?>

Tech Leaders Collective: Agentic AI – Moving from Possibility to Practical Value

Posted on 5 Nov 2026

Industry Insights & Interviews

AI tools are becoming part of everyday working life but helping an individual write an email or summarise a document is different from using AI to improve an entire business process. For technology leaders, the challenge is understanding where that next step can deliver meaningful value, and how to take it responsibly. 

Our October Tech Leaders Collective meetup explored that shift through a practical session on agentic AI. Combining live demonstrations, examples from healthcare and group discussion, the event examined how AI agents can support operational workflows, improve access to services and reduce administrative effort, without losing sight of the people using them.

Understanding Agentic AI: Beyond a Single Task

The session explored the progression from basic task automation to more connected, context-aware workflows. A task agent might extract information from an email and enter it into a system. A workflow agent can take that process further, using the information to trigger the next appropriate action. 

More autonomous agents can work towards a defined goal, drawing on connected systems and tools. However, the discussion distinguished these possibilities from the reality of deployment: organisations need to understand what an agent can reliably do, what access it requires and where human oversight remains necessary. 

For tech leaders, the starting point is not how much autonomy a system can offer, but what level is appropriate for the problem being solved. 

Practical Examples: Improving Access and Reducing Administration 

Healthcare examples brought the discussion to life. A virtual agent used by NHS 111 Wales illustrated how people can ask questions in natural language and receive information drawn from a clinically approved knowledge base, rather than the open internet. 

The session also explored a GP practice’s use of virtual assistance to support patient registration and flu vaccination requests. These examples showed how relatively focused changes can make services easier to access while reducing the need for routine phone calls. 

A live voice demonstration extended that idea, showing an appointment-booking interaction and answers to practical questions about a clinic visit. It also exposed the limits of the demonstration’s configuration: the agent could not send the requested reminder. This reinforced the importance of testing real interactions and being clear about what a solution can and cannot do. 

Designing Around People, Not Just Technology 

A recurring theme was the need to start with the user experience. Patients, customers and employees have different preferences, needs and levels of confidence with digital tools. A solution that works well on a website may need a different approach on WhatsApp, over the phone or within Microsoft Teams. 

The speakers emphasised that introducing AI should not mean removing every other route to support. Human assistance still has a role, particularly when a query falls outside the system’s capabilities or someone simply needs to speak to a person. 

Successful adoption depends on delivering useful answers through accessible channels, not expecting everyone to adapt to the technology in the same way.

Building the Foundations: Purpose, Platforms and Data 

The session outlined a practical framework for considering AI adoption. First, there must be a clear business purpose: improving service, reducing administrative work or addressing a specific operational bottleneck. Defining that purpose also makes it possible to measure whether the implementation delivers the intended improvement. 

Leaders then need to consider the people using the solution, how it connects with existing systems and the quality of the information behind it. An AI agent cannot make an outdated knowledge base reliable simply by presenting it more conveniently. 

The examples highlighted different approaches to data readiness, including a practice that refreshed its website before introducing a virtual agent. The wider lesson was clear: technology choices and implementation plans need to reflect the organisation’s starting point. 

Start Small, Learn and Build 

The event’s central takeaway was practical: organisations do not need to automate everything at once. Starting with a clearly defined inefficiency gives teams an opportunity to test the technology, understand the user response and build confidence before expanding its role. 

Agentic AI offers opportunities to connect tasks and improve services, but its value comes from solving real problems. For tech leaders, the next step is to identify a useful starting point, establish the right safeguards and measure what changes. 

At Tech Leaders Collective, we bring technology leaders together to share experience, challenge ideas and explore what emerging technology means in practice. Join us at a future meetup to continue the conversation. 

Written By:
Tech Leaders Collective: Agentic AI – Moving from Possibility to Practical Value
Stella Webster

Connect on LinkedIn