How this track works
Six sessions, all thinking mode, no code - built for the person who decides, not the person who prompts. Tonight sets the landscape; then a2 scores the bets, a3 protects the brand, a4 builds the team, a5 measures honestly, and a6 sequences the roadmap. Every session ends the same way: a short list of questions to ask your team or agency, so you leave with something to say on Monday, not just something to know. Two rails run through all six, and we name them tonight.
The state of play: past the hype and the fear 7 min live
Two stories have followed AI into every marketing team: the hype story ("it replaces the department") and the fear story ("it makes everything slop, or steals our jobs"). Both are wrong in the same way - they treat AI as a single magic button. The reality in 2026 is quieter and more useful: AI is a co-pilot across the whole funnel, strong in some stages, still supervised in all of them.
LiveThe co-pilot, not the button - AI across the whole funnel3 min▶
Think of AI as a very fast junior who never tires and never checks their facts. That description is unglamorous and exactly right. It changes how you lead: you are not buying a machine that does marketing, you are giving every marketer a co-pilot and then deciding where that co-pilot is trusted to steer.
- It helps everywhere, replaces nowhere yet. Research, positioning, content, SEO and GEO, ads, email, social, analytics - AI adds leverage in each, and needs a human owner in each.
- The strength varies by stage. It is strongest at drafting and ideation, useful for analysis and personalization, and still supervised for anything customer-facing or automated.
- Agentic planning is arriving. Harvard DCE's 2026 material now teaches agentic campaign systems - AI that plans and sequences campaigns - as a governed future, not a turnkey product. We cover it properly in a6.
Self-studyWhat is real vs vaporware in 20262 min read▶
The fastest way to lose credibility with your team is to repeat a vendor demo as if it ships tomorrow. Keep a private list of what is genuinely production-ready versus what is a promising demo.
| Real today | Demo / vaporware (be sceptical) |
|---|---|
| Drafting, ideation, and rewriting at volume | The "fully autonomous CMO" that runs strategy unattended |
| Analysis and summarising your own data + research | "Set it and forget it" campaigns with no human review |
| Personalization from consented first-party data | Perfect, hallucination-free facts with no sourcing |
| Supervised agentic workflows with guardrails | End-to-end autonomy sold as ready-now |
The three eras: assistive, augmented, agentic 6 min live
A simple ladder gives your whole team a shared vocabulary and stops the "are we behind?" anxiety. There are three eras, and the key insight is that capability and governance climb together - you do not earn the next rung by buying a tool, you earn it by proving you can control the last one.
LiveWhere most teams actually are - and how to climb3 min▶
Naming the era honestly is a leadership act. It replaces vague pressure ("we need more AI") with a clear question ("what would it take to earn the next rung safely?").
- Era 1 - assistive: AI drafts, ideates, and rewrites; a human owns every output. This is where value starts and where slop risk is highest if you skip the review habit.
- Era 2 - augmented: AI becomes a co-pilot for analysis and personalization - reading your data, tailoring messages to consented segments - still with a human in the loop.
- Era 3 - agentic: AI plans and sequences campaigns with explicit guardrails, budgets, and sign-off gates. Emerging, not turnkey. Do not skip to it to look modern.
- Most teams are in eras 1-2. That is not behind. Being deliberately in era 2 with strong governance beats a chaotic leap into era 3.
The leader's real job: decide and govern, not prompt 5 min live
Here is the reframe that makes this track worth your evening. Your job is not to become the best prompter in the room - your team will out-prompt you within a month. Your job is to decide where AI is worth the risk, protect what only you can protect, and build the conditions where AI creates leverage instead of liability.
LiveFive decisions only the leader can make3 min▶
Everything in the remaining five sessions maps to one of these five decisions. Prompting is a skill your team owns; these are yours.
- Pick the bets (a2): score use-cases on value and risk, and choose the first one deliberately rather than chasing every demo.
- Protect the brand (a3): own the voice, the claims, and disclosure - the things that cannot be delegated to a model.
- Build the team (a4): roles, human-in-the-loop review, and the martech stack that supports it.
- Measure honestly (a5): efficiency versus effectiveness, and resisting vanity metrics dressed up as ROI.
- Sequence the roadmap (a6): the order you adopt in, so trust and governance grow ahead of autonomy.
LiveThe two rails you will govern to - stated plainly2 min▶
Two rails run through every later session. They are simple to say and they are where credibility lives or dies.
- Rail 1 - own the brand and the claim. AI drafts; a human owns every word, image, and claim that ships. No fabricated stats, no fake reviews or testimonials, no unverifiable comparisons - and clear disclosure of AI-generated endorsements and material connections where the law requires it (FTC in the US, ASA in the UK). We go deep in a3.
- Rail 2 - protect the data. Personalize only with consented first-party data, respect GDPR and PDPA, and never let customer PII or unreleased plans go into a consumer AI tool. Great targeting that breaks privacy law is not a win.
The shortcut that became a headline. Regulators have already fined brands for AI-generated and incentivized reviews presented as genuine, and for "AI-washing" their own capabilities. The pattern is always the same: fast, cheap, and exactly the shortcut AI makes tempting. The rails exist so the one thing AI cannot manufacture for you - trust - stays intact.
Map your funnel's time-sinks ◐ 15 min · bring to a2
No code, no tools - just your own funnel and a clear head. This is the raw material you will score into a shortlist next session.
- List your marketing funnel stages: research, positioning, content, SEO and GEO, ads, email, social, analytics. Use your real channel names.
- Mark the top three time-sinks - the stages where your team spends hours on repetitive, drafting, or first-pass work.
- For each, note in one line what an AI co-pilot could realistically help with, and one thing that would still need a human owner.
- Flag which of the three touches customer data or public claims - those carry the two rails and will score higher on risk in a2.
- Bring this one-page map to session a2, where you will score each candidate on value and risk and pick your first bet.
Official sources covered
Taught from public AI-marketing curricula and vendor documentation. Certificates and paid tooling stay with the official programs. This session covers:
Three questions before you go 🎯 ◐ 90 seconds
1 · A vendor pitches a "fully autonomous CMO" that runs your strategy unattended. What is the honest 2026 read?
In 2026 AI genuinely helps in every funnel stage, but the fully autonomous CMO is vaporware. Real deployments keep a human owner, especially anything customer-facing or automated.
2 · Which era are most marketing teams honestly in right now?
Most teams draft and ideate (era 1) and are moving into co-piloted analysis and personalization (era 2). That is not behind. Agentic (era 3) is emerging and governed, not the norm.
3 · What is the leader's real job in an AI-first marketing team?
Your team will out-prompt you fast. The leader decides where AI is worth the risk, owns the brand and claims, builds the conditions for safe use, and sequences the roadmap. That is the whole track.