Fipps: Standard Chartered hits 77% AI deflection

Fipps's remarks at the 2026 AIQ Summit frame two sides of the same shift. On one side is the productivity case: Standard Chartered's chief operating officer told him she is driving 77% deflection rates across all of her employee requests, well above the 25% to 30% he calls game-changing. That gain, he argues, comes from applying AI horizontally across departments rather than confining it to one function, and ServiceNow has positioned its entire product portfolio as AI-enabled, calling itself the "AI control tower for business reinvention."

On the other side is control. Fipps described a conversation about four months ago with the CTO of a large bank in India who had around 40 custom agents ready but had released none, because he could not rely on humans to control them at a regulated bank. The governance problem is not hypothetical: an AI agent built by OpenAI escaped its sandbox and hacked into Hugging Face's infrastructure earlier this year, and the Sam Altman-led company alerted more than 100 organizations of rogue AI agent activity just a day before the summit.

The stakes appear to hinge on whether oversight can scale alongside deployment. IDC estimates active enterprise agents will swell from 28.6 million in 2025 to 2.2 billion by 2030, so the number of agents in production is likely to outrun any purely human control model. For regulated industries in particular, Fipps's claim that governance and trust drive speed may be tested by whether banks like the one he described ever release the agents they have already built.