Currently it seems every organisation I talk to wants to talk about AI. But then, if you tell any charity or membership body that they may be able to ease pressure on over-stretched resources or deliver a little more for their beneficiaries they are going to be keen to know more.
But, things are not as simple as they sound… Most organisations we work with have a lot already on their plates – competing priorities, projects backing up and a stack of legacy systems.
A lot of organisations have started with Copilot, run a few training sessions and then got a little stuck. There is nothing wrong with that as a starting point but what can you do next? We’ve been working with several organisations on exactly this question and I’ve developed what I think is a deeply unglamorous – but very pragmatic – approach to the problem.
The good news is that it’s straightforward, you don’t need to be technical to understand it and it is aimed at getting you up and running rapidly.
At its heart, the approach comes down to a few simple steps:
- Look for the problems most worth solving – the biggest opportunities often lie in your most repetitive, frustrating or time-consuming work.
- Figure out if the cure is better than the disease – find the organisation’s best bets by establishing an early view of what it would take to change and what it could deliver.
- Don’t eat the elephant whole – start with a manageable pilot, learn from it and build on what works.
- The future is in the foundations – use those early projects to develop the knowledge and momentum to take the next steps.
1. Look for the problems most worth solving
A good problem is likely to be one that is currently the source of significant organisational pain – one that takes up a disproportionate amount of time, uses a lot of resources or causes significant friction for your supporters or beneficiaries.
AI, and technology in general, is really good at certain types of work. AI is particularly useful for things like pattern recognition, processing large amounts of information, summarising and drafting content, retrieving information and making sense of data. It’s also great at repetitive tasks and never loses the will to live, no matter how dull the task in front of it.
Of course, if a process follows a predictable set of rules, conventional automation might do the job just as well, potentially at a lower cost to both your budget and the environment.
What we’re really looking for are those opportunities where the potential benefit is significant enough to justify the time and money involved in making a change.
2. Figure out if the cure better than the disease?
So far, in the charities and membership bodies we’ve worked with, we’ve found long lists of problems worth solving. It’s amazing how many unloved and dated processes hide in plain sight across organisations.
Understandably though, organisations want a reasonable degree of confidence that this kind of process re-engineering work is a good use of their constrained resources.
To work out which projects are likely to be winners, we normally use some key criteria – tailored to the organisation – but typically looking like this:
- What resources does the process currently take?
- What cost, process or user benefits will it deliver?
- Can we pilot it easily and cheaply?
- What risks or barriers would we need to overcome?
It should be possible to make a decent estimate of current costs, based on resource usage and team budget.
But you want to understand whether the cure is better than the disease – there is little point spending a fortune automating a process that only takes a few hours a year – so you also need to do a very high level estimation of what it might cost to run a project, whether it is possible to do a pilot and what the ongoing level of cost would be after.
This can be done at an order of magnitude level and can be based on experience of similar projects etc.
For one large organisation, where an initial 90 ideas were submitted, we were able to use this process to filter down to those projects with the potential to deliver the most benefit relatively quickly. We then went through the results with the organisation’s director, agreed which were worth pursuing and worked five of them up into more detailed project outlines.
3. Do not eat the elephant whole – start small and learn as you go
One of the things we look at is whether we can run a small pilot or proof of concept before committing to a larger project. This allows us to make progress quickly, test whether the expected benefits are realistic and build the knowledge and confidence to take the next steps.
We are currently working on one of the first pilot projects to emerge from the prioritisation process: a way for a large membership organisation to deal with thousands of queries.
The first stage is to implement a pilot for one particularly resource heavy process. It will involve a trial of some leading customer service software and introduce some straightforward automation.
Oddly, I’m quite pleased to say that it doesn’t currently involve AI. The work should deliver some relatively quick benefits through conventional automation without taking on the additional cost, risk and complexity of an AI project.
But it also gives them a better foundation for the future. Once the process is running through a proper customer service platform, with workflows and information better organised, then trialling AI in a controlled way is much more achievable.
It’s not particularly glamorous, but that’s rather the point. We can test the benefits of a relatively straightforward change now, show impact, and build the case for delivering even more.
4. The future is in the foundations
There is some interesting research suggesting that organisations that are already good at digital and data are also much better at putting AI to work.
Research from BCG found that organisations with more mature data and AI capabilities had scaled four times as many use cases as those lagging behind. McKinsey has also found that redesigning workflows is strongly associated with getting better business results from AI.
That doesn’t mean you need to fix every legacy system before you can get started. But it does suggest that identifying your highest priority processes and redesigning them is likely to be a key step in ensuring you are ready for the future.
What to do next?
For me, the starting point for any digital project should always be getting the clearest understanding possible of the problem you are trying to solve.
AI is no different. You can get started with some structured thinking, sensible prioritisation and a willingness to test things out.
That will build confidence, knowledge and momentum to take on more ambitious projects.
We’ve been helping charities and membership organisations work out where to start, what is worth doing and how to turn those ideas into achievable projects. If you’d like to explore how that might work for your organisation, we’d be very happy to have a chat.
Next up, I am considering a piece on the issues you should be considering before you go wild with AI. If you have views on what would be useful to you or some thoughts you want to share, please email me at ali@wearemc2.co.uk.