Assessing Business Value for AI Projects

Hi there,

I'm writing this newsletter from my hotel room in Mountain View, where I'm speaking at the annual Springboard Rise conference.

Silicon Valley is probably the place with the most shiny new toy syndrome. If there's a new technology out there, everyone wants to have it (and make money with it) - too often it's solutions looking for problems.

But unless you're working with VC money and trying to launch the next "big thing," you should start not from the shiny object, but from the business value - not only, but especially in AI projects.

Today, I want to share with you my strategies for assessing that business value.

Let’s start!

Business value = people + money

Estimating business value is tricky, but important. The most straightforward way to think about business value is to think in costs saved or additional revenue gained. 

And it’s true, there are AI use cases which are that “simple”:

• Recommendation engines that will make people spend X% more. That’s a top line growth right there. 

• Use AI to optimize your ads budget. That’s an immediate cost benefit. 

But there are also many more AI projects where the concept of “value” is much harder to measure:

• Support chatbots: Using an AI chatbot for customer service can free up your staff and increase customer satisfaction, but it's difficult to measure the exact financial impact.

• ChatGPT: Letting an employee use AI to write "better" business documents or prepare presentations. But "how much" better is it really?

• Lead Scoring: Implementing AI to prioritize leads can significantly enhance the efficiency of sales teams, but again – it’s hard to measure.

So why don't we just focus on the first category of use cases? Because those "simple" use cases get exhausted pretty quickly - especially for non-tech companies. The second type is just much more common.

How do we approach this?

There are several ways.

What I do is try to assess the business value along two components - the financial side and the human side. This approach has proven to work quite well - especially for the more complex use cases.

Let’s take a look:

Component 1: The People Aspect

"I finally enjoy doing my job again!" said a seasoned copywriter to me after exploring the capabilities of ChatGPT.

AI solutions not only optimize processes, but also make the work experience more enjoyable. By taking over day-to-day tasks, they allow people to focus on more creative and meaningful work, bringing back more passion and productivity.

Without a doubt, a happy and energized workforce translates into greater business value, and creates a win-win cycle of higher performance and improved bottom-line results.

And apparently, this doesn’t only span copywriters, but almost every other “knowledge” worker task as well:

A recent BCG study showed how Generative AI can be a powerful lever of performance:

The insights of this study are astounding: With the help of Generative AI, people can not only increase the speed at which they can get tasks done (efficiency), but also improve the quality of their work.

From my own experience - I can confirm this observation. And it’s notable that in most cases, the quality enhancement is not realized because the AI suggested some super intelligent thing. It’s because the AI assisted in completing the “mundane” tasks so the human can focus on high-level, quality work.

Examples?

• Consider a copywriter who - instead of starting with a blank paper - uses generative AI to “prep the page” and suggest a compelling outline, which the author can then modify and populate with their own content.

• Imagine an insurance worker who needs to determine if a damage is covered by the insurance contract. They can ask a Generative AI system about the obvious reasons why the damage may not be covered, allowing them to concentrate on the more detailed aspects.

• Think of a data analyst who uses Generative AI to create problem statements, issue trees, and analysis frameworks to make sure they follow a strong and reliable path to insights.

While it is difficult to measure the exact impact of these activities, we must not discount their importance and consider them a critical part of our value equation - in some cases, the most important part.

Component 2: The Financial Aspect

Now that we covered the “people” component, let’s see how we can gauge the financial impact of a new AI project. Here are some common frameworks to get started:

1. Return on investment (ROI) analysis: The ROI evaluates the financial gains or savings achieved by implementing an AI project compared to the costs incurred. Typically, you discount your investment over a period of time and calculate the net present value (NPV) of the project. This means that the faster you can realize financial benefits and cost savings, the higher the ROI. Check this ROI beginner article to learn more.

2. Risk assessment: Identifies and evaluates the potential risks associated with the AI project, both from a financial perspective and in terms of other impacts on the organization.

3. Benefit assessment: In addition to immediate (direct) benefits such as increased revenue or improved cost savings, the AI solution could also lead to delayed (indirect) benefits such as improved customer satisfaction, brand image, or increased market share. They give your company a strategic advantage and make it more competitive in the long run.

Which framework to look at depends not only on your use case, but also on your organization's situation and priorities. But in my experience, being able to present a compelling business case that shows a quick, positive ROI (even if it is not that high) is the fastest way to get an AI project off the ground.

Tip: Don’t do it alone

Now, before you start calculating the human and financial impact of the AI project in your head, consider the following:

We tend to overestimate the returns and underestimate the risks and costs of projects we cook up.

The best way to avoid this trap is to review your idea or business plan with a group of other people in your organization who have no direct stake in the project's success or failure.

Some companies have created what's called a "business value committee," which is a group of experts from different parts of the company who are tasked with assessing whether the proposed value of a new AI project is reasonable - even (and especially) if it's outside their immediate area of expertise.

Conclusion

In short, gauging the business value of a new AI project is difficult, but important. We must not fall in love with financial metrics - especially the ones we have come up with - and look at the bigger picture:

It's not just about hard numbers - it's about enriching the human experience.

That's why my favorite AI use cases are not the ones that just drive bottom-line KPIs, but the ones where people tell me, "This made me love my job again."

These are the transformative experiences that redefine companies and the people within them.

If you have the opportunity to pursue a use case like this - go for it!

And as always, I'm here to help. Just book a call* and I'll get you started building your first atomic AI prototype in less than 20 days and under $20k.

See you next Friday!

Tobias

*use code FREEFLOW to book the call for free

PS: If you found this newsletter useful, please leave a feedback! It would mean so much to me! ❤️