A survey by Arctera, released July 21 with Hanover Research, polled 500 compliance professionals across the Americas and Europe, in sectors like finance, healthcare, and energy. The result exposes a gap many companies haven't yet noticed: 78% expect AI-related communications risk to increase over the next 12 to 24 months, but fewer than one in five (19%) can actually prove their AI governance is working.
What Changed
According to the report, 55% of companies already have governance basics in place — policies, training, review steps. But only 19% have the evidence layer needed: logging, retention, detection, and risk scoring that let them prove, after the fact, what AI produced, who reviewed it, where it went, and whether the record was kept.
Why It Matters
Having an AI policy on paper isn't the same as being able to demonstrate to a regulator, an auditor, or a customer that the policy was actually followed in a specific decision. Most companies today can say "we have rules" — but can't show the trail that proves the rules worked in a specific interaction.
The Impact for Brazil
Brazilian companies in regulated sectors — banks, insurers, healthcare — advancing AI adoption for customer service, credit analysis, or client communication carry the same risk the survey identified: writing the policy is the easy part. What actually protects a company in an audit or a legal dispute is having an auditable log of every relevant AI-generated decision — and that's rarely in an AI project's initial budget.
Entercast's Take
This data confirms what we've argued in recent posts: AI governance isn't a document, it's infrastructure. A company that invests in policy without investing in audit capability is building a defense that won't survive the first real challenge. Before scaling AI use in processes that touch customers or regulated decisions, it's worth asking: if a regulator asked to prove what happened in a specific interaction six months ago, could the company answer?