We hear a lot about AI transformation — but how do you actually implement it?
That’s a struggle many businesses are facing. While everyone’s well-versed on the benefits of AI digital transformation, like trading costs and repetition for better efficiency and personalized experiences, companies face frustrations with the “practical” part, or “moving the needle.”
McKinsey researchers wanted to know how to fill in the blanks in putting together the systems, people, and processes necessary to leverage AI’s benefits most effectively.
So they wrote a book called Rewired: The McKinsey Guide to Outcompeting in the Age of Digital and AI. Here are some highlights from a recent podcast, hosted by Roberta Fusaro, discussing the book with McKinsey authors Eric Lamarre, Kate Smaje, and Rodney Zemmel.
Digital transformation isn’t a project to be completed, it’s ongoing
“For companies feeling stuck, some pixie dust probably sounds pretty good. But companies can’t just wave a magic wand and end up with architecture overnight…What really hobbles companies is the lack of focus on longer-term measures and the focus on value. A company has to measure its digital and AI transformation with the same degree of rigor that you would measure any cost or revenue transformation.”
- Roberta Fusaro and Rodney Zemmel
Finding the right talent requires evolved recruitment practices
“Who can I upskill? Who could really get to a different level on this? How do I take some of the capabilities I have and maybe make them more technology-enabled without having to throw the baby out with the bathwater?
- Roberta Fusaro
“Rethinking some elements of talent management can help, such as how you remap career pathways, for example. How you think about compensation models, and really rewarding skills versus tenure, or time-in-role and so on.”
- Kate Smaje
Data democratization is key
“At McKinsey, we find that the solutions we develop for our insurance clients have a more than 50 percent overlap with the code base that we use with our mining clients. There are big components that are fully reusable. How do you make sure that when you’ve developed your model, it’s stable? That, as the world changes or the data changes, it’s not doing crazy things. How do you make sure that it continues to be valid?”
- Rodney Zemmel
“This is where data products come in. Data product is that small, cross-functional team, optimized for data. But their job is to curate customer data product, if that’s what we focus on, or operations data product or supply chain data product… it helps really accelerate the deployment of business intelligence, or even AI models, because now I have made it possible for everybody in the organization to consume something where the data is well-structured.
- Eric Lamarre
You can’t just throw money at it
“A simple test question for whether companies are on the right side of how to do this is to ask, “Who’s responsible for adoption?” If the answer to that is the digital team or the chief digital officer, that’s the wrong answer. Adoption needs to be owned by the business owner of that area.”
- Rodney Zemmel
“So you’ve got to put yourself in the mindset of, “How is this technology going to be used?” Because, ultimately, if it can’t be used differently and you can’t change a business model around it, then it’s just technology.”
- Kate Smaje
Read the whole podcast transcript here.
