Right now, everyone's gotten into genAI, and that includes organizations. But the three defaults are ignore, ban, and "all hands". These aren't viable or backed by the science. What is?
Excellent thought provoking piece. In my experience working with dozens of companies evaluating how, when and where to use GenAI, once we get past the "will it work? can I trust it? will it embarrass me?", we focus on the transaction economics like any tech implementation: What KPI are you try to change? What is it now? What do you want it to be? By when? and then do the math. Does your idea of using GenAI help? does it help enough? If you have 60,000 customer service calls a week at $13.75 per call, and GenAI can deliver quality and reduce that by 15%, it's a simple business decision to do give it a shot. It is actually much much harder to get the money to invest, the people to build and test, and the adapt adjacent processes, get them all on-board at the same time. Then you have to pick and integrate the right tech solution into your current systems. Many ideas presented are just bad ideas because you do something new that is not actually really needed and does not impact above KPIs: "who thought this was a good idea?" (the committee, not the end users) or "great but that only saved me 3 minutes this week." (Not interesting enough).
Appreciate that and agree, Mike. And the KPI complexity is precisely what makes Nelson and Don's suped-up A3 process really powerful. It makes sure you're asking the right question and selecting the right problem before getting to work.
Great post. This maps to much of my thinking, and in fact I'm noodling on a similar post right now. Your point about thin margins in small enterprises is an underappreciated one. Many of the salivating venture capitalists and technologists who think that AI is going to take over everything all at once don't really seem to understand how non-tech non-venture scale companies operate or the constraints under which they operate. Hilarity ensues when startups try to sell into these companies.
The smartest operators I have seen in this space pursue AI applications in one industry vertical. They get close to their customers, they intimately learn these customers' problems and pain points, and they build an AI solution to solve that very specific problem. AI tech is, in theory, a very generlizable technology but it is not yet generalizably plug and play like, say, electricity.
Glad you appreciated it, Dave. And couldn't agree with you more on the smart operators (i.e., AI tech vendors) bit. That takes ruthless insistence on superb, immediate customer value, doing things that don't scale, and often keeping AI out of the solution to start.
It is hard for large orgs to get their processes changed around the current speed of development around some of these tools. As soon as you have started the standard process to consider whether to start investigating if this is a tool you want to consider seriously, there’s a new leader in the field.
I've created an AI narration of this article, if that's OK?
Excellent thought provoking piece. In my experience working with dozens of companies evaluating how, when and where to use GenAI, once we get past the "will it work? can I trust it? will it embarrass me?", we focus on the transaction economics like any tech implementation: What KPI are you try to change? What is it now? What do you want it to be? By when? and then do the math. Does your idea of using GenAI help? does it help enough? If you have 60,000 customer service calls a week at $13.75 per call, and GenAI can deliver quality and reduce that by 15%, it's a simple business decision to do give it a shot. It is actually much much harder to get the money to invest, the people to build and test, and the adapt adjacent processes, get them all on-board at the same time. Then you have to pick and integrate the right tech solution into your current systems. Many ideas presented are just bad ideas because you do something new that is not actually really needed and does not impact above KPIs: "who thought this was a good idea?" (the committee, not the end users) or "great but that only saved me 3 minutes this week." (Not interesting enough).
Appreciate that and agree, Mike. And the KPI complexity is precisely what makes Nelson and Don's suped-up A3 process really powerful. It makes sure you're asking the right question and selecting the right problem before getting to work.
Great post. This maps to much of my thinking, and in fact I'm noodling on a similar post right now. Your point about thin margins in small enterprises is an underappreciated one. Many of the salivating venture capitalists and technologists who think that AI is going to take over everything all at once don't really seem to understand how non-tech non-venture scale companies operate or the constraints under which they operate. Hilarity ensues when startups try to sell into these companies.
The smartest operators I have seen in this space pursue AI applications in one industry vertical. They get close to their customers, they intimately learn these customers' problems and pain points, and they build an AI solution to solve that very specific problem. AI tech is, in theory, a very generlizable technology but it is not yet generalizably plug and play like, say, electricity.
Glad you appreciated it, Dave. And couldn't agree with you more on the smart operators (i.e., AI tech vendors) bit. That takes ruthless insistence on superb, immediate customer value, doing things that don't scale, and often keeping AI out of the solution to start.
Great piece thanks Matt
It is hard for large orgs to get their processes changed around the current speed of development around some of these tools. As soon as you have started the standard process to consider whether to start investigating if this is a tool you want to consider seriously, there’s a new leader in the field.
I've created an AI narration of this article, if that's OK?
https://askwhocastsai.substack.com/p/how-to-handle-genai-in-organizations