-->

KMWorld 2024 Is Nov. 18-21 in Washington, DC. Register now for $100 off!

Trees, chains, and brains

Article Featured Image

The advances in AI keep coming. Each new release raises the bar—and the hype—to ever-higher levels. As a KM’er, you might be wondering: What are the questions most asked of AI? More importantly, are they quality questions? Are they questions that actually help in planning for a future that’s changing faster and growing more complex with each passing day? And perhaps most important, do they help us make better choices?

When I asked what question it was asked the most, ChatGPT replied that the number-one question was, “Can you help me with [specific problem or topic]?” In second place was, “How do I improve my productivity?” And among the top four, as you might expect given today’s politically charged discourse, “What’s your opinion on [current event or controversial topic]?” It’s interesting that we’ ve gotten to the point where millions would actually ask a computer algorithm for its “opinion.”

Today’s AI has many different flavors and architectures, along with massive amounts of memory and processing capacity. We could probably make better use of this computational power by looking at how we can improve the quality of our queries and, as a result, make better quality decisions. Retrieval-augmented generation (RAG) is a step in the right direction.

Since Day 1, KM has engaged in a never-ending quest for better, faster search and retrieval and obtaining the right information at the right time. Yet we must keep reminding ourselves that all the information in the world is of little value unless we know how to properly act on it. But according to McKinsey (“Untangling Your Organization’s Decision Making,” McKinsey Quarterly, June 2017; mckinsey.com/capabilities/people-and-organizational-performance/our-insights/untangling-your-organizations-decision-making), approximately half of all business decisions made by organizations turn out to be poor decisions. Given that’s the case, we might first try to figure out why, despite all the technology advancements, we continue to make so many poor decisions.

To do that, let’s draw from our vast body of KM experience, put together a simple framework of different decision-making approaches, and think about how they could be enhanced by applying KM. And just for fun, let’s give it a catchy title. Borrowing from the movie Planes, Trains, and Automobiles, let’s call it “Trees, Chains, and Brains.”

Trees

We’re talking decision trees, which have been around since the early 1960s. They’re a useful tool in which the decision maker outlines the choices and possible outcomes, along with probability estimates for each outcome. For example, you’re planning a weekend outdoor sports tournament, with ticket sales, concessions, and both players and spectators travel- ing many miles to attend. But there’s a hurricane forming off the coast, with the National Weather Service assigning a 30% probability of its passing through the area at the time of the tournament. If it hits, the event is a washout, and you incur a loss from wasted advertising expenses, ticket refunds, and the like. If it misses, the weekend is a huge success. Do you postpone/ cancel, or go ahead as originally planned?

KMWorld Covers
Free
for qualified subscribers
Subscribe Now Current Issue Past Issues