Watch the TechSmart Q&A: What membership organisations need to know about AI

Tim: An AI agent is autonomous software that can follow rules or make its own decisions. For example, it could listen to customer calls and act on behalf of the company — sending emails or making purchases. Agentic AI has only become viable in the last six months. While some people know how to use it, most don’t yet understand it. Adoption starts with awareness of risk. Two things are key: basic AI literacy through training and small, safe tests before rolling out wider systems.

Alex: Websites won’t disappear, but their role will change. Browser makers are driving this shift — agentic browsers like Perplexity already act on users’ behalf. In future, associations may publish tools and capabilities that AI agents can use directly.

Tim: Websites will remain as information hubs, but the biggest changes will be in search and advertising.

Tim: AI could transform the NFP sector, especially where resources are limited. Building confidence starts with a mindset shift — encouraging curiosity and safe experimentation. Next, ensure people understand risks. This combination builds trust and helps them find effective solutions.

Alex: Associations are great at sharing knowledge. Openly sharing ‘recipes’ for success and using synthesised data or local models to reduce personal data risks helps build confidence across the sector.

Alex: It depends on the risk. If the impact of an error is low — such as assigning the wrong membership level — a single check at the end may be enough. High-risk cases, like medical uses, need more oversight. Multiple AI agents can also cross-check outputs.

Tim: Humans must know what ‘good’ looks like. Every process needs clear goals, defined outputs, and human review. Include regular audits and spot checks. Humans need to stay in the loop because they understand emerging risks. Build risk management in from the start.

Tim: Think long term. There’s extra effort at first, but it’s a one-off. Within six to twelve months, efficiency gains become clear. AI agents can hugely increase productivity once embedded in teams.

Alex: It’s like onboarding a new colleague. With clear processes and SOPs, change becomes much easier — whether it’s people or technology.

Alex: Keep sharing what works and what doesn’t. That shared learning builds collective knowledge and improves results for everyone.

Tim: That’s our focus — connecting commercial and not-for-profit organisations of all sizes to improve AI understanding and adoption.

Tim: The UK should lead in ethical AI. We can do that by rewarding companies that build ethical, sustainable tools and empowering consumers to choose them.

Alex: AI usually does what it’s told. If instructions are clear and grounded in organisational knowledge, nine times out of ten it works well. But errors happen. Make it easy for people to flag issues and use AI agents to audit conversations for recurring problems.

Tim: Hallucination is common in generative AI. Build processes early — regular checks, audits and back-and-forth questioning of the AI. Understanding risks and verifying accuracy must be ongoing.

Alex: Boards cautious about rapid AI adoption are often right to be. The pandemic proved digital change can move fast, but AI needs solid foundations first.

Tim: Many organisations haven’t finished digital transformation. Those that have built cultures of testing, learning and agility are adapting well. Others struggle when leaders chase AI for novelty. Start by identifying real problems, then build the right mindset.

Alex: It’s about iteration. Start with a valuable problem and build practical, affordable solutions around it. Small steps lead to meaningful progress.

Alex: Focus on knowledge-sharing. Start with AI tools that help staff answer member or stakeholder questions. Begin with a small pilot where staff validate responses before sending them. As accuracy improves, open it up to members. Start small, test, and refine.