In the first 90 days, an AI-first marketing leader maps the team's workflows, finds the slowest repetitive tasks, and automates one or two with AI before touching anything else. Days 1 to 30 are for auditing and quick wins, 31 to 60 for building and testing workflows, and 61 to 90 for measuring results and training the team to run them.
In this blog post I'm going to walk you through the practical AI playbook for the first 90 days as a marketing leader (CMO, VP Marketing, Head of Growth), the version with the actual milestones and the specific decisions you need to make. Not the version that tells you to "build an AI roadmap" without saying how.
Most "first 90 days" playbooks for marketing leaders predate AI. Most "AI for marketing" content ignores the unique constraints of being new in role. Together this leaves new marketing leaders with two incompatible workstreams to manage.
This article fixes that. The 90-day plan below assumes you're new to a marketing leadership role AND you want AI to be part of your operating model, without sacrificing the credibility-building work you need to do as a new leader.
I've been a marketing consultant for twenty-one years. I went all in on AI in 2024. I've helped about a dozen newly-appointed marketing leaders (CMOs, VPs Marketing, Heads of Growth) build their first 90 days around AI. The patterns below repeat.
By the end of this blog you'll have the week-by-week playbook, the milestones, the specific AI investments to make in your first 90 days, and the AI investments to delay until day 91+.
TL;DR
Days 1-30: Listen, audit, baseline. Use AI for accelerating learning. Don't build anything customer-facing yet.
Days 31-60: Identify the 1-2 highest-use AI workflows. Build them with strict scope. Communicate the strategy.
Days 61-90: Ship the workflows. Measure impact. Plan day 91+.
What NOT to do in first 90 days: full AI marketing transformation. Mass AI content. Sweeping team changes based on AI capability.
Why the first 90 days is the wrong time for big AI bets
Three reasons new marketing leaders should be cautious with AI in their first 90 days:
1. You don't know what you don't know. You've inherited tools, workflows, and conventions you haven't fully understood. Big AI bets based on incomplete diagnosis become expensive corrections later.
2. Your team is reading you carefully. Your first AI decision sends signals about your style, your priorities, your judgement. Make those signals deliberate.
3. Credibility compounds. The leader who ships 2 careful AI wins in 90 days has more authority for the bigger AI bets in months 4-12 than the leader who tried everything at once.
The playbook below balances these constraints.
Days 1-7: Pure listening
Goals: Understand the current state. Don't change anything yet.
Activities:
- 1:1s with every direct report (60 minutes each)
- 1:1s with every cross-functional peer (45 minutes each)
- Read every existing marketing document (brand, strategy, brief library, reporting)
- Review every active campaign
- Sit in on at least 3 customer-facing conversations (sales calls, customer interviews, support tickets)
AI use: Allowed. Use AI for synthesis of what you're hearing. "I've had 8 1:1s, here are notes from each, what patterns do you see?" Saves 4-6 hours of synthesis time.
Don't: Announce any changes. Don't sketch any AI initiatives publicly. Listen.
Days 8-21: Quiet diagnosis
Goals: Build your private hypothesis about what's broken, what's working, and where AI fits.
Activities:
- Run the AI marketing audit framework on the business, privately
- Map the current marketing tech stack (every tool, what it costs, who uses it)
- Document the current marketing workflows end-to-end
- Identify the 5 biggest bottlenecks (volume? quality? strategy? operations? relationships?)
- Identify the 3 biggest current bets and their performance
- Get raw data access, full historical performance, not summarised dashboards
AI use: Heavy. Synthesise audit data, propose hypotheses, run scenario analysis. AI-accelerated diagnosis fits your timeline.
Don't: Share your diagnosis externally yet. Continue listening.
Days 22-35: First public communication
Goals: Share what you've learned. Validate hypothesis with stakeholders. Set expectations for what's coming.
Activities:
- Present diagnosis to your CEO/CRO (45 minutes, candid)
- Present diagnosis to your direct reports (90 minutes, candid)
- Present diagnosis to peers (30 minutes, slightly more polished)
- Define your operating principles publicly (what you'll prioritise, what you won't)
- Announce no big changes yet, but tell people why
AI use: Help with presentation prep. Don't AI-generate the actual communication. The voice matters; it's yours.
Don't: Promise specific AI initiatives yet. Don't commit to deliverables you haven't scoped.
Days 36-50: Pick your two
Goals: Identify the 1-2 highest-use AI workflows for the next 60 days. Scope them tightly.
Criteria for selection:
- Addresses a real bottleneck (not a "we should have AI" want)
- Outcome can be measured within 90 days
- Doesn't require restructuring teams yet
- Internal team supports it
- Budget approval feasible
Typical first AI workflows by business type:
- B2B SaaS: AI lead enrichment + qualification (workflow 1+2 from my B2B lead gen guide)
- Service business: AI sales call prep + follow-up (workflows 1+4 from my sales workflows guide)
- E-commerce: AI product description engine
- Content-led business: AI content production workflow (per my 80/20 content guide)
AI use: Heavy. Scope the workflows. Get vendor quotes. Build the business case.
Don't: Pick 5 workflows. Don't promise transformation. Pick 1-2 and ship them well.
Days 51-60: Build the case + sign off
Goals: Get formal approval. Set up measurement infrastructure. Recruit internal champions.
Activities:
- Present the 1-2 chosen workflows with: cost, timeline, expected impact, measurement plan
- Get CFO/CEO sign-off on budget
- Identify the internal team member who'll own each workflow long-term
- Set up baseline metrics for the chosen workflows (you need pre-AI data to measure improvement)
- Choose your build approach: consultant, in-house, hybrid
AI use: Help with the business case writeup. Help with measurement infrastructure setup.
Don't: Skip the baseline metrics. Without pre-AI baseline, you can't prove impact.
Days 61-75: Build
Goals: Ship the workflows. Don't add scope.
Activities:
- Weekly check-ins with the build team
- Stay close to implementation, don't delegate fully and disappear
- Resist the temptation to add scope ("while we're at it, could we also...")
- Communicate progress to your team and stakeholders
AI use: As the build progresses, use AI for whatever the workflows need.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Don't: Add scope mid-build. The first 90 days are about a win, not a transformation.
Days 76-90: Ship + measure + plan day 91+
Goals: Workflows in production. First measurement data. Plan for next quarter.
Activities:
- Production deployment of the 1-2 workflows
- 2-week parallel run (AI + manual) to validate quality
- First measurement vs baseline
- Communicate results internally, be honest about what didn't work
- Sketch the day 91-180 plan based on what you learned
AI use: Measurement, reporting, analysis.
Don't: Declare premature victory. Don't promise the next quarter's bigger transformation yet, you don't have the data.
What NOT to do in your first 90 days
Five specific traps:
Trap 1: Replace existing AI tools that work
You'll inherit some AI tools. Some are good. Some are bad. Don't change anything in the first 30 days. Sometimes you'll find tools that look mediocre but serve a real purpose. Sometimes you'll find expensive tools nobody uses.
The right move: audit, decide in days 31-60, change in days 61-90 only if necessary.
Trap 2: Mass AI content rollout
The temptation: "I'll show impact fast by doubling content output with AI."
The reality: Google penalises this. By month 4-6 you've made things worse, just as everyone is evaluating your impact.
Trap 3: AI for outbound at scale
The temptation: "We'll generate pipeline fast with AI-powered outreach."
The reality: brand damage by month 3. Senior leaders in your industry remember.
Trap 4: Team restructuring based on AI capability
The temptation: "Now that AI does X, we don't need person Y."
The reality: AI workflows are fragile in production. Restructuring before AI is stable causes pain.
Trap 5: Promising transformation in 90 days
The temptation: "I'll deliver our AI marketing transformation in my first 90 days."
The reality: 90 days isn't enough for transformation. Promising it sets you up to fail or to fake it.
How to communicate AI in your first 90 days
Three communication frames that work:
Frame 1: "Tools, not transformation"
Position AI as specific tools you're evaluating, not as a strategic direction. This buys you time to diagnose before committing.
Frame 2: "Quick wins first"
Position the 1-2 workflows as quick wins to build momentum. Set expectations for what's coming after.
Frame 3: "Measurement before scale"
Position the first 90 days as proving the value before scaling. CFOs love this. Boards love this.
Avoid: "AI-first organisation." "Digital transformation." "Disruption." These signals sound impressive in interviews but invite scepticism once you're in the role.
What to learn in your first 90 days
Beyond the playbook, what to internalise:
- Which AI vendors you'd never buy from. Build a personal blacklist.
- Which AI vendors you'd recommend. Build a personal shortlist.
- Which tools your team relies on most. Defend those if budget cuts come.
- What your team is good at. Don't AI-replace their best work.
- What your CEO believes about AI. Their public statements vs private views may differ.
- What your board expects. AI commitments at board level often have specific timelines.
- Where your company's data is messy. Inform future AI roadmap.
Frequently asked questions
What if my CEO wants AI transformation in 90 days? Push back. Promise diagnosis + first wins in 90 days. Transformation in 12 months. If CEO won't accept that, you have a bigger problem than AI.
What if my predecessor already started an AI project? Audit it carefully. If it's going well, continue and credit them. If it's failing, decide whether to fix or kill in days 31-60.
Should I hire an AI marketing person in my first 90 days? Probably not. Wait until day 91+ when you understand what you need.
What if my team resists AI? Listen first. Resistance often points to legitimate concerns (job security, quality risks). Address those before pushing AI.
Can AI help me write my first board update? Yes for structure. No for content. The content needs to be yours.
What's the single best thing to ship in 90 days? A measurable improvement in one specific metric, achieved through a small AI workflow your team is proud of. That builds credibility for everything that comes next.
Want a first-90-days plan tailored to your situation?
If the conclusion is "wait, hire X first," that's the conclusion. You walk away with the diagnosis and no further pitch.
I'm Lilach Bullock. I've been a marketing consultant for twenty-one years. I went all in on AI in 2024. I work with founders and marketing leaders who want AI to move their numbers, not just their tool stack.
Related: running a guest posting campaign
Want the complete version? Read where I break down AI marketing.
