“AI consulting” has become such a broad label that it covers everything from someone who teaches you to write better prompts to someone who redesigns how your whole team works. If you’re thinking of hiring one for marketing, you should know what’s reasonable to expect and what’s hype. Here’s what a serious consultant does, and what they don’t.
What an AI consultancy for marketing is (and isn’t)
It’s a judgment service, not a tools service. A consultant reviews how marketing gets done in your business today, identifies where AI saves time or improves results in a measurable way, and leaves you with a plan to put it into practice. The valuable part is knowing where it works and where it doesn’t.
It isn’t a subscription to the trendy app of the moment, a generic one-afternoon course, or a package of “automations” that gets installed the same way at every company. If the proposal you receive would be identical for a dental clinic and a clothing store, it isn’t consulting: it’s a product with a different name.
The first thing they review: your business, not the technology
A good consultancy doesn’t start by asking which tools you want to use. It starts with what already exists:
- How your customers find you today. Which channels bring real inquiries and which only bring visits.
- Where your team’s time goes. Which tasks repeat every week and how many hours they eat up.
- What content you already have. Copy, photos, emails, listings, answers to frequently asked questions: AI performs far better when it starts from your own material.
- How results are measured. Without a baseline number, nobody can say later whether anything improved.
- Who decides and who executes. An improvement that depends on a single person falls apart the day that person goes on vacation.
This phase is usually the most uncomfortable, because it almost always reveals that the problem wasn’t a lack of technology but a process with no owner or an unclear message.
What you get at the end
Here’s what should end up in your hands, concretely:
- A written diagnosis, in plain language, covering what was found.
- A prioritized list of opportunities: what to do first because it delivers more for less effort, and what to leave for later.
- An implementation plan with owners, steps, and realistic timelines.
- Measurement criteria: which number to watch, how often, and what kind of movement counts as success.
- Hands-on support at launch, so that what was agreed actually gets done and doesn’t stay in a nice-looking document.
Notice what’s missing: “a chatbot that answers everything” or “endless content.” Those things can be part of the plan, but a deliverable isn’t a tool that’s been switched on: it’s a well-made decision, put into action.
Where AI does help in marketing
There are tasks where the gain is real and shows up quickly:
- First drafts of copy, emails, descriptions, and scripts, which a person reviews and signs off on.
- Organizing and summarizing information: reviews, customer messages, long reports.
- Adapting the same message to several formats and languages.
- Answering repetitive questions quickly, leaving people the work that requires judgment.
- Preparing and keeping up to date routine work that used to be done by hand and kept getting put off.
The pattern is always the same: AI speeds up the mechanical work and leaves the judgment to the person. When that relationship is reversed, the problems begin.
What a serious consultant doesn’t do
This is the part that helps you choose well:
- They don’t promise rankings, sales, or exact numbers. Nobody controls a search engine’s algorithm or your customers’ buying decisions. They can talk about ranges and ballpark timelines, always with a caveat.
- They don’t replace your team overnight. If someone sells you “fire your agency and let AI do everything,” be wary.
- They don’t publish without human review. A text that’s generated and published as is usually sounds like everyone else’s, and factual errors go straight to the customer.
- They don’t hide what they don’t know. If something can’t be verified, they say so.
- They don’t tie you to a trend. Tools change every few months; what lasts is a well-thought-out process.
- They don’t use your information without explaining how. What’s shared, with whom, and for what purpose should be clear from the start.
How to tell a real consultancy from an empty promise
Before hiring, ask whoever is proposing the service these questions:
- What will you review about my business before proposing anything?
- What will I have in hand when we’re done, specifically?
- How will we know it worked, and from what baseline number?
- Which part is done by a person and which by AI?
- What happens when the support ends: can I keep going on my own?
- What would you recommend I not automate?
The last one is the best test. Someone who never tells you “don’t do this with AI” is probably selling you something else.
How long it takes to show
Treat these timelines as a guide, not a guarantee: internal productivity gains (fewer hours spent on repetitive tasks) usually show up within the first few weeks. The effects on lead generation and sales depend on your industry, your market, and how consistently you apply the plan, and they’re measured in months, not days. Anyone who promises commercial results in a week is selling you expectations, not a method.
Who it makes sense for and who it doesn’t, yet
It makes sense if you already have a business that sells, a small team short on time, and marketing tasks that repeat. It doesn’t yet if you still don’t know who you’re selling to or if there’s nobody who will put the recommendations into practice: in that case, the first step is to sort out your offer, and AI can wait without costing you anything.
A well-done consultancy sometimes ends with “it’s not right for you today.” That’s also a result, and it’s worth more than an implementation nobody is going to use.
If you want to know what we’d do with you
The most useful thing is to start with a no-obligation conversation about your case: what you sell, how customers find you, and where you feel your time goes. Send us a message with that, and we’ll tell you frankly whether AI can help you right now, where, and what isn’t worth touching. And if this article cleared something up for you, share it with someone who’s weighing the same decision.
This article describes a general approach to the work; it includes no results figures. The timelines mentioned are approximate and vary by business.
Versión en español: Qué hace un consultor de IA para marketing (y qué no)
