables

Deliverables are the artefacts the buyer receives that produce the promised outcome. They are the layer new agencies over-invest in, because they are the easiest to write. A page of deliverables feels substantial. It rarely is.

Two disciplines make the deliverables layer strong.

The first is subtraction. A first package is stronger with three well-explained deliverables than with twelve listed ones. Each deliverable should be small enough to describe in a sentence and clearly connected to the promised outcome. If you cannot say, in one line, how a deliverable moves the outcome forward, it does not belong in the offer.

The second is buyer legibility. A deliverable that only another practitioner would recognise — "structured data audit", "entity consolidation pass", "citation-surface diagnostic" — belongs in the statement of work, not on the offer page. On the offer, deliverables are named in the buyer's language. "A single document that shows how AI systems currently describe your business, what they get wrong, and what to change first" is a deliverable a founder can picture. "AEO baseline audit" is not.

A useful test: give the deliverables list to someone outside the industry and ask them, without prompting, what they think each item is for. If they cannot answer for any single item, that item is not yet written for a buyer.

Layer 4 — Cadence

Cadence is the rhythm at which the work happens. It is the layer most often chosen for the agency's comfort rather than the client's evidence of progress, and it is the layer that most quietly determines whether a client renews.

AI Visibility work is inherently continuous — answer engines change, the client's competitors publish, the client's own story evolves — which makes cadence more consequential than it appears. A first package should distinguish clearly between the initial engagement, which produces the first version of the outcome, and the ongoing rhythm, which maintains and compounds it.

Two cadence questions decide most of the design.

• How often does the client see visible progress? Not a report — a moment where something they care about has demonstrably moved. Monthly is a reasonable floor for a retainer; quarterly is a ceiling.
• How often does the agency make a decision that affects the outcome? This is usually more frequent than the client-facing cadence, and it is the rhythm the agency's internal systems have to sustain.

A first package that promises a strategic AI Visibility outcome on an ad-hoc cadence is telling the buyer that the outcome is not really the point. A first package that promises the same outcome on a monthly cadence, with a clear description of what happens in each month, is telling the buyer that the discipline has been thought through.

Layer 5 — Price

Price is the summary of the four decisions above. If they have been made deliberately, price is the least difficult part of the Blueprint. If they have been skipped, price becomes the whole argument — which is how agencies end up defending a number they have not earned the right to charge.

Three principles govern price in a first package.

The first is that price should be defensible without discounting. If the offer only closes when the price is reduced, the previous four layers have not done their work. The correct response to a first-package price failure is almost never a discount; it is a return to Layer 1 or Layer 2.

The second is that price should be legible. A buyer who has read the positioning, the outcome, the deliverables and the cadence should reach the price and think "that is roughly what I expected this to cost". Prices that surprise the buyer — in either direction — are signals that a layer above is not communicating what it should.

The third is that price should be a decision, not a negotiation. A first package with three deliberate prices — an initial engagement, a monthly retainer and an occasional advisory rate — is easier to hold than a first package that quotes bespoke figures for every enquiry. The market rewards agencies that appear to have already decided what their work is worth.

The Mental Model — Confidence Before Capability

Every layer of the Blueprint is, underneath, a confidence signal.

New agencies tend to over-invest in demonstrations of capability — credentials, technical depth, methodology diagrams — and under-invest in the signals that generate confidence. Clarity. Specificity. Restraint. A package that says less, but says it precisely, generates more confidence than one that lists everything the agency can do.

This is the reason the Blueprint refuses to start with a rate card. A rate card demonstrates capability. It does very little for confidence, because it does not appear to have been designed for the buyer in particular. The Blueprint's five layers, resolved in order, are the shortest route to an offer that reads as considered — which is another word for confident.

Two Worked Scenarios

Theory without application ages quickly. Two short scenarios show how the Blueprint resolves in practice.

Scenario one — The specialist consultant

A former in-house SEO lead is starting an AI Visibility practice solo. She has one testimonial from a previous employer, a good network in the B2B SaaS category, and no full-time capacity.

Working through the Blueprint, her positioning becomes: "For Series A and Series B B2B SaaS companies whose founders have started to ask why AI systems describe their competitors more clearly than them." The positioning refuses e-commerce, refuses enterprise, and refuses tactical buyers.

Her promised outcome becomes: "Within ninety days, your company is described accurately and favourably in AI-generated answers to the category questions your buyers actually ask, and your team knows how to maintain that."

Her deliverables collapse to three: a Representation Baseline (how AI currently describes the company), a Category Answer Map (the questions the agency will optimise for), and a Maintenance Playbook (how the client's own marketing team keeps it moving after month three).

Her cadence is a ninety-day initial engagement, followed by an optional monthly review retainer. The monthly review is deliberately small — one hour of decision-making, one page of change — because her capacity is finite and she would rather be honest about that than promise a volume of work she cannot deliver.

Her price is a single fixed fee for the initial engagement and a single monthly retainer for the review. Both are set at the top of the range her network would find defensible for a specialist. The package reads as considered because it has been.

Scenario two — The generalist agency's first vertical

A three-person digital agency wants to add AI Visibility as its second service line. Its existing work is across categories; its first AI Visibility package will not be.

Positioning is chosen deliberately narrower than the agency itself: "For independent professional services firms — accountants, consultancies, law firms — whose partners have realised their next best clients are now asking AI for recommendations."

The promised outcome names a change the partners will recognise: "Within one hundred and twenty days, your firm is accurately represented in AI-generated recommendations for the services your partners most want to win, and you have a monthly rhythm for keeping it that way."

Deliverables are three: a Partner Recommendation Audit, a Firm Representation Rewrite (across the firm's own site and the third- party sources AI systems draw from), and a Monthly Representation Review.

Cadence is a four-month initial engagement plus a monthly review retainer. The monthly cadence is chosen because professional services partners renew based on visible discipline, not on volume of activity.

Price is deliberately positioned at the top of the professional services range the agency's existing clients would recognise. The firm can charge it because the previous four layers have earned it, not because the number is impressive on its own.

Two agencies. Two packages. Same Blueprint.

Common Failure Modes

A first package tends to fail in one of four familiar ways. Each failure corresponds to a specific layer of the Blueprint.

The Wide Positioning Failure. The offer describes the discipline rather than a buyer. The agency then attracts enquiries from every direction and cannot serve any of them well. Fix by rewriting Layer 1 until at least three plausible buyers would recognise that the offer is not for them.

The Activity-as-Outcome Failure. The promised outcome describes what the agency will do, not what will change for the buyer. Symptom: clients ask "so what will we actually see?" three months in. Fix by rewriting Layer 2 against the observable, transferable and honesty tests.

The Everything-Deliverable Failure. The package lists twelve deliverables, each written in practitioner language. Symptom: the sales call is spent explaining the deliverables rather than the outcome. Fix by cutting the list to three deliverables in the buyer's own language.

The Ad-Hoc Cadence Failure. The initial engagement is carefully designed and the ongoing rhythm is left as "as required". Symptom: clients drop off after month four because they cannot see what the retainer is for. Fix by designing Layer 4 with the same care as Layer 3.

If a first package is not converting, the failure is almost never at Layer 5. It is almost always at Layer 1 or Layer 2, and the price is simply carrying blame for a design decision that was missed further up the stack.

Testing the Package Before Publishing

A first package does not need a launch. It needs three quiet tests.

The first is the five-buyer test. Show the package — not the website, the offer itself — to five people who resemble the buyer you have positioned around. Ask each one to describe, in their own words, what the package is and who it is for. If four of the five describe it back accurately, Layer 1 and Layer 2 are working. If they do not, the package is not yet ready for a website.

The second is the ninety-second test. Time a non-technical reader as they read the offer page. If they cannot describe the outcome and the cadence after ninety seconds, the offer is too long, too jargon-heavy, or both. Cut until they can.

The third is the silence test. Send the package to one plausible buyer with a short, honest note asking whether it is a fit. Then stay silent for a week. What comes back — including the silences — tells the agency more about the offer than any amount of internal polishing.

Only after those three tests does the package deserve a permanent home on a website.

The Package as a Living Document

A first package is not a launch; it is a first version. The most useful first packages are revised roughly every quarter for the first year and then annually thereafter. Revision is not a sign of weakness. It is a sign that the agency has been listening.

Revisions usually flow upwards through the Blueprint. Something about the price fails, which prompts a rethink of cadence, which prompts a rethink of deliverables, which occasionally forces the agency to admit that the promised outcome was too generous, or the positioning too wide. That upward flow is healthy. The agency that resists it, and defends its first package as though it were a constitution, is the agency that eventually finds itself competing on price alone.

The Blueprint's job is to make revision easy. Because the layers are named and ordered, an agency changing its offer knows exactly which layer has moved, and which layers must adjust in response. The alternative — a rate card that gets edited every few months without anyone quite knowing why — is how new agencies drift into commodity pricing without noticing.

Methodology & Sources

This guide is compiled from the LaunchAnAEO Business Knowledge Engine and consumes the following canonical documents:

• fw-ai-visibility-package-blueprint — the framework introduced in this guide.
• mm-clients-buy-confidence-before-capability — the mental model that underwrites the Blueprint's design decisions.
• mm-clients-buy-outcomes — the mental model that separates a package from a rate card.
• gov-editorial-strategy and gov-editorial-style — the editorial standards this guide is written against.
• val-originality, val-cornerstone-editorial-quality and val-public-publication — the validators this guide has passed before publication.

Related frameworks consulted during preparation: fw-agency-value-ladder, fw-service-design-journey, fw-discovery-conversation-map.