How Comcast Solved the AI Productivity Paradox

Circling back to that stat that I opened up with, it looks like McKinsey's latest survey is showing that 80% of orgs see individual productivity gains from AI, but only 37% report any earnings impact.

Why do you think that is?

I think a lot of it has to do with the planning that kind of goes around as AI is implemented and what people attempt to use the savings for.

Members of the organization begin using AI tools, begin using things like Copilot and Cloud and ChatGPT.

there's really not a plan in place for how that saved time is going to be used, right? So, you know, you save an hour here, you save an hour there, the job becomes easier, right? But I don't really think there is a plan for how do we kind of roll that back into the organization to use that saved time.

I also think part of it, too, is that the savings and the efficiencies that organizations are realizing from AI adoption, it is landing unevenly, right? You know, a few hours here and a few hours there, it scatters across many calendars.

But, you know, how do you kind of bring that together to really kind of report the savings and things like that? I think the main evidence today is that people are feeling more productive from using AI and but like we haven't seen the results yet.

And I think that's just because of the uneven application, how organizations plan to leverage the efficiencies that they gain from adopting AI.

I think part of what, you know, what could be done is, you know, as you go down these AI initiatives, as you launch these programs in your organizations, you want to define the measurement upfront.

Don't do it after the fact and try to reconcile the savings to the, to the the efficiencies put a stake in the ground say like i expect to achieve these savings i expect it to look like this and and so on and so forth and then making sure that you have a sponsor and an owner around that too to to really kind of drive that drive that narrative drive those savings across the organization so it's it's not really so much that they aren't seeing impact i think the impact is there i think it's really more of like you know how are we going to capture this how are we going to to to measure this

Agreed.

I think folks are struggling with that measurement aspect.

Do you have any examples of what that might look like, what directly tying an AI gain to a financial metric might look like?

How do you tie this back to, you know, the output that an organization can do? Because, you know, a lot of times when you're looking at the gains, the efficiency gains you're seeing from adopting AI, a lot of it is done in, you know, you know, back office type functions initially, right? Because you're going to see a lot of efficiencies there, you know, data pulls, analyses, things of that nature, right? And I think it's more of, you know, that time saved and the redeployment of that time to other tasks.

So, you know, the increase of the throughput.

You know, for things like that, it's difficult to kind of tie that to, you know, revenue gains and things like that.

But like, you know, I think what we can do is look at the throughput of the work.

I think what we can do is look at the number of things that are completed and along those lines to really kind of help drive the efficiency and the measurement of how AI is working for organizations.

Thank you for that.