​Read online​

Welcome to Unpacking Meaning. If you received this from a friend and enjoy it, ​subscribe here.

Stop selling AI. Start selling Tuesday afternoons.

Here’s the number that should terrify every AI SaaS founder right now: 88%.

That’s how many companies McKinsey surveyed say they’re using AI. And only six percent — 6 — see any meaningful financial value from it. The median company has deployed it, but they haven’t gotten anything out of it.

The gap we’re seeing is not between AI or no AI. It’s perception. Lots of companies are not seeing any value from AI, hence get more and more skeptical of what it’ll do fore them. Which for us marketers, also means we need to be able to communicate AI’s value well enough in this rough landscape.

Elad Gil put his finger on it on the Tim Ferriss Show recently. His point: the problem for AI adoption isn’t how good is the AI. It’s how much do your people have to change the way they work to actually use it?

The issue is almost always change management. Not the tech.

This is the thing the AI SaaS industry does not want you to understand, because it ruins the product pitch.

The adoption crisis nobody talks about

S&P Global found that 42% of AI initiatives get abandoned before they ever reach production. Companies fund them, build them, announce them internally — and then let them die because they struggle to adapt their workflows or the status quo they’re used to.

Microsoft Copilot is the perfect case study. Microsoft deployed Copilot across 450 million commercial seats. By any measure, the most aggressive enterprise AI rollout in history. And the adoption rate came in around a measly three percent. They discount forty to sixty percent on every deal to close it! Which tells you everything about whether people actually want it.

Now compare that to GitHub Copilot with twenty million developers:

• ​81% install on day one​
• No change management required
• No rollout
• No training sessions

It works inside the IDE, where the developer already lives. And it works because it doesn’t ask the workflow to change.

The difference between three percent and eighty-one percent is whether you build to be part of their workflow or demand that they change it all to fit you. And, of course, how you communicate it.

The moat that isn’t a moat

The uncomfortable truth about building a product in 2026 is that whatever you’re running today, someone else will run cheaper, faster, or both within 18 months.

Your advantage is how you’re cutting the switching cost (real and perceived) that comes from embedding your tool into how someone does their job.

The thing is, the switching cost is not your contract, your data, or your integrations anymore, but it’s more operational. More tied to their day-to-day work lives. You know you’re winning at it, when your customer’s team quickly reorganizes how they work around your tool, and when you ask them what they’d do if your tool stopped working today, they’d feel lost.

This is why traditional SaaS companies ran seventy-two percent gross revenue retention, and AI-native companies are running forty percent.

The old model for switching costs got wiped out. The new one has to be earned one “Tuesday afternoon” at a time.

But understanding whether you’ve achieved that — really understanding it — takes more than a pulse survey. It means getting in there: interviewing customers, going deep into their day-to-day lives, understanding the psychology behind their decisions and how their workflow actually flows. And then you have to be able to communicate it, both internally, so your team knows what they’re building toward, and externally, so prospects can see exactly why leaving would cost them.

How to actually do it: the five-framework playbook for earning operational switching costs

1. The Interview Framework — job stories, not surveys

Exit interviews and NPS scores tell you what already happened. Customer interviews tell you why — and whether it will happen again. Run them like an investigative reporter, not a marketer.

The opening move: Tell me about the last time you almost stopped using us. Then shut up and listen.

Follow with:

• What almost made you leave? What was the thing that nearly broke it?
• What specifically stopped you? What was the alternative you considered, and what made it lose?
• Walk me through a Tuesday. What does your afternoon actually look like now?
• When is the last time you recommended us to someone — and what did you say exactly?

What you’re building: a story, the narrative of why they stayed.

2. The Switching Cost Audit — map where you live in their workflow

You can’t defend what you can’t see. Weave in a live switching cost audit with each customer interview:

For each tool or process your product touches, ask: If a competitor offered this tomorrow — same features, lower price — what would it take for you to actually switch?

You’re looking for three types of switching costs:

• Data switching costs: Do they have history with you that would be lost? Reports, training data, outputs, a record of decisions made?
• Workflow switching costs: Have they built their process around your workflow? Do they organize their day around how you work?
• Relational switching costs: Do they know your team by name? Does institutional knowledge live in that relationship?

If you’re only scoring on integrations and contract length, you’re missing the real moat — whether they’ve reorganized their “Tuesday afternoon” around you.

3. Internal Communication Playbook — make churn cost visible to your team

Your sales team doesn’t feel churn. Your engineers don’t feel churn. They feel feature requests and pipeline reviews. The churn is abstract until it isn’t — until the ARR drops and everyone scrambles.

Make the switching cost real for every internal stakeholder:

• For your sales team: Stop reporting churn as “one logo left.” Report it as: This customer represented $47K ARR. They had three active integrations. Two of their team members had been using us for 18 months. They left because we weren’t in their specific XYZ workflow — we were a one-time check dashboard. That’s a different kind of loss. It changes how your team thinks about the deals they’re chasing.
• For your product team: Run the “Switching Cost Audit “ quarterly on your top accounts. Publish the answers internally. If a customer says “I’d feel lost without you,” that’s not just a retention metric but a product roadmap signal too. It tells you what not to break.
• For your executive team: Translate churn into compound terms. One customer leaving at $40K ARR doesn’t just cost $40K. It costs the expansion revenue they would have contributed over 24 months, plus the acquisition cost to replace them, plus the credibility hit when references dry up. Run the full math, not just the ARR line.

4. External Communication Playbook — show prospects why leaving would cost them

Here’s the mistake most SaaS companies make: they try to prove their value with feature lists. “We have 47 integrations! We’re SOC2 compliant! We have 24/7 support!”

Prospects tune that out in the first slide.

Instead, lead with the switching cost story:

Some of our customers almost left in the first 90 days. Here’s why they stayed — and why they’re still here 18 months later.

Then show the specific shape of what they’d lose:

• What their workflow looks like with us versus what it would look like without us
• The specific process they’d have to rebuild if they switched
• What 18 months of history with us actually gets them

This reframes the conversation. From competing on features to competing on what it would cost them to leave — and the answer, if you’ve done the work right, is: a lot.

The question you should be asking

Ultimately my point today is, stop asking whether your AI is better than the next one. Start asking how you can get into your prospects’ “Tuesday afternoon”.

That’s what makes an AI-powered product actually defensible. Not the model. Not the integrations. Not the contract. Whether your customer has easily built their day around the way your tool works — and whether leaving would break something they actually care about.

DISCOVERY

Marketing in the age of AI

I sat down with Emanuel Rose to talk about what separates AI-generated copy that sounds like everyone else from copy that actually converts. Generic AI output is built for volume, not for your specific market — and the moment you let a machine write in generic SaaS language, you’re racing to the bottom.

→ Listen to the episode​

Learn to Trust Yourself Again

After years, I decided to pick back up my personal Substack Negative Capability and published a new essay. In it I tie it the chaos of running my team of agents to what AI struggles to do that makes us truly still valuable. Our ability to remain in contemplation and reflection while letting opportunities emerge.

This is my corner of the internet to think in public, learn and connect with like-minded curious humans. I’d love it if you decided to follow.

→ Check it out here​

What It’s Like to Be a Writer

Jasmine Sun wrote a super honest take on her first year as a full-time writer. I loved the three-question exit criteria she set for herself at the start of the year and what those answers actually looked like by December. It’s a good reminder that the professional uncertainty we all feel isn’t a bug, but actually a useful signal to inform us on how to move forward.

→ Read at jasmi.news​

Teach Your Agents to Manage Up

Hilary Gridley wrote a prompt you can give your AI agent right now that will make it fundamentally more useful. Agents fail the way junior employees fail — they either run ahead with bad assumptions or they flood you with open-ended questions. Useful for managing both agents and humans.

→ Read at hils.substack.com​

RESONANCE

Claude Hopkins, Scientific Advertising

Have a great weekend!

Cheers,

Chris

Chris Silvestri Founder & conversion alchemist

🙌🏻 Let’s be friends (unless you’re a stalker)​

When you're ready, here's a few ways I can help

Not sure where to start? Take our free message-market fit scorecard.​