On this page14
- Which content types matter most, ranked by funnel stage?
- The four building blocks of the plan
- Pillar pages
- Comparison pages
- Founder point-of-view posts
- Distribution
- The 90-day plan, week by week
- Who does what each week
- How to run the research weeks
- A content brief template for every piece
- The repurposing system
- Distribution: the half most teams skip
- What to measure at day 90
- When to get help
90-Day Content Marketing Plan for an AI Startup (Template)
A 90-day content marketing plan for an AI startup should produce three things by day 90: two or three pillar pages that own your core category terms, four to six comparison and alternative pages that catch buyers mid-decision, and a steady weekly rhythm of founder point-of-view posts that build trust. Month one is research and foundations, month two is publishing the bottom-of-funnel pages first, and month three is distribution and repurposing. Plan for roughly 25 to 35 pieces of content in total, most of them repurposed from about ten core assets.
Most AI startups I talk to publish the wrong things in the wrong order. They start with top-of-funnel thought pieces because they're fun to write, and they publish the comparison page their sales team begs for six months later. Flip it. Write the content closest to revenue first.
What follows is the template I'd hand a seed or Series A AI team with one marketer, a founder who can give two hours a week, and no agency.
Which content types matter most, ranked by funnel stage?
Not every format earns its keep at the early stage. This is how I rank content types for an AI startup with limited hours, ordered by how directly they influence a buying decision.
| Rank | Content type | Funnel stage | Why it matters |
|---|---|---|---|
| 1 | Comparison and alternative pages ("X vs Y," "X alternatives") | Bottom | Buyers search these right before choosing |
| 2 | Use case and integration pages | Bottom to middle | Shows exactly how your product fits their stack |
| 3 | Case studies and worked examples | Bottom to middle | Proof, in their language |
| 4 | Pillar pages on your core category | Middle | Owns the terms buyers and AI assistants learn from |
| 5 | Founder point-of-view posts | Middle to top | Builds trust and a recognisable voice |
| 6 | Original data or benchmarks | Top, with high earned-media value | Gets cited by journalists, analysts and AI search |
| 7 | How-to tutorials and templates | Top | Drives traffic and sign-ups when tied to your product |
| 8 | Generic industry news commentary | Top | Lowest priority; easy to skip |
If your hours are tight, do ranks 1 to 5 well and skip the rest until month four.
The four building blocks of the plan
Pillar pages
A pillar page is a long, definitive guide to a core topic you want to be known for. For an AI evaluation platform, that might be "LLM evaluation: the complete guide." For an AI security tool, "Prompt injection: how it works and how to prevent it." Each pillar links out to smaller supporting posts and receives links back from them.
Aim for two or three pillars in 90 days, not ten. Each should run 2,500 to 4,000 words, include original diagrams or tables, and get updated quarterly.
Comparison pages
Comparison pages catch buyers when they've narrowed the field. Write them honestly. Name where a competitor is better. Buyers can tell a fair comparison from a hit piece, and AI assistants increasingly favour the balanced one when they summarise.
Typical set for an AI startup:
- Your product vs the incumbent everyone knows
- Your product vs the closest venture-backed competitor
- Your product vs the open-source option
- Your product vs building it in-house
- "Top [category] tools" roundup where you appear honestly
Founder point-of-view posts
These are short, opinionated posts in the founder's voice: what they believe about the category that others don't, what they learned shipping the product, where the market is wrong. They're the content that gets shared, quoted and picked up by reporters. One a week is the target. Even 600 words counts.
Distribution
Publishing isn't distribution. Every core piece needs a distribution plan: newsletter, founder LinkedIn and X, relevant communities, partner newsletters, and a pitch to a reporter or podcast if the piece has a real point of view.
The 90-day plan, week by week
| Week | Focus | Deliverables |
|---|---|---|
| 1 | Research | Interview 5 customers or prospects; list 30 questions they ask; pull search data for your category |
| 2 | Positioning and keywords | Choose 2 to 3 pillar topics and 4 to 6 comparison targets; map keywords to pages |
| 3 | Foundations | Set up blog structure, author page, analytics, newsletter sign-up; write the content brief template |
| 4 | First bottom-funnel pages | Publish 2 comparison pages; founder POV post 1 |
| 5 | Bottom-funnel pages | Publish 2 more comparison or alternative pages; founder POV post 2 |
| 6 | Use cases | Publish 2 use case or integration pages; founder POV post 3 |
| 7 | Pillar 1 | Publish pillar page 1; founder POV post 4; first newsletter issue |
| 8 | Proof | Publish case study 1 (or a labelled worked example); founder POV post 5 |
| 9 | Pillar 2 | Publish pillar page 2; founder POV post 6; second newsletter issue |
| 10 | Original data | Publish a small benchmark or survey result; pitch it to 5 to 10 relevant reporters |
| 11 | Repurposing sprint | Turn pillars and data into 10+ social posts, 2 short videos, 1 slide deck, 1 guest post pitch |
| 12 | Pillar 3 or refresh | Publish pillar page 3 or update the first comparisons with new info; founder POV post 7 |
| 13 | Review | Measure, prune, and plan the next 90 days |
That's roughly 7 founder posts, 4 to 6 comparison pages, 2 use case pages, 2 to 3 pillars, 1 case study, 1 data piece and 3 newsletter issues before repurposing. With repurposing it lands around 30 pieces.
Who does what each week
The plan only holds if the time is protected. A realistic weekly split for a lean team:
| Role | Weekly time | Owns |
|---|---|---|
| Content marketer | 20 to 30 hours | Briefs, drafting, editing, publishing, distribution, reporting |
| Founder | About 2 hours | One 45-minute recorded interview for the POV post, final review of pillars and data |
| Product or engineering lead | About 1 hour | Technical accuracy on pillars, use case and integration pages |
| Sales or customer success | About 30 minutes | Feeding in buyer questions and objections, flagging what they use |
The founder interview is the engine. Record it, transcribe it, and pull the POV post, two social posts and a newsletter paragraph from the same conversation. Founders who write from scratch every week usually stop by week five. Founders who talk for 45 minutes usually don't.
How to run the research weeks
Weeks one and two decide whether the next ten weeks work. Skip them and you'll write a polished blog nobody needs.
- Interview at least 5 customers, prospects or lost deals about how they found and evaluated tools
- Collect the exact questions they asked during the sales process
- Pull search data for your category, competitor names and problem phrases
- Ask ChatGPT, Perplexity and Claude the questions your buyers ask, and note which sources they cite
- List your top 5 competitors and what content ranks for their brand plus "alternative"
- Pick the 3 claims you want to be known for, in one sentence each
If you haven't nailed positioning yet, the positioning statement template for AI startups is a good week-two exercise.
A content brief template for every piece
Every piece in the plan gets a brief before anyone writes. It takes 15 minutes and saves a full rewrite.
Title (working):
Target query or question:
Funnel stage: top / middle / bottom
Reader: role, company type, what they already know
The one thing they should believe after reading:
Proof we can use: data, customer quote, example, demo
Internal links: pillar page, product page, related posts
Distribution: newsletter, founder social, community, pitch target
Owner and due date:The repurposing system
Ten strong core pieces can feed a whole quarter. The trick is deciding up front how each core piece gets broken down.
| Core asset | Repurposed into | Channel |
|---|---|---|
| Pillar page | 5 to 8 social posts, 1 carousel, 1 newsletter section | LinkedIn, X, newsletter |
| Comparison page | 1 sales one-pager, 2 social posts, 1 FAQ block | Sales team, LinkedIn |
| Founder POV post | 1 LinkedIn post, 1 X thread, 1 podcast talking point | Founder channels, podcast pitches |
| Original data | 1 press pitch, 3 to 5 chart posts, 1 slide deck, 1 webinar | Media, LinkedIn, events |
| Case study | 1 short video, 1 sales email, 2 social posts | YouTube, outbound, LinkedIn |
| Customer interview | 3 quote graphics, 1 FAQ answer, 1 newsletter story | Social, site, email |
A simple rule: never publish a core piece without listing its five repurposed outputs in the brief. If you can't think of five, the core piece is probably too thin.
Distribution: the half most teams skip
A useful benchmark is to spend about as much time distributing a piece as you spent writing it. For each core piece:
- Send it to your newsletter with a short personal note from the founder
- Post a native version on LinkedIn and X (not just a link)
- Share in 2 to 3 relevant communities where it's genuinely useful and allowed
- Send it directly to 5 to 10 customers or prospects it would help
- Pitch the strongest pieces to a reporter, newsletter author or podcast host
- Add internal links from at least 3 older pages
- Brief the sales team so they can use it in follow-ups
Original data and strong founder opinions are where content and PR overlap. A benchmark nobody else has is the easiest kind of story to pitch, and earned coverage of that data sends links and AI-search citations back to the page. The guide to getting cited by ChatGPT and Perplexity explains why that matters more every quarter.
What to measure at day 90
Content compounds slowly, so 90-day metrics are early signals, not verdicts.
| Metric | Healthy early signal | Where to look |
|---|---|---|
| Comparison page conversions | Some demo or trial starts attributed to these pages | Analytics plus CRM |
| Organic impressions | Steady week-over-week growth | Google Search Console |
| AI assistant mentions | Your brand appears for 1 or more buyer questions | Manual checks in ChatGPT, Perplexity, Claude |
| Newsletter growth | Consistent net new subscribers each week | Email tool |
| Sales usage | Reps share content in deals | Ask them; track link clicks |
| Earned pickups | At least 1 reporter, newsletter or podcast references your data or founder | Media monitoring |
Prune anything that got no traffic, no sales use and no links after 90 days. Update it or fold it into a pillar.
When to get help
The plan above is designed to run with one marketer and a founder. The part that usually breaks is founder time. The POV posts and data pieces need the founder's thinking, and that's the first thing to slip when a fundraise or a big customer lands.
If that's your bottleneck, a ghostwriter who can interview the founder for 45 minutes and turn it into a publishable piece keeps the rhythm going. That's the work behind my content writing service, including placed founder essays in tier-1 contributor outlets. But plenty of teams run this well in-house once the brief template and interview habit are in place.
For the first month of founder posts, start with the free 30-day founder content calendar, which gives you a prompt for every day. The companion post on the 30-day founder content calendar explains how to use it.
Ninety days won't make you the category's default source. It will tell you which three pieces your buyers actually read, and that's the list you build the next year on.
Want your 90-day plan reviewed against your category and buyer? Book a 30-minute teardown.