Getting cited by AI when your competitors are the incumbents.

The challenge

Hemut is a YC-backed modern TMS, built by founders who lived the problem. On Google they ranked well. In AI search they were invisible. Ask Perplexity for the best TMS for mid-sized carriers and you got Rose Rocket, McLeod, Trimble, and Samsara. Those incumbents appeared in 67 to 71% of answers. Hemut appeared in 8%, usually in a list with no reason attached. The better product was missing from the systems that shape early buying decisions.

What we found

We audited 47 real carrier queries across four AI engines. The problem was not content quality. It was citation logic. AI does not rank on Google's authority signals. It cites sources that clearly and consistently explain a problem, which is the core idea behind Answer Engine Optimization. Hemut's content was written for search engines, not for how AI decides what to quote. Three query clusters should have been theirs and were not: comparison, problem-focused, and buying-stage.

What we did

Three moves over 11 weeks. Most of it was repositioning, not new content. The story was already there. We made it legible to the right systems.

1. 1

   Built useful content on the exact problems Hemut solves: dispatcher overload, enterprise software that overkills mid-sized carriers, and profitability tracking that works. Written to cite them, not to sell.

2. 2

   Restructured their founding story, YC profile, and founder interviews into the formats AI treats as authority, instead of one isolated blog post.

3. 3

   Created workflow content showing how their AI agents solve real coordination problems, built to be cited for specific capabilities.

The outcome

In four months, Hemut went from 8% to 23% of AI answers. The context changed too. They stopped showing up in generic lists and started getting named as the solution when buyers asked about eliminating check calls or tracking profit per load. AI traffic grew 180% and converted at 4.7%, versus 1.9% from traditional search. These were later-stage buyers who already knew they needed a change. MQLs climbed from 4 to 9 a month and held, because visibility was spread across many systems, not one publication or algorithm.