Skills a Marketing Team Needs for AI Search: An Eight-Week Plan and a Learning System That Keeps Pace

The six skills a marketing team needs for AI search, how to assess them, an eight-week plan using the team's own website, who owns what, and how to separate the knowledge that lasts from the part that changes every month.

A marketing team that works well in AI search needs six skills: understanding how search engines and AI assistants find and choose sources, writing answer-first content backed by evidence, enough technical literacy to spot crawling and rendering problems, measuring visibility across Search Console, analytics and AI answer tests, earning mentions through PR and listings, and working with subject experts. Most teams already have some of these. An eight-week plan, a few hours a week, can close the gaps.

The six skills

Skill What it looks like in practice
1. How AI search works Can explain retrieval, query fan-out and why passages matter, and knows which crawlers each assistant uses
2. Answer-first writing with evidence Writes sections that open with a direct answer, include specifics and name sources
3. Technical literacy Can read robots.txt, view page source, run a crawler check and brief developers clearly
4. Measurement Uses Search Console, GA4 channel groups and repeated AI answer tests; reports share of voice honestly
5. Mentions and listings Manages review profiles, pitches editors, keeps company details consistent
6. Working with experts Interviews subject experts and turns their knowledge into accurate content they approve

Assess the team first

Ask each person to rate themselves from 1 to 4 on the six skills, then check with a short practical task per skill: explain why a page is not cited, rewrite a weak paragraph, find a crawler block, build a share-of-voice calculation, draft a pitch, turn an expert interview into a section. The gaps between self-rating and task results show where training is needed.

The eight-week plan

Week Topic Reading on this site Practical task
1 How AI search works How AI search works; GEO vs SEO Explain to the team how one of your pages could be cited
2 Crawlers and access AI crawler access checker Run the checker on five of your pages
3 Answer-first writing 10 rewrites; quotability Rewrite three sections of a service page
4 Evidence and expertise E-E-A-T; first-party evidence Interview an expert and write one section they approve
5 Questions and topics Query fan-out research Map sub-questions for one service
6 Measurement Tracking mentions; dashboard Run the question set and calculate share of voice
7 Mentions and listings review sites; digital PR Audit your listings and draft one pitch
8 Reporting and planning reporting to leadership; 90-day plan Present a one-page report and next quarter’s plan

Plan about three hours a week: one hour of reading, two hours on the task using your own site.

Make every week end in something shipped

Training that finishes with a certificate and no change to the website has not worked. The reason the plan above uses your own pages rather than worked examples is that the output is the proof: at the end of each week something should be live, or a number should be written down.

A simple rule keeps it honest. A week is not complete until one of these exists:

  • A page on your site that reads differently from how it read on Monday
  • A fix shipped by a developer, with the before and after recorded
  • A row added to the measurement sheet, dated
  • A pitch sent, or a listing claimed and corrected

Teams that skip this step tend to know the vocabulary and own no results, which is also the hardest position to defend when the budget is reviewed.

Who owns what, and what each role learns first

Nobody needs all six skills. Splitting them by role is faster than training everyone in everything, and it makes the monthly work survive one person being busy.

Role What they own What to learn first
Content lead The pages themselves, and the working relationship with subject experts Answer-first writing, evidence, and how to interview an expert (skills 2 and 6)
Measurement owner The question set, the monthly run and the report that goes upward Search Console, analytics channel groups and repeated answer tests (skill 4)
Technical contact, in-house or agency robots.txt, rendering, schema and indexing fixes How retrieval works and how to read a crawler check (skills 1 and 3)
Senior sponsor Press relationships, listings, and sign-off on any claim the company publishes Mentions and listings, and what makes a claim defensible (skill 5)

In a two or three person team one person covers several roles. What matters is that the measurement owner is a named individual, because shared ownership of a monthly task is how the monthly task stops happening.

Separate the stable knowledge from the volatile

The reason a one-time course stops being enough is not that the subject is hard. It is that half of it barely changes and the other half changes every month, and teams that treat all of it as urgent end up reading announcements instead of working.

Stable, learn once Volatile, check monthly
How retrieval works, and why a passage rather than a page is what gets used Which crawlers each AI company runs, and what they are currently named
Answer-first writing, and why evidence beats adjectives Which assistants show citations, and in what format
Reading robots.txt, page source and a render check Answer features rolling out, and which markets they reach
Keeping company facts identical everywhere Which review sites and directories assistants are favouring
Measuring on a fixed question set over time What the tools can automate this quarter

The left column is what the eight-week plan teaches, and it will still be true next year. The right column needs a routine, not a course.

The monthly routine

One named person reads the official sources once a month: Google Search Central, OpenAI’s and Anthropic’s crawler documentation, Perplexity’s documentation and Bing’s webmaster blog. They circulate a short note covering only what changes how the team works, and ignore the rest. Vendor blog posts and conference takes are not a substitute for the primary documentation.

The quarterly re-check

Once a quarter, re-run the short practical task for each of the six skills and compare with the previous quarter. This catches the two failure modes that a monthly news summary hides: a skill that was never really learned, and a skill that has quietly become one person’s private knowledge.

Frequently asked questions

Do marketers need to learn to code?

No. They need to read page source, understand robots.txt and brief developers precisely. The technical fixes can stay with developers.

Should we hire a GEO specialist?

For many teams, training existing SEO and content staff is faster. A specialist or agency helps when there is no one to own measurement and technical work. See briefing an agency for GEO work.

Is there formal training?

Brilad runs AEO and GEO training workshops for in-house teams, built on the same material.

How long before the team works without prompting?

The eight-week plan covers the working knowledge at about three hours a week. Fluency comes from running the measurement cycle a few times rather than from the reading, so expect roughly a quarter before the monthly routine happens without being chased.

Who should own this in the company?

One named person, usually whoever already owns SEO or content. The specific thing that needs single ownership is the monthly measurement run, because that is the task that silently stops when it belongs to everyone.

Is a one-time course enough?

It covers the stable half well. The volatile half, crawler names, citation formats and which listings matter, moves faster than any course is revised, which is what the monthly routine and quarterly re-check are for.

Can a two-person team do this?

Yes, with less breadth. One person takes measurement and the other takes content, technical work goes to a developer or agency with a named contact, and the question set shrinks to the ten questions that matter most. Consistency matters more than coverage.

Sources

  1. Google Search Central, “AI features and your website”
  2. OpenAI, “Overview of OpenAI crawlers”