As you start to put together your AI strategy, you need to understand whether you’re pursuing a purposeful AI strategy or an emergent AI strategy. This will enable you to find the best route to value. What will determine your approach is your organisation, your position within it, and the culture it exhibits. In this article, I’ll cover what an emergent and purposeful AI strategy is and how you can turn the most common type of strategy into the most effective one.
First, let’s define what we mean by the two.
What is a purposeful AI strategy?
A purposeful AI strategy is one that is led from the top, defined by leadership, with a dedicated budget, support, targets and direction. It’s aligned with your company’s strategic objectives and is pursued meticulously and with focus. With a purposeful AI strategy, you’ll have at least one team with the sole objective of bringing AI to fruition at your company.
That’s not to say that the strategy will be holistic, meaning spanning the width and breadth of your organisation. It may well be confined to a specific business unit initially. What’s important about the purposeful strategy is the mandate from leadership to spend the time and money and make the changes required to bring it to fruition.
Over time, that purpose will spread to other areas of the business methodically while being driven by a dedicated team with the full backing of the organisation.
What is an emergent AI strategy?
An emergent AI strategy is one that’s led from the grassroots. Someone in IT plays around with some technology and tries it out in a use case that works out. So, they keep going onto another use case and so on. Or, someone in the contact centre buys a vendor solution to automate live chat conversations for customer service to solve that particular pain point.
These are emergent because they’re siloed groups of people trying to solve their own pain points in the only way they can control. It’s all about individuals or small teams or divisions, with little to no budget, doing what they can to positively impact their specific areas of the business.
Which approach is right?
There isn’t a right or wrong answer for where to start with your AI strategy. You have to start from where you are, based on your role, influence and what you can control.
If you’re a senior leader, your role lends itself to you being responsible for putting together a purposeful strategy. If you’re an IT manager, for example, you might pursue an emergent strategy.
Of course, where you want to end up is having a holistic, purposeful AI strategy that has the full support of the organisation. How you get there, though, will differ from company to company, person to person.
The vast majority of organisations start with an emergent AI strategy to prove the value of AI, which turns into a purposeful strategy as leadership begin to understand the opportunity and fold it into the wider organisational transformation plan.
Your emergent AI strategy is an experiment
It’s useful to view your emergent AI strategy through the lens of experimentation. You are not committing to realising any specific monetary or any other type of value. You are testing the hypothesis through experimentation to see whether a specific value can be realised. This experimentation mentality enables you to have a test-and-learn approach and provides freedom from the organisation’s scrutiny.
One of the ways that you can make sure to be diligent and purposeful with your experimentation is to use a hypothesis test card to track your hypothesise and results.
The dangers of a persistent emergent AI strategy
So, having an emergent AI strategy is completely normal and common and perhaps even an ideal place to start, depending on where you are. The trick is in spotting when it’s time to make it purposeful. This is because keeping your emergent strategy emergent for too long will wind up costing you. Why? Because you’re probably not the only division doing it. Perhaps marketing also has an emergent AI strategy. Maybe customer support does too. And HR. And IT. And sales. And so on.
That means you’re going to end up duplicating costs and capabilities. That means you’re adding waste. Your customer experience isn’t joined up, and you have 4 different technology infrastructures and teams that do four different things, none of which talk to each other. It’s also not clear whether there’s alignment with wider strategic goals.
This is the time to wrap a purposeful strategy around it, consolidate capability and spending, and then centralise governance, best practice dissemination, training and impact measurement.
If you’re a leader in your business, it’s your duty to sniff this out because this is a sure sign that someone needs to step in, unite the sporadic efforts, consolidate capabilities and spending, create consistency, and fully represent value.
How to turn an emergent AI strategy into a purposeful AI strategy
At some point, you’re going to need to turn your emergent AI strategy into a purposeful one. But how do you do that?
As mentioned, the specifics are different in every organisation, depending on your size and culture, but generally speaking, you need to accomplish some variant of the following:
1. Show promise.
The first thing you need to do is show the organisation that the technology has promise. It can be used to solve a specific problem you have, and it can scale up to solve further problems in the future.
This is typically done through research, experimentation, prototyping and eventually the creation of a proof-of-concept (POC).
A POC is a basic working version of a potential solution that aims to prove that your idea is likely to work in practice. This POC doesn’t need to have all the bells and whistles of a production system. It can be super-basic. It’s not something you’ll put live to customers either (at least not all of them), so you’ll likely go through many iterations to validate that you can design the intended experience.
Your goal here is to understand the viability of the solution and build the scope and requirements to take it to production. You will test it, however. This can be done with customers or intended end users (ideally) in a controlled environment.
You’re seeking to learn whether your POC meets the success criteria you set out. Depending on the company and use case, this will differ, but generally, you want to learn whether it can do the job you intend.
This can stem from anywhere in the organisation. The innovation team, IT, marketing, sales, HR, product, you name it.
2. Prove value.
A Proof-of-Value (POV) is where you take your solution into a live environment to test and learn whether it really solves the problem you intend it to in the wild. You may have heard this called a Minimum Viable Product (MVP) before.
3. Share results.
A good percentage of managing an AI programme is about telling the story to stakeholders. This is how you’ll educate them on what’s possible and show the business value you’re achieving.
- Demonstrate wider application. Then, you need to show the organisation that this technology isn’t just confined to the specific area of focus in your proof of value. You need to show that it can be used in many more areas in your department and across the organisation.
- Communicate. Returning to your RACI matrix, who needs to see this plan? Who can give the go-ahead for the next step? Who is responsible for the strategic direction? Spend some time with them to show them the potential applications for this technology. Your goal here is to get the resources or funding to work on a more detailed business case for a priority use case.
- Develop a business case. A business case is a document that highlights the problems you’re trying to solve, the high-level solution you propose, the technology, resources and cost of making it happen and, crucially, the outcomes you intend to realise.
- Prioritise the next 12 months. Now, take a look at your proposed business case and use cases and split this up into some achievable milestones you plan on hitting over the next 12 months. You can use the framework in this piece to prioritise use cases. It’s always easier to ask for a little to take a step forward than it is to ask for everything on day 1. Having a plan for where to start will help you kick on quickly once the ball is rolling. We’ve seen some organisations take a 12-month initial view to prove value at a relative scale before moving ahead with more concrete operational plans. This means that the budget you ask for is a 12-month budget, and the technology you procure is for the same duration. This reduces the reliance on long-term commitments and outlays and will enable you to develop sufficient maturity so that you can make more informed and long-term technology and resource decisions as you develop your maturity. (Don’t be pressured into vendors asking you to sign 3-year deals. At this stage, you want 12 months and 12 months only.)
- Make your ask. Now, you need to land your plan. Make your ask. Ask for the budget, mandate and resources to make it happen. This isn’t as straightforward as a single meeting with your boss. You’ll likely have to go on an internal roadshow, presenting to senior leaders and teams across the organisation to sell your story and get them on board.
- Prepare for launch. The next step is to prepare to launch your programme. You’ll need to determine the people and technology needed to deliver your initial use cases. Then, you’ll need to put together an approach to AI delivery that mitigates risk and focuses on proving value quickly.
From theory to action
Not every company has the skills needed to deliver successfully, and not every company can afford to take the time to learn on the job. Some would benefit from a kickstart. At VUX, we offer end-to-end AI design and delivery training services with on-the-job support to enable the team to be successful from day one.
Technology choices can also be a complex minefield if you don’t know what you’re looking for. There’s snake oil, fakers and the like out there, so separating the wheat from the chaff and selecting is important. Again, the service we offer at VUX helps you define your requirements based on your use cases and future vision, then helps you find, vet, and select the appropriate technology for your needs.
If you need any support in taking this approach from theory to practical, feel free to drop us a line, and I’m sure we can have you on your way successfully in no time.
