Turning AI Investment Into Measurable Business Value With HP

How can businesses turn growing investment in AI infrastructure, cloud capacity and devices into outcomes that employees, customers and finance teams can actually measure?

In this episode of Tech Talks Daily, I speak with Neil Sawyer, who manages HP's business across Europe, the Middle East and Africa. Neil works with companies across one of HP's largest global regions as they move from AI experimentation into wider deployment, making him well placed to discuss what happens when early enthusiasm encounters cost, security, governance and the realities of the workforce.

We begin with the gap between building AI capacity and applying it to a business problem. Data centers, models and powerful devices provide options, but the investment only becomes useful when a company identifies the workflow it wants to improve. Neil argues that leaders should begin with the outcome, understand where AI can remove friction and decide how they will measure productivity, employee experience and business performance before buying another layer of technology.

That raises a difficult question about productivity. If AI helps someone complete a task faster, does the organization use that saved time to improve the work, develop new ideas and give employees room to think, or does it simply add another task to the queue? I share my own experience as a business of one, where every efficiency gain has a habit of becoming extra output rather than a Wednesday afternoon at the cinema. Neil compares the current moment with earlier periods of industrial change and makes the case that automation should release people from repetitive administration so they can contribute creativity, judgment and higher value work.

We also discuss why AI costs are becoming a boardroom issue. Token based services and agentic systems can produce growing and unpredictable bills as adoption spreads across a company. Neil explains why every workload does not need the same model or environment. Large language queries may benefit from cloud capacity, while sensitive data, company specific information and some recurring tasks may be better suited to local or on device processing. The decision affects cost, responsiveness, privacy, security, data sovereignty and environmental impact.

Neil describes HP's view of hybrid AI, including devices with neural processing units and Z by HP Boost, which can connect available workstation GPU resources. He also explains why device refresh decisions should reflect workforce personas. A data scientist, account manager and office administrator may work for the same company, yet their computing needs can be very different. Mapping technology to the employee's role can help a business spend with greater discipline while giving people the performance they need.

The conversation also covers governance and measurement. Informal use of public AI services can be difficult to see, assess or manage. Neil recommends giving employees an approved AI toolkit, using enterprise services that provide telemetry and examining how technology availability and performance affect the employee experience. Adoption figures can show that a tool is being used, but they do not prove that it is improving an outcome.

We finish with a practical checklist for leaders. Define the outcome, identify the workflow, determine where each workload should run, calculate the cost, agree the measures of success and put clear controls around data, privacy and cybersecurity. That approach gives cloud and device based AI distinct jobs within the same business strategy.

How is your organization deciding which AI workloads belong in the cloud, which should run closer to the employee, and whether the investment is producing measurable value? Share your thoughts with me.