The disruption brought about by AI is making many organizations rethink the way they have been operating. Questions such as ‘where to start’ and ‘what areas to change first’ are commonplace if they are contemplating an enterprise-level shift.
“Every enterprise, every business, is asking that set of questions, and it became pretty clear to us that there wasn’t a disciplined way to answer them,” said China Widener, Deloitte vice chair and U.S. technology, media and telecommunications sector leader. “It varied by stakeholder. It varied by maybe even leader. It might have even varied by a priority in a given moment.”
“And so the question became, ‘how do you create a disciplined approach that allows you to answer that core question: What should I be focused on identifying in my organization if I am looking to make an enterprise-level shift?'” she said in an interview with The AI Innovator. “We believe that the market needed a structured answer that would improve its confidence that it was making good choices.”
The result is Deloitte’s Enterprise AI Navigator, which it built to assist consultants on the ground as they help clients make detailed changes tailored specifically to their context and data while benchmarking them against the industry standard.
To build the tool, the consulting firm first looked at what factors were key to making sound decisions. The first was depth of industry knowledge. “It’s hard to talk about what you should do if you don’t have a true, deep, working knowledge of what it takes to run a particular business in a particular industry,” Widener said.
Deloitte takes that knowledge and couples it with the firm’s data to advise clients on which “patch of the forest you should be paying attention to, and with some exchange of information, we can talk about which tree in the forest you should climb in order to think about agentification in your enterprise,” she added.
In practice, the tool takes a look at a company’s decisions in areas such as marketing workflows and benchmarks them to prevailing industry knowledge. “Every client has the opportunity to have all the best thinking at scale all the time, because it’s now embedded right in this tool,” she said.
Decision-making must be looked at in a nuanced fashion, Widener said. For example, for a cost-conscious company, it’s not as simple as finding which action generates the most savings.
“You might be able to generate a lot of savings, but it might be hard to do,” Widener said. On the other hand, an action might be easy to deploy but saves less. “What you want is the intersection of both those things – what to prioritize and then to understand the potential implication downstream.”
Widener said it’s also important to consider the future impact of enterprise-wide actions before making a full commitment. That’s one reason why many pilots do not make it to production. “You actually need to think about, before you get there, what’s the impact (of the project and) whether it’s worth scaling — let alone the complexity required to actually scale.”
The final step is to actually build or buy the system. “They might be acquired off-the-shelf, but you have to actually make a determination consistent with your technical architecture: What makes sense to you as an organization? What commitments have you already made? How does it fit into the architectural flow or your future-state architecture?” she asked.
Deloitte’s Navigator tool helps clients move through these phases of decision-making in a “disciplined and consistent way,” Widener said. “If you’re a heterogeneous organization, you may have multiple CIOs, not just a single CIO. You have multiple stakeholders,” she added. “Everybody comes through the same process of analysis relative to their idea. That breeds a level of consistency.”
Knowing what to fix – and how to fix it – is critical to get companies out of AI pilot limbo. “We know that almost 85% of organizations are investing more in AI but only 20% are really seeing transformation today,” she said. “Transformation is critical if you’re going to get real value out of AI.”
Be First to Comment