Localised AI: Why bringing AI closer to the user matters | Softcat
Skip to main content

Localised AI: Why bringing AI closer to the user matters

AI doesn’t always need the cloud, here’s why local processing deserves a place in your workplace strategy.

Softcat PPT Background Corner Lit Radial Plum Gradient RGB Softcat PPT Background Corner Lit Radial Plum Gradient RGB

Daniel Paton

Workspace Advisor & Deputy Team Leader

AI is quickly becoming part of our everyday working life, but as organisations adopt more AI-enabled applications, there’s another question worth considering:

Where does all that processing actually happen?

Traditionally, many AI workloads have relied heavily on cloud infrastructure, with data sent to remote services for processing before a response is returned to the user. Localised AI offers another option, allowing supported AI workloads to run directly on the PC, using dedicated hardware such as Neural Processing Units (NPUs). An NPU is a processor specifically designed to handle AI-related tasks, taking some of the workload away from the CPU and GPU.

This isn’t about replacing cloud AI. It’s about making better use of both cloud and local processing, depending on the task. As AI becomes more widely used and embedded into everyday applications, that flexibility could become an increasingly important part of future device and workplace strategies.

Why does localised AI matter?

There are several benefits to bringing AI processing closer to the user:

Reduced reliance on the cloud: Not every AI workload needs to be sent to a cloud service. Where applications support local processing, AI-capable devices can handle certain tasks directly on the PC, reducing reliance on connectivity and remote computing resources.

Potential cost efficiencies: Cloud-based AI relies on computing resources to run workloads. Moving suitable processing onto endpoint hardware could reduce the amount of cloud capacity organisations need, which may become particularly valuable as AI usage grows.

Keeping processing local: For supported use cases, data can be processed directly on the device, rather than being sent elsewhere. This gives organisations another option when thinking about data, privacy and compliance.

A more responsive experience: Local processing means supported AI requests don’t always have to travel to the cloud and back. For users, this can mean faster responses and a smoother experience when using applications designed to take advantage of on-device AI.

Greater resilience: AI features that run locally can be less dependent on a constant connection to cloud infrastructure. This can be particularly useful for mobile workers or people working in locations where connectivity isn’t always reliable.

In practice, depending on the applications and hardware involved, localised AI can add great value to everyday tasks. This includes elements such as creating live captions and transcriptions, as well as suppressing background noise on video calls. It can also help with image processing and enhancement – tasks that historically have taken a significant amount of time.

What does this mean for your device strategy?

Until now, device refresh decisions have largely focused on familiar considerations such as processor performance, memory, storage and battery life. AI throws another factor into the mix: how well can the device handle AI workloads locally?

Many newer AI PCs include dedicated NPUs alongside the CPU and GPU, giving them additional capacity to run supported AI tasks; this is worth considering when planning your next Windows 11 device refresh. We know that devices purchased now could remain in use for several years, during which AI-enabled applications are likely to become a much bigger part of everyday working life. Choosing hardware that can support these developments, and any further changes, could help organisations avoid being caught out as requirements continue to evolve. It's not just about what users need their devices to do right now, but whether those devices will be capable of supporting the next generation of workplace applications.

Cloud AI vs localised AI: It doesn’t have to be one or the other

The future of workplace AI is unlikely to be entirely cloud-based or entirely local. Both have an important role to play.

Cloud AI remains essential for workloads that need significant computing power, access to wider organisational information or cloud-hosted services. Localised AI complements this by allowing suitable workloads to run closer to the user. The key is knowing which approach works best for each workload. Getting that balance right could help organisations make better decisions around performance, connectivity, data security and the cost of delivering AI at scale.

Bringing AI closer to the workplace

Localised AI isn't simply about making PCs more powerful. It gives organisations another way to deliver AI capabilities, with more choice over where processing takes place. For the right workloads, this could mean less reliance on cloud resources, improved responsiveness and greater control over where data is processed.

As AI becomes a bigger part of everyday working life, it’s worth looking beyond the tools themselves and considering the hardware that supports them. The question isn’t just, “What AI tools should we adopt?” but also, “Where should our AI workloads run?”

If you’re reviewing your device strategy, or looking at how AI-ready PCs could support your organisation, speak to your Softcat Account Manager to explore the options, or contact our Sales team.