One of the key challenges with the application of AI is that it is such a broad technology with such wide potential applications in so many areas how do you prioritise where to place the effort and resource? I’ve had a go before now at categorising potential application of AI through the lens of seven outcomes, but it’s still a complex task to know what to focus on first.
Good prioritisation is often a balance between value (the impact or return that doing something will result in, its importance and urgency) and complexity (how hard or lengthy the task is to complete). The prioritisation technique (often used in SAFe agile methods) called Weighted Shortest Job First (or WSJF) positions this as ‘cost of delay’ against job size or duration. Cost of delay is the cost of NOT doing something, which is another way of thinking about both value and urgency. Job duration and size is akin to complexity – the more complex the task the longer it might take and the more resources it may suck in. I like it because it takes account of many of the most important factors to consider in prioritisation – urgency, time, value, impact, resources, difficulty and complexity.
The idea with WSJF is that tasks which have the highest cost of delay but lowest duration or complexity are your quick wins. Those tasks which conversely have a high degree of complexity but a low return or value should be deprioritised.
Of course, we need to consider carefully how we are categorising value and complexity. Relevant to value may be how aligned the application of AI is to achieving organisational objectives, the payback and return of investment, the strategic benefits in terms of areas like customer experience, avoidance of risk, identification of opportunities, or productivity gains. It’s worth considering value from the perspective of audience (value to who?): customers, staff or partner organisations, as well as scale. The ease of scaleability (is this something that can easily be scaled to impact multiple customers or teams) should also be considered.
Relevant to complexity may be how resource-intensive the change or implementation is projected to be. How long the change will take to show benefits, and whether the skills and knowledge exist within the business to actually do it. For example, many of the applications of AI will likely come from AI capabilities which are embedded into upgrades of services, systems and technologies that the teams already use. This may require training or adaptation of processes but is a relatively quick win.
I think this is a useful lens to think about all the potential applications of AI. There will be areas which inevitably are of high value and easy to implement, and others which will require significant time and resource where the benefit is more questionable. The watch out here is that there may be areas which have high complexity or duration but which also have a high value (the box at the top right of our 2 x 2). These areas may not be urgent but they are important in building long term capability. Similarly there may be things that we can do quickly and easily, and whilst the value may not be the highest, they occasionally may float to the top of the list (the box at the bottom left of the 2 x 2). But as long as we take this into account this is a simple but pragmatic way of understanding application.
Of course, experimentation (and the knowledge that we can gain through doing it) is important in building our understanding of both value and complexity which is partly why it’s so important to test and learn.
Pop your email into the box below to get posts emailed straight to your inbox
Leave a Reply