The road to AI adoption may never be clear – so how do we move?
The path to AI transformation is as yet unclear. Many organisations are waiting for things to settle before they take the first step. But what if that never happens? Here’s how to move forwards regardless.
The need to embrace AI at an organisational level is now unavoidable. Beyond market and board pressures, informal use by staff (shadow AI), and the fact that it’s increasingly embedded in the tools we use – opting out is now impractical, if not impossible.
But the associated risks haven’t gone away either. Ethical and data issues, cost, and the complexities presented by AI transformation are overwhelming. We’re told the possibilities are endless – but in reality this means there is no playbook to work from. On top of that, the pace of change means the goal posts are constantly moving. Where do we even start?
Our human tendency, when presented with uncertainty, is to do nothing – even when the potential risks of inaction are greater than taking action (the Ellsberg paradox). In other words – when the path is unclear, we’d rather stay still.
But what if the path is no clearer in five years time? Will we still be stuck in this same thought process?
You can’t plan for what you don’t know – but you can adapt
The pace of change is unlikely to settle down any time soon. But when your organisation’s approach to new technology is driven by strategy and, critically, a good handle on data (it may be dull and painful, but it’s essential), it’s much easier to adapt to change triggered by AI.
The main focus in 2024 was to create an AI policy. Last year it was trying it out for efficiency. This year it’s more about experimentation. An approach that includes and encourages your people to experiment with AI tools is more likely to bring them with you, so change is done with them rather than to them.
A health charity I worked with last year started off by aligning the whole organisation behind how AI should be used – they created an AI policy for staff use. Next they spent a year experimenting with different AI tools, keeping a log of what data they used and how useful the outputs had been. Then they asked staff to list all the problems they wish AI could solve. They took all this information and came together in a workshop to discuss which experiments with AI should be prioritised. Finally, they ended up with a roadmap to work from – choosing which AI project to pilot first.
This simple approach works because it helps the organisation adapt to change in logical, practical steps and it does not push organisations to innovate before they are ready. It also considers people alongside technology.
An AI policy is a start. But having a solid understanding of what’s happening within your organisation in terms of AI usage is now essential – as shared in this excellent article by Ryann Miller, written after speaking to over 150 nonprofit leaders.
Ideally we need to understand the landscape before we can dive in, and we can’t do that unless we open up a dialogue with the people working in our organisation, to learn:
How they’re already using AI in their roles
Their ideas around how the technology can be used in a way that it isn’t right now
The frustrations they’re experiencing that AI could help with
Once we get a sense of what’s already happening and what are the problems colleagues wish AI could solve, it’s time to decide what can actually be done or tested.
The experimentation phase is when things really start to get interesting…
You can only be strategic about tech when you know what it can do
In a recent event on AI transformation I organised with my colleagues from Charity Change Collective, I loved hearing examples of what non–profits have already been doing or thinking about:
Using AI to draft emails in an organisation's tone and style.
Spotting patterns in help line queries and developing content that's missing.
Cutting a grants process from two weeks to half a day.
And my personal favourite – finding every place KPIs are referenced/reported to show how differently they’ve been defined (the team is now working to standardise those KPIs one by one).
AI is a brilliant tool for spotting patterns, surfacing detail and sorting out your data mess. It can make an invisible problem visible.
Many moons ago I said to my manager – there’s a lot of detail in digital but you kind of need to get it in order to be strategic about it. I changed my mind a few times about it, thinking that knowing principles might be enough. But with the increasing need to make decisions about AI, I’m back to that thought.
You need to know what it’s capable of (the good, the bad and the ugly) to be creative about how you apply it to the problems you have. And a lot of that starts by learning and trying it out to see what it can do.
Your tech stack is only as good as your people stack
AI and digital transformation are similar in that the new tech surfaces and multiplies structural and organisational issues. It does not solve them.
With digital transformation you could get away with a sticking plaster approach, where you do the tech bit but not look at related processes, skills and capabilities of people who are meant to use them (I describe this scenario as Mini drivers at the wheel of a Porsche).
Partial transformation is not possible with AI. That’s because AI is going to be everywhere, at the core of products and tools we use – at work and in life. We are already using it and being exposed to it without even noticing (e.g. social media and shopping algorithms).
With digital transformation non-profits could get away by not fully changing. With AI transformation, if we don’t look at how we need to adapt - re–design the way we communicate, manage our data, deliver services, work, the skills we develop and recruit, behaviours we nurture and reward – I’m not sure we’ll be able to survive.
Change is continuous – so manage it
Organisations are now facing constant change, so the best response is to build resilience and manage the change continually: be clear why change is happening, what's the urgency, what will be changing and what's staying the same, what skills do people need, how will they get them, prioritise and pilot new ways of working and new products. This should be embedded in organisational thinking not just for the odd technology or AI project.
In a world that's constantly changing, organisational adaptability and resilience are vital. The tech has only ever been part of the story.
