Is workflow embedding a moat against AI?
Weakened, not gone: AI makes migration cheaper, but the workflow tax of re-learning and re-integrating still slows customers down.
Weakened by AI
Updated November 19, 2026.
What is it?
Your product is woven into how customers work: their processes, integrations, training and habits all assume it. Leaving means rebuilding all of that.
Does it hold against AI?
Weakened. Hockenmaier still counts switching costs as one of four compounding moats, because the workflow tax remains. But Lemkin notes AI lowers switching costs across the board: agents migrate data, map fields and rebuild integrations far faster than people did.
Encoded business logic on its own no longer holds. Bustamante's point is blunt: custom workflows can now be written down as a skill file.
How do you know if you have it?
Would a customer need weeks of effort and retraining to leave, even with AI helping? Is the embedding in your logic (copyable), or in their people, processes and data (stickier)?
How do you deepen it?
Embed where the work is irreducibly coordinated across people and systems. Pair embedding with a moat that holds, such as data effects or being the system of record.
Questions people ask
- Are switching costs still a moat?
- Smaller ones. Count on the workflow tax to slow churn, not to stop a determined customer.
Sources
- How to tell if a startup has a moat (Lenny's Newsletter): https://www.lennysnewsletter.com/p/how-to-tell-if-a-startup-has-a-moat
- What's your moat? (SaaStr): https://www.saastr.com/whats-your-moat/
- Ten vertical-software moats under LLMs (Bustamante, summary): https://hq-raydata-co.pages.dev/vault/06-reference/2026-02-16-vertical-software-selloff/
Does your business have a moat?
Five AI attackers, live web research and a published rubric. Free, about five minutes, no account. Or talk it through with Matt on a strategy call.