No tags match.
UK financial firms check AI thoroughly before launch, but only about half keep monitoring it once it’s live. Here’s how to run an AI workflow after go-live: one named owner, four quality signals, a running cost per outcome, a rule for when to scale or stop, and a second workflow faster than the first. Part 4 of Pilot to Production.
The FCA’s October good-practice paper found the firms doing AI well tested, challenged and signed it off before launch, while retrofitting governance afterwards left gaps. Here are four gates to pass before you switch an AI workflow on, from acceptance criteria to a rollback you’ve actually used. Part 3 of Pilot to Production.
We compared the free AI readiness assessments from Microsoft, Cisco, Thoughtworks, Google and ourselves on what each one scores. Most cover strategy, data, skills and governance. None asks about the AI agents you’re already running, so here are five questions to add.
In a researcher’s test, one prompt was enough to take over every AgentCore agent in an AWS account region because the default role was too broad, and OpenAI found its own agents weakening the monitoring meant to watch them. Here are seven controls an AI workflow needs before anyone relies on it. Part 2 of Pilot to Production.
Half of Gartner’s top strategic technology trends for 2027 describe controls, not capabilities. Turn AI agent identity, coding agent harnesses, exposure remediation, context platforms and SovereignTech into five questions your estate should answer with evidence, and score yourself.
Anthropic’s IPO filing says it may imminently hit its own threshold for AI-accelerated progress, and admits it has spotted past speed-ups only after they happened. As AI builds more of your systems, here’s how to see the change before the lagging numbers do.
Microsoft has told enterprises to treat AI agents as insiders. This month’s breaches show the real risk is access nobody governs, and the fix is the joiner-mover-leaver lifecycle identity teams already run, automated for a new class of identity.
Brussels just deferred the AI Act’s high-risk duties to December 2027, and the exhale from compliance teams was audible. But models degrade on their own calendar, and the monitoring evidence you’ll need in 2027 can’t be backdated. What actually moved, what didn’t, and what to do with sixteen months.
Servers crash loudly. Models fail silently – they keep answering, slightly more wrongly, week after week. Most organisations that shipped AI in the last two years have no way of knowing whether it’s still right today.
The CFO question has changed from ‘what is our AI strategy’ to ‘show me where it is running’. Mark Beard, Founder and CEO of Vertex Agility, on why pilots die, why regulated businesses will pull ahead, and the honest test of any AI programme: count the workflows, not the pilots.
Nvidia and six of the world’s largest investors will mobilise more than $500 billion of third-party capital to finance AI compute, while OpenAI’s latest adoption data shows frontier firms pulling 8.3 times ahead of typical enterprises on actual use. Money has stopped being the constraint on enterprise AI – execution has not.
Every statement in the AI Leadership Readiness Scorecard is asked twice, once about you and once about your organisation. Leaders fixate on the two scores. The finding is the distance between them.
