How to become an AI marketing consultant (without the hype)

Woman working at a desk in a cozy home office.

There is no exam, no license, and no certificate that makes a client hire you as an AI marketing consultant. Clients hire judgment. Specifically, judgment about which parts of their marketing AI should touch and which parts it should never go near. Right now, most people selling AI marketing consulting have the tool fluency and none of the judgment, and that gap is where the real opportunity sits.

Lilach Bullock, an AI implementation consultant who has been doing this work with paying clients, wrote up the straight version of how this path actually works. What follows is the operator-relevant core of it.

Why the crowded end of the market is not worth entering

In the past two years, thousands of people with no marketing background completed a weekend prompt engineering course and started pitching themselves as AI marketing consultants. Some charge the equivalent of £500 a day to tell a small business owner to use ChatGPT for social captions, which the business owner could have figured out in an afternoon on their own.

That end of the market is already saturated and getting cheaper fast. Prompt writing is not a scarce skill anymore. GPT-5 and Claude produce a decent first draft without hand-holding.

What is not saturated is knowing which parts of a marketing function AI should touch and which parts it should stay away from. AI can quietly wreck brand voice, customer trust, or email deliverability if left unsupervised. Knowing that distinction, and being willing to act on it in front of a client, is the actual job.

What clients are actually paying for

Not prompts. Judgment, structure, and accountability. A client paying £800 to £1,500 a day (the realistic UK range for someone with a genuine track record) wants three things:

  • An audit of where AI can save real hours without degrading quality
  • A system they can run without you once the engagement ends, built in tools they already use like HubSpot, Klaviyo, or ClickUp
  • Someone who will say “don’t automate that” when the tempting answer is to automate everything

That last point matters more than most consultants admit. Crisis communications, real complaint handling, and a founder’s voice on a sensitive announcement should never be run through a model first. Clients who hire you expecting you to automate everything are testing whether you’ll tell them the truth or just take the fee.

laptop computer on glass-top table

A real client engagement, with real numbers

Bullock worked with a 12-person B2B services firm doing roughly £1.8m in revenue. Their marketing manager had left and the founder’s instinct was to buy every AI tool available and let it fill the gap.

Within a week, the engagement established two concrete wins: case study drafting time dropped from roughly six hours to ninety minutes, and social content repurposing fell from a full day a week to about two hours. Those are real, measurable, worth paying for.

But the sales-qualified lead follow-up emails stayed entirely human. A draft the tool generated had quietly overpromised a delivery timeline nobody could hit. Catching that single mistake likely saved a client relationship, and it is exactly why that firm kept Bullock on retainer for four more months at £1,000 a day. The tool handled the volume work. The consultant handled the judgment work. That split is the whole business model.

️ The six-step path if you’re starting from scratch in 2026

Bullock recommends resisting the urge to skip steps, because the shortcut version is exactly what’s flooding the market with consultants clients don’t trust.

  1. Get 2 to 3 years of hands-on marketing experience first. In-house or agency, ideally across content, email, and paid. Without this, you’re a tool demonstrator, not a consultant, and clients can tell within one conversation.
  2. Spend 60 to 90 days going deep on 4 to 5 tools maximum rather than shallow on twenty. One for content (Jasper or Copy.ai), one for research and SEO (Surfer or Clearscope), one for meeting and call notes (Otter or Fireflies), and one automation layer (Zapier or Make). Learn their limits, not just their features.
  3. Run three unpaid or heavily discounted pilot projects for businesses you already know, and document the before-and-after numbers: hours saved, output volume, response rates, anything concrete.
  4. Write up one of those pilots as a proper case study with real figures. This becomes your entire sales pitch for the next twelve months.
  5. Set your day rate based on what similar consultants charge for similar seniority, not what you feel comfortable asking. Undercharging in year one trains every future client to expect a discount.
  6. Pitch five specific businesses a week rather than posting generically and hoping. Cold outreach, specific, referencing something real you noticed about their marketing.

If you have no marketing background at all, don’t skip straight to consulting. Build the underlying skills first. Consultants who never did the operational work tend to give advice that sounds smart and breaks in practice.

3D rendered ai text on dark digital background

️ Pricing without underselling yourself

Day rates in the UK run from about £400 for someone one year in, up to £1,200 to £1,500 for someone with a strong client list and case studies to show. In the US, figures fall between $500 and $2,000 a day depending on region and specialism. Retainers tend to work better for both sides once you’re past your first year, typically £2,500 to £6,000 a month for ongoing strategic oversight plus quarterly audits.

Beginners often price at £150 to £250 a day because it feels safer to win the client. The problem is that clients read price as a signal of confidence and competence. A low rate can actively cost you the pitch rather than win it. Bullock reports seeing this happen twice with people she mentored: they dropped their rate to get a foot in the door, the prospect went quiet assuming low price meant low skill, and raising the number got the reply.

Building a portfolio before you have clients

Turn every pilot project, paid or not, into a one-page case study with a before number, an after number, and one sentence on what you deliberately chose not to automate. That last line does more for your credibility than any certificate, because it shows restraint, and restraint is rare in this field right now.

Skills you need before you use the title

None of these appear on a certificate. All of them show up in a client’s second meeting with you.

  • Solid grounding in core marketing: funnels, positioning, copy, and channel strategy, not just tool operation
  • Comfort reading basic analytics including GA4, email open and click rates, and ad platform reporting
  • Working knowledge of data privacy basics, since AI tools and customer data collide constantly and clients will ask
  • The ability to say no to a client’s request when it’s the wrong use of AI, and explain why clearly
  • Enough writing skill to spot when an AI draft is flat, generic, or factually wrong, because you will be the last check every single time

Common questions

Do you need a certification?

No formal certification is required or widely recognised. Courses from HubSpot Academy or Google can build useful vocabulary, but clients hire based on case studies and demonstrated results, not badges.

How much can you earn in 2026?

UK day rates typically run £400 to £1,500 depending on experience and track record, with monthly retainers of £2,500 to £6,000 for ongoing work. In the US, day rates commonly fall between $500 and $2,000.

What’s the biggest mistake new AI marketing consultants make?

Selling prompt writing as if it’s a rare skill. It isn’t anymore. The valuable thing is judgment: knowing precisely which tasks to automate and which to keep entirely human, and being willing to tell a client “don’t automate this” even when it costs you part of the fee.

Stay on top of AI & Automation with BizStack Newsletter
BizStack  —  Entrepreneur’s Business Stack
Logo