Your team is the real AI strategy
It is tempting to think the AI question is "which tools do we buy?" The harder and more important question is "what do our people need to become good at, and how do we help them get there without fear or chaos?" The tools are commodities - your team's judgement, taste, and speed at using them is the moat. This session is about building that moat deliberately: naming the roles that are shifting, the skills worth investing in, and the change management that decides whether AI adoption feels like a threat or an upgrade.
The roles are shifting 13 min live
Nobody's title has to change tomorrow, but the center of gravity of the work does. The grind of production moves to the machine; the value of the marketer moves up the stack toward strategy, judgement, and taste. Here is the shape of that shift.
LiveFour roles, one shift up the stack7 min▶
These are not four new hires - they are four kinds of work that need an owner, and often one person wears more than one hat.
- The AI-literate marketer. Spends less time producing first drafts and more time on strategy, briefs, and judgement. The output per person goes up; the skill that matters shifts from typing to thinking and editing.
- The brand guardian / editor. The person who owns voice, claims, and final quality. In an AI world this role gets more important, not less - someone has to be the taste and the truth check on volume.
- Marketing ops + prompt libraries. The connective tissue: shared prompts, approved tools, and workflows so the whole team benefits from what works instead of each person reinventing it.
- The creative / agency partner. Specialist craft and surge capacity. AI raises the floor on production, so partners shift toward the high-craft and high-scale work that still needs a human specialist - and toward being accountable for disclosure and rights.
Self-studyWhat does not change - and why that is the point2 min read▶
It is easy to over-rotate on what AI changes and miss what it makes more valuable.
- Deciding what is worth making is still a human strategy call - AI cannot tell you what matters to your market.
- Judging whether the work is good is taste, and taste is the scarce skill when everyone can produce infinite drafts.
- Owning the claim and the brand stays with a person, always - the buck cannot stop with a model.
The skills to build 11 min live
Four skills carry most of the value. The good news for a leader: three of them are teachable to people you already have, and the most important one is a judgement skill your best marketers already have the raw material for.
LiveThe four skills, and how to build each6 min▶
| Skill | How to build it |
|---|---|
| Editing + brand judgement (the #1 skill) | Practice on real AI drafts - critique, cut, and rewrite to voice. This is coaching and reps, not a course; pair newer people with your best editor. |
| Prompt literacy | Short internal workshops plus a shared prompt library. Learn by adapting prompts that already work for your brand, not from scratch. |
| Claim-verification discipline | A habit, not a class: every stat gets a source before it ships. Build it into the review checklist until it is automatic. |
| Data / analytics fluency | Enough to read results and question AI output honestly. Lightweight training plus making numbers part of every campaign retro. |
Notice that editing and brand judgement sits at the top. When production is free, the marketer who can tell shippable from slop - and fix the gap fast - is worth more than one who can produce more of it.
Self-studyReskill vs hire - a leader's rule of thumb2 min read▶
Most of this is a reskilling problem, not a hiring one - which is cheaper, faster, and better for morale than assuming you need new people.
- Reskill for prompt literacy, claim discipline, and editing judgement - your existing marketers already know your brand and market, which is the hard part.
- Hire or partner for genuine capability gaps - deep analytics, specialist creative, or marketing-ops leadership you do not have at all.
- Beware "hire an AI person" as a reflex. A single AI hire rarely fixes a team-wide skill gap; distributed literacy usually beats one specialist in a silo.
Change management + operating model 15 min live
Skills do not spread on their own. The difference between a team that adopts AI well and one that either fears it or misuses it is almost entirely change leadership - and then a clear decision about who owns what.
LiveFear to curiosity - the moves that work6 min▶
Fear comes from two places: "will this replace me?" and "what if I get it wrong?" Good change leadership answers both directly.
- Name champions. Find the early adopters, give them air cover, and let peers learn from peers - it lands better than a top-down mandate.
- Build safe sandboxes. A space to experiment with non-confidential data, where getting it wrong is learning, not a career risk. Fear of mistakes is what keeps people from trying.
- Chase quick wins. One visible time-saving win beats a strategy deck. Let the team feel the leverage before you ask them to change how they work.
- Make brand-review a badge of honor. Tie AI use to the quality bar from session a3, so "used AI well" means "passed brand review", not "cut a corner".
LiveCoE vs embedded - who owns what5 min▶
There are two ways to organize AI capability, and the right answer is usually a blend that shifts over time.
- Center of Excellence (CoE). A small central team owns tools, prompt libraries, training, and standards. Strong for consistency and governance early on; risks becoming a bottleneck if everything routes through it.
- Embedded. AI capability lives inside each function - content, social, ads, email. Strong for speed and ownership; risks fragmentation and inconsistent brand safety without a shared spine.
- The usual path. Start CoE to set standards and tools, then embed the capability into teams as literacy spreads - keeping a light central owner for tools, brand voice, and training.
| What needs an owner | Who typically owns it |
|---|---|
| Approved tools + data safety | Central (CoE) - one clear list, one standard |
| Brand-voice guide + prompt library | Central to define, embedded to use |
| Training + reskilling | Central to run, managers to reinforce |
| Claim verification + sign-off | Embedded in each function, per session a3 |
Team skills-map + two change moves ◐ 30-40 min
Turn the session into a plan you can start this quarter. Map the skills honestly first, then pick just two moves - two done beats ten planned.
- Build the skills-map. List your team down one side and the four skills across the top (editing + brand judgement, prompt literacy, claim-verification discipline, data fluency). Score each person green / amber / red. Patterns will jump out.
- Mark reskill vs hire. For each red column, decide: can we coach our way there, or is this a genuine hire-or-partner gap?
- Pick your #1 skill owner. Name the person strongest on editing and brand judgement, and give them a coaching role for the quarter.
- Choose two change moves. From Part 3, pick exactly two to start this quarter - for example, a non-confidential sandbox plus a named set of champions. Write who owns each and what "done" looks like.
Official sources covered
Taught from public AI-marketing skilling and enablement curricula. Certificates and paid tooling stay with the official programs. This session covers:
Three questions before you go 🎯 ◐ 90 seconds
1 · Which individual skill matters most on an AI-ready marketing team?
When production is nearly free, taste is the scarce skill. The marketer who can judge whether AI output is good and correct it to brand voice is worth more than one who simply produces more of it.
2 · What is the strongest set of change moves to get a team adopting AI well?
Fear of replacement and fear of mistakes are what block adoption. Sandboxes make experimenting safe, quick wins prove the leverage, and champions let peers learn from peers - together they turn fear into curiosity.
3 · What is the core tradeoff between a Center of Excellence and an embedded model?
A CoE centralizes tools and standards for consistency but can become a chokepoint; embedding capability into each function is faster and more owned but can fragment brand safety. Most teams start CoE and embed as literacy spreads.