Where AI pays back in your marketing, and where it does not.

Five things: a readiness assessment across data, process, skills and governance; a ranked shortlist of use cases scored on value against effort; two or three of them run as measured pilots with baselines and kill criteria; a written roadmap for the next two to four quarters with a tool recommendation including cancellations; and the governance layer, meaning a usage policy, a named owner and role-specific enablement. Four to six weeks, fixed scope, all files editable and yours.

That is the first thing we test, and the answer is usually partly. Content and research use cases need very little. Predictive analytics, lead scoring and real time personalisation need a clean data model, consistent definitions and a working join between marketing activity and revenue. 62% of organizations already report losing revenue directly because of poor CRM data quality, and a model built on that repeats the errors faster. When the data is the blocker we say so and scope marketing operations consulting first rather than selling you a roadmap you cannot run.

A fixed fee, quoted after a scoping call, with deliverables and dates written down before you commit. The number moves with team size, how many use cases you want piloted and whether governance has to satisfy a compliance function, so publishing a rate would mislead most readers. For budget context, Gartner reports CMOs allocating a mean 15.3% of marketing budget to AI, rising to 21.3% among organisations with mature AI processes. Book a meeting for a scope and a number.

Not in this engagement, and for most marketing teams that is the right answer. The value is in choosing use cases, sequencing them, measuring them honestly and governing them, and the majority are served by tools you can already buy plus workflow design. Where a pilot proves that a bespoke build genuinely pays back, we specify it, help you brief it and review the vendor's work, and we take no fee from whoever builds it.

Whichever ones fit the use cases that survived scoring, and the recommendation is written after the pilots rather than before them. We hold no reseller agreements, take no referral fees and mark up no licences, so the advice is free to say that the tool your team likes is not worth its seat cost, or that two of your four subscriptions overlap. Consolidating and cancelling is a common outcome and often the fastest return in the whole project.

By taking a baseline before each pilot and agreeing the success threshold in advance, on measures that matter to the business: production time, cost per asset, cycle time, conversion or pipeline where the volume supports it, plus a quality assessment by your own reviewers. Vanity outputs, such as volume of content produced, are excluded deliberately. Context worth keeping in view: only 49% of marketing and finance leaders measure how marketing is driving business outcomes at all, so the measurement habit often outlasts the pilots as the more valuable change.

The baseline is included: we record how answer engines and AI overviews currently describe, cite or ignore your brand across a set of prompts your buyers would actually use, and keep the prompts so the test can be repeated. Improving that position is a separate discipline and it is real work on your content, entity data and citations, delivered through our GEO and AEO service alongside SEO. Nobody can guarantee a mention in an AI answer, and anyone who does is selling something.

Our experience is that it changes the mix rather than the headcount, and we do not take redundancy work. The roles that gain value are the ones that judge output, own data and manage specialists; the roles defined purely by throughput change most. Anxiety is real and worth naming: 71% of professionals believe AI will eliminate more jobs than it creates over three years, and 48% describe their own sentiment toward AI as neutral, negative or unsure. Where the honest answer is a structural one, marketing team structure advisory is the engagement that addresses it properly.

The usage policy is explicit about it: which categories of information may never be entered into a general purpose tool, which tools are approved for what, where data is processed and retained, what needs a data processing agreement, and what has to be disclosed to customers or clients. Pilots run on either synthetic or properly permissioned data. Where your sector adds obligations we are not qualified on, we say so and recommend your counsel reviews the policy before it is adopted.

A short interim policy first, the full one after. Teams are already using these tools, so a one page interim rule on what may be entered and what must be reviewed protects you during the engagement, and the durable policy is written once the pilots show which workflows are actually in use. Just 13% of organizations have all four governance foundations in place, an AI council, a roadmap, generative AI policies and an ethics policy, and 39% have a council, so most companies are starting from nothing here.

Less time than you would expect. We need the marketing team members who do the work, whoever owns data and reporting, the person accountable for brand, the sales lead where AI touches lead handling, and someone senior enough to approve the policy. Expect a kickoff, two working sessions, a review per pilot and a final presentation. Pilots do consume some of your team's attention by design, because the point is to test the tools inside your real workflow.

Marketing strategy consulting decides where to compete, what to say and which channels to back. This engagement is narrower: given your strategy, where does AI change the economics of executing it. If you are unsure the strategy itself is right, start there, or with a marketing audit if the question is what is currently working. An AI roadmap bolted onto an unclear strategy just industrialises the wrong activity.

Only if you ask, and it is quoted separately. This engagement deliberately excludes execution so the roadmap is free to recommend fewer tools, a slower pace or no further spend. Plenty of clients take the roadmap, the policy and the enablement pack and run everything with their own team, which is a good outcome. Where you want ongoing marketing leadership rather than a project, a fractional CMO is the engagement to look at.

Usually three things were missing: a baseline, a decision date and an owner. Pilots without those drift until someone loses interest, which reads as failure but was never a test. We also sequence deliberately, so the first pilots are the ones your data can support, and the harder use cases wait until the constraint is fixed. The pattern is common: only 19% of organizations describe their AI progress as accelerating, while 46% call it inconsistent or siloed.

More than most buyers expect, and it is the cheaper half. Only 49.7% of companies say their marketing teams have the skills and training to use their own marketing systems, down from 54.1% in 2022, and CMOs rank a lack of internal talent as their top barrier to AI-driven efficiency, first for 19% and in the top three for 38%. So the enablement pack is role-specific, built around the workflows we piloted with your own material, and kept short enough that people actually use it.

You own the roadmap, the policy, the prompts and the visibility baseline, and your named owner runs the next quarter. Most clients book one check-in ninety days later to review what was adopted, what was quietly abandoned and what the next two use cases should be, which takes half a day and is optional. Nothing in the engagement obliges you to buy anything from us afterwards, including the execution work.