Even the technology visionaries famed for their near psychic understanding of their customers create a few loser products, but they can recover quickly (think of the late Steve Jobs and several unsuccessful Apple products). Unfortunately, most software companies don’t have the resources to take that kind of chance.
Adding incremental features and functions is not complicated. Your sales and support people hear tons of ideas to add to your software application. As a company, you understand your customer’s needs for software to operate their business.
One thing you are likely hearing about, and maybe experimenting with already, is Artificial Intelligence (AI). You and your technical staff likely see AI as something wholly new and different. Your marketing people are likely all-in on the term AI and want to use it as quickly as possible to create some market differentiation.
While it seems that Artificial Intelligence is a technological leap, AI which is useful for business processes is more of an incremental step. Consultants have been doing what is now classed as AI projects for targeted business purposes for years.
Adding AI functionality usually takes one of two forms for ISVs. Either you incorporate specific AI functions into the existing application, or you create a separate module within a suite of applications. Nothing precludes doing both as well.
Therein lies the problem. How do you know how to proceed?
The fundamental question is “can your customer benefit from machine learning and predictions using the data in your applications?” That is the essential question of AI in enterprise software purposely not using the term “AI”.
For most industries and most enterprise applications, the quick answer is – probably.
If you are indeed one of the aforementioned visionary technologists, you already understand your customers well enough to answer the question for them.
On the other hand, if you are like most software executives, you need to do some research to answer the question. You need to talk to your customers and get their opinions. Critically, you need to find out what AI functions they expect, how should they be delivered, and how much are they willing to pay.
The problem for doing quick and dirty market research into AI for your customers is the massive amount of confusion around the term AI. Sorting through that confusion in a market discovery interview requires a quick re-definition and some different terms.
We have had great success in our market discovery calls for ISV clients by quickly talking about AI to get attention in the very first part of the interview. We then promptly re-define our topic to either “predictive models for business operations”, or to “machine learning with transaction data”. For the rest of the conversation, we use either “predictive models” or “machine learning”.
Getting away from the term AI lets the customer talk about their business needs without the baggage of using the term AI.