01
The shift from social to AI
In 2018, your brand voice was on Twitter. In 2026, it's an LLM replying to your customers at 3am. The listening problem is the same. The channel is new.
2026 Whitepaper
Listening to Your AI.
Your reputation is still your currency. In 2018 that meant listening to customers on social. In 2026 it means listening to the AI speaking on your behalf — to every reply, summary, approval, and handoff it ships before a human ever reviews it.
What's inside
Four answers you'll take back to your team.
02
Six signals that predict a bad AI decision
Concrete patterns we've seen stop problems before they reach customers. Tone drift, policy slippage, citation collapse, and three more.
03
A decision ledger auditors will actually accept
What to capture for every AI decision so that, a year later, you can answer: who decided, why, against which policy, with what evidence.
04
A 30/60/90 plan to start listening this quarter
Pragmatic sequencing for teams that can't freeze AI rollout but can't keep flying blind either.
The 2018 original comes with it.
We first wrote about listening in 2018, when the channel was social media and the risk was a tweet gone wrong. Every reader of the 2026 edition gets the original PDF too — useful context for how the thesis traveled, and why the principles held up even as the channel changed.