Agency vs. consulting: the real distinction

#The distinction, stated in mechanism terms, not marketing terms

AI Agency draws a clean line between its "niched-vertical implementation" sub-path and its "standalone paid AI-readiness audit" sub-path: one builds a working system and is accountable for whether it works; the other diagnoses and recommends, and is not accountable for whether the client's team actually implements the recommendation well. That's not a difference in prestige or sophistication — it's a difference in who carries execution risk, and it is the exact same fork that separates a marketing agency from a marketing consultant, including the fractional-CMO version of consulting this course's Module 4 covers in depth.

An agency is paid to produce a result through its own delivery. It writes the copy, builds the campaign, sends the email, manages the ad account. If the campaign underperforms, that's the agency's problem to fix, inside the scope the client already paid for — the client's exposure is bounded by the contract, but the agency's own margin absorbs the cost of getting it wrong. A consultant is paid to produce a recommendation. The strategy document, the channel plan, the positioning framework, the 90-day roadmap — once delivered, the engagement's core deliverable exists, and whether the client's own team executes it well, badly, or not at all is now the client's risk, not the consultant's. This is the same distinction Marketing's own start-here lesson draws between marketing and sales at a different seam ("marketing generates and qualifies demand; sales converts it") — a real functional boundary, not a hierarchy, and confusing it costs a founder months the same way confusing marketing and sales does.

#Why this is a genuinely different business, not a lighter version of the same one

The practical consequences run in every direction that matters for actually building the business:

Neither carries less real skill than the other — a genuinely useful strategy document requires as much expertise as running a genuinely good campaign, arguably more, since a consultant has no execution feedback loop correcting a bad strategic call the way a live campaign's own performance data corrects a bad targeting choice. What differs is the business model, not the competence required to run it well.

#Why a founder has to pick one, on purpose

The instinct to do both — "I'll consult, and if they want execution too, I'll just do that as well" — is the single most common way this specific business gets muddled before it has a chance to work, and it's worth naming here, at the mechanism level, before Module 4 names the failure mode it produces with a source. The moment a consulting engagement starts absorbing execution work, the pricing that made sense for advisory-only work (a flat retainer priced against a few hours a week of strategic judgment) stops covering the actual hours going into the relationship, because execution work is priced and staffed completely differently — that's the entire reason Module 3's client-acquisition lesson and Module 4's rate data land on such different numbers per hour of the founder's own time. A business that's structurally unclear about which of the two it's selling ends up pricing like a consultant and staffing like an agency, which is a margin problem with a name: it's covered directly, with a real source, in Where advisory-only fails.

#Where the broader-mix-agency thread (Module 3) and the fractional-CMO thread (Module 4) actually connect

They're not fully separate businesses in practice — a mature operator frequently runs both, sequentially or in parallel, and this course treats that as the normal path rather than an edge case. A common, credible progression: start as a consultant (lower capital, no team to manage, immediate proof of judgment), and once a client relationship and a track record exist, either bring on delivery capability and convert the relationship into agency work, or deliberately hold the line at advisory-only and specialize as a fractional CMO long-term. The reverse also happens — an agency operator who's tired of the labor-margin structure and the client-concentration risk in Module 5 repositions as advisory-only, trading a higher revenue ceiling for near-100% margin and a smaller number of clients they can actually serve well. Module 3 and Module 4 are written to be read independently, in whichever order matches which side of this fork you're actually evaluating, but neither is a smaller or easier version of the other — they're different answers to the same client-need, and the right one depends on whether you want to build a delivery team or not, not on which one sounds more prestigious.