The Shortlist Stack: a B2B SaaS marketing framework for AI startups
The Shortlist Stack is the four-layer framework I use for B2B SaaS marketing, built for a time when buyers decide which tools to evaluate before they ever talk to sales, often inside ChatGPT, Google and their LinkedIn feed. The layers are Story, Surfaces, System and Signal. Each answers one question, and they work in that order.
Why build marketing around the shortlist?
Most B2B deals are decided by the shortlist a buyer builds alone. If your product is on it, you get a fair hearing. If it isn’t, you rarely hear the deal existed. More and more of that shortlist is now written by AI assistants, so being clearly described across the places they read has become part of marketing.
95%
of the time, B2B buyers choose a vendor from their day-one shortlist. 6sense, 2025
51%
of B2B software buyers now start research in an AI chatbot more often than in Google. G2, 2026
The four layers
Layer 1
Story
Positioning a buyer and an AI model can repeat.
Most AI products get described by what the model does. Buyers shortlist by the problem it removes. Story finds the sentence that connects the two, tests it against how customers already talk, and makes it the spine of the site, the sales pitch and the founder’s posts. When the story is clear, AI assistants repeat it too.
What I do
Interviews with customers and lost deals
Category and competitor mapping
Positioning and messaging workshop with the founders
What you get
Positioning statement
Messaging guide for site, sales and social
The founder’s point-of-view pillars
How to tell it’s working: Can a buyer, and ChatGPT, describe what you do in one sentence that matches yours?
For support leads at B2B SaaS companies who lose hours to repetitive tickets, Acme is the AI agent that resolves them inside your help desk.
Who it’s forThe problemWhy now
Layer 2
Surfaces
Be present where buyers and AI assistants look.
B2B buyers do most of their research before they speak to a vendor: in LinkedIn feeds, newsletters, communities, review sites, comparison pages and, more and more, in ChatGPT and Perplexity. Surfaces picks the two or three places that matter for your buyer, plus the pages AI answers cite for your category, and makes sure you appear there with something worth reading.
What I do
Buyer research: where your best customers learned about you
An audit of your site, LinkedIn and AI answers for your category
A focused channel plan and citation source map
What you get
Channel map
AI search visibility baseline
A 90-day plan for two or three channels
How to tell it’s working: How often you are named for the questions buyers ask, and where.
A rhythm the founder can keep through launches, hiring and fundraising.
Marketing compounds only when it keeps going. System builds a publishing and launch rhythm around the time your founder really has: an idea bank, a drafting workflow, a calendar tied to the product roadmap, and a launch plan for the moments that matter.
What I do
Founder content workflow and idea bank
Launch and release planning
Templates your team can run without me
What you get
Publishing cadence and calendar
Content workflow and templates
Launch playbook
How to tell it’s working: Weeks in a row the founder published, and launches shipped on plan.
Measure what turns into pipeline, then double down.
Every month, Signal looks at what brought demo requests, which pages buyers read before a call, what they say about where they heard of you, and what AI assistants now say about your product. Then we cut what isn’t working and put more into what is.
What I do
Attribution setup, including a “How did you hear about us?” field
A monthly review with the founders
AI answer tracking for your key buyer questions
What you get
Pipeline attribution dashboard
Monthly review and recommendations
A plan to double down
How to tell it’s working: Demo requests by source, including AI search, month over month.
Each layer depends on the one before it. More channels can’t fix an unclear story. A good story with no rhythm behind it fades after one launch. And without measurement, there’s no way to know which surface deserves more time. Most startups are strong on one layer and missing another, so an engagement starts with the first layer that fails the self-check.