Outcomes at Work, Part 2: Where Advantage Comes From Now

There are plenty of AI ROI surveys around, but I would not get too caught up in the exact numbers. The pattern is more interesting. The gains are showing up where organisations have moved beyond pilots and built AI into the way work actually happens. Elsewhere, it is still mostly experimentation. That is the real divide: using AI to speed up old work or starting to rethink the work itself.

In most knowledge work, augmentation is the more useful line of thinking. Full automation is rarely an advantage if competitors have access to similar models. The stronger position comes from knowing where AI is reliable, where it is not, and how to redesign the human role around that reality.

Some will, and in some cases they already are, but that does not make it a good strategy. Cutting headcount is easy to explain to the market; building a better operating model is much harder.

In software engineering, the work is already shifting from writing every line by hand to reviewing outputs, framing problems properly and catching what the machine misses. In finance and accounting, the same pattern is emerging: less manual preparation and routine analysis, more interpretation, orchestration and advisory work. My view is that organisations that are deliberate enough can hold both: stripping out the drudgery while protecting the conditions in which judgement develops.

It is becoming a category in its own right, sitting between workforce strategy, operating model, and execution. It is precisely where Vertage is leading in the market.

Because they add tools without redesigning the system around them. A pilot can make an individual task faster and still create almost no enterprise value if the workflow, incentives, accountability, and measures around it stay the same. The biggest multiplier is institutional rather than individual: culture, management discipline, decision rights, and workflow design determine whether AI becomes real operating leverage or just another layer of activity.

It means organisations need to think much harder about how people build judgement and experience early in their careers. If some of the routine work starts to disappear, the old apprenticeship model becomes less reliable on its own. So leaders need a clearer answer to how people learn, where they get exposure, and how they grow into more valuable roles over time.

Both. Reskilling is no longer a fixed programme aimed at a stable target; the target keeps moving. That is why organisations need two things at once: hands-on experimentation with AI inside the flow of work, and stronger individual responsibility for learning where these tools help, where they mislead, and where human oversight matters most.

You do not start with a giant training catalogue. You start with the work changing fastest and the roles closest to value creation. The organisations moving quickest build a baseline level of AI fluency, then embed learning into live workflows. This is where Vertage can play a practical role: bringing knowledge, experience and oversight together so organisations can scale people's impact, while helping them steer an agentic AI workforce with clear guardrails, quality checks, and release models.

It means being able to zoom out, rethink the process, and move quickly on the use cases where reliability is good enough and the operating benefit is real. In practice, the strongest examples often come from small teams close to the business, with clear executive backing, strong guardrails and the authority to redesign how work gets done.

They should look beyond activity. The right measures are outcomes: profitability, reinvestment capacity, customer experience, operating performance, speed of execution, and the organisation's ability to move people into higher-value work.