Companies are moving beyond AI pilots, but access to tools alone is not producing productivity or business value. At Fortune’s AIQ Summit, executives described a more demanding test: redesign workflows, train people and put governance and costs on the same footing as deployment.
Watch Desk analysis
What happened
The summit brought together leaders from technology, finance, industry and other sectors to discuss how organisations are putting AI into practice. Their examples range from workforce training to customer service, risk controls and model spending. The figures below come from the summit report, which attributes them to speakers and named company research.
Our top picks
- Training, not just access
ServiceNow’s maturity index found 57% of “pacesetter” organisations invest in ongoing AI upskilling, compared with 4% of other organisations. - Hiring is part of the divide
The same index found 68% of pacesetters attract, hire and retain AI talent, against 10% of other companies. - A specific workplace deployment
Palo Alto Networks’ chief people officer said a custom AI agent had reduced IT tickets by 83% and supported workflows in HR, finance and legal. - Automation can cut costs, but model choice matters
Ecolab’s chief AI officer said the company reduced token costs by roughly 70% to 80% after optimising its model use. - Governance can set the pace
Bank of America’s technology leader said the bank assesses projects across 16 risk pillars, including privacy, bias, workforce impact and intellectual property.
Why it matters
These examples point to a practical distinction: buying AI access is not the same as improving work. The summit report says Booking Holdings is routing simpler tasks to cheaper models, while reserving more capable ones for work that needs them. Its CFO also said the company’s roughly 9,000 engineers are getting about 30% more code into production, counting merge requests that pass testing and quality controls.
That is a more useful measure than how many tools an organisation has switched on, though the figures are company-reported examples rather than a like-for-like comparison. The report also notes that AI-generated output can create extra work for colleagues, and that in safety-critical industrial settings the reliability bar is much higher than for a desktop assistant.
Our read
The most convincing message here is refreshingly unglamorous: start with a problem, redesign the workflow, train the people doing it and measure whether the result is better. AI is not a productivity wand; it is a capable new component in a workplace that may still need rewiring. Leaders should be able to explain what changed for staff and customers, not just how many pilots made it through the slide deck.
What to watch
- Whether organisations publish outcomes beyond access and usage figures.
- Whether training and workflow redesign accompany new deployments.
- How companies measure quality, costs and extra work created by AI-generated output.
- Whether governance and human review remain in place as systems take on more tasks.
Discussion spark: Should employers judge AI success mainly by productivity gains, or by whether the work becomes better for staff and customers?
Sources and evidence
- Source update (12 October 2026, 20:04 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.