The practical bit: From AI hype to AI reality
Okay, this is where we get to the practical stuff. We reckon there are four areas to focus on that can get you out of this trough of AI disillusionment.
1. Identify the value of generative AI
We see it time and again with our clients: they don’t have a handle on how to measure return on their investments in AI. This starts with aligning your AI strategy with your strategic priorities. Yes, it makes sense to trial AI in pockets. But you then need to move on quickly to identifying where AI can deliver the greatest impact. That enables you to get to work on a portfolio of initiatives that are most likely to deliver a strong ROI – and to make sure your efforts are aligned.
Sounds good. But measuring the value of GenAI isn’t always straightforward. Some employers may question whether the time their people are saving is being used to do more work or to take longer lunch breaks. We’d say it’s providing time to think and have more in-depth discussions with clients.
How do we do it at KPMG? We build a bottom-up view of your organisation, preferably using your workforce data at a role and salary level. And we talk to you to identify your critical priorities. We can then look at our data on AI uses cases to identify what will drive most value and the level of change required.
2. Build trust in AI
Who do you trust more? A human who answers your query to a call centre, or an AI-enabled chatbot? We’ve talked to clients who’ve keenly promoted the use of GenAI in their businesses. Then when their employees start using it, they start questioning their work.
Building trust in AI is critical to delivering results. You, your people and your customers need to be able to trust your AI. We’ve identified ten factors that are key to delivering trusted and ethical AI:
- Fairness – ensure there’s no bias
- Transparency – be clear on what the AI solution is doing
- Explainability – ensure it’s clear how and why AI has come to a certain conclusion
- Accountability – embed human oversight to manage risk
- Security – bake in cybersecurity
- Privacy – ensure compliance with data privacy regulations
- Sustainability – limit the environmental impact of your AI
- Data integrity – build in strong governance and data quality measures
- Reliability – ensure a high level of performance from your systems
- Safety – safeguard against harm to people or property
3. Think about the impact on your workforce
How do your people feel about AI? Are they embracing it as a tool that augments their abilities. Or are they concerned about what it means for their job prospects?
You need them to lean towards the former. How do you do that? It’s about communication – having a clear and compelling story around AI. And it’s about retraining and upskilling. Through a culture of continuous learning, workshops and building awareness, you can embed AI into everyday working life.
That upskilling will be vital because AI will change the way we work and what we do. That goes for everyone - whether it's your IT, finance or even your sales and procurement functions.
While you’re considering the skills you’ll need, you also need to look at AI’s impact on role design. If you’re going to harness AI to increase productivity, you need people with the right skills in the right jobs. Be honest with yourself – what is this going to mean for your people and how are you going to manage the change?
4. Put in place the tech foundations to enable AI
Do you have the tech foundations in place to harness AI?
We’re not saying stop experimenting with AI while you sort out your infrastructure and platforms. We say start with the easy wins. Jumpstart your AI future with rapid sprints that establish the value of your priority use cases. But while you’re doing that don’t neglect the platforms and processes that need transforming to reap the full value of AI.
Start by establishing a baseline. Where is your technology infrastructure today? Where do you want it to go with AI? And what are the gaps you need to close? Take the opportunity to look under the hood and see what you already do well and what you need to improve. It will save you time, money, and a lot of headaches.
Given that AI is only as good as the data you feed it, prioritise putting in place robust data management practices. Being confident in your data will help address some of those trust issues we’ve mentioned – and it will mean better insights and results.
There’s a big question here of whether you build or buy. We’ll talk about that in-depth another time. But even if you’re working with partners who give you access to the latest tech, it pays to keep yourself up to date with the rapid advancements in AI.