What It Takes for AI to Deliver Real Productivity

Christine Calhoun:

What do organizations need to do to make sure their AI tools actually deliver on the promise of productivity, efficiency and better outcomes, Shishank?

Shishank Gupta:

I think the first thing that needs to happen is to really identify and begin with an objective of what are they trying to do with AI, especially in the context of business applications, where these are applications that are really running the business in many ways. So identifying goes back in a way to your vision of what areas of business that you really want to amplify with AI. And within that, it then comes down to identifying specific scenarios, use cases, and then really going all in to make an impact. What happens most times we've seen is, there's a lot of experimentation in different areas. And the experimentation seems to be in areas that is low risk, possibly also low reward. So even if it works, at the end of the day, the reward of implementing that into production is not that high. And therefore, it fizzles away. I think the key is to identify those impactful areas, which will make a huge difference to the business if AI was implemented the right way and then implemented with the right guardrails, whether it is data governance, whether it is privacy, security, any of the legal compliance aspects that come with it. And then obviously training the people and everything. I think that is the key for businesses to look at in I think the first thing that needs to happen is to really identify and begin with an objective of what are they trying to do with AI, especially in the context of business applications, where these are applications that are really running the business in many ways. So identifying goes back in a way to your vision of what areas of business that you really want to amplify with AI. And within that, it then comes down to identifying specific scenarios, use cases, and then really going all in to make an impact. What happens most times we've seen is, there's a lot of experimentation in different areas. And the experimentation seems to be in areas that is low risk, possibly also low reward. So even if it works, at the end of the day, the reward of implementing that into production is not that high. And therefore, it fizzles away. I think the key is to identify those impactful areas, which will make a huge difference to the business if AI was implemented the right way and then implemented with the right guardrails, whether it is data governance, whether it is privacy, security, any of the legal compliance aspects that come with it. And then obviously training the people and everything. I think that is the key for businesses to look at in adopting AI is identifying the right areas, which will make an impact to the business and making sure that they're implemented in true spirit and later with all the guardrails that are needed so that there is no looking back once you do it. Because as I said, even if the business case, what you achieve from the business case is slightly lesser, it’s still a significant impact to the top line and bottom line. And those are the ones that we have to do and not the ones which are easy to do, but the ones that really need to be done. So identifying the right ones is a key area in my mind.

Christine Calhoun:

Jared, how about you?

Jared Spataro:

I'd add another thing that people may not want to hear, but you have to increase your discipline and rigor. I think we're at kind of peak knowledge work inefficiency. We look around and it's unclear if we had another tool or another process or improve something if we're really improving the quality, the output at a departmental, at a company level. And so I think that means that we really need to take a step back and recognize that the industrial revolution was all about adding rigor and adding discipline to the manufacturing process. This revolution is going to be about adding discipline and adding rigor to what we think of as cognitive work. That will likely prove a little bit unpopular amongst all of us who have had a little bit freer range to roam in, but it's going to be really important. Only then with this discipline and rigor are you going to be able to measure the outcomes, the effects. Only then will you be able to kind of drive the incremental grinding improvement that you're going to need to see the results. But my prediction is, you know, welcome to the moment when all of sudden we're going to start to add discipline and rigor to knowledge work, a domain that has largely been unstructured, ad hoc, not well measured, not well managed until this time. So that's the beginning, the opening of a new door and a bit of a new era for us.