We’re deep into the AI hype-cycle, and it’s often hard to separate signal from noise. We’ve seen major advancements in generative AI across several interesting use cases.
If you’re not a tech company, you’re probably not sure how AI may benefit your business, how to quantify the value from investments in AI, and how to go about adopting AI in your company.
All too often, I’ve seen companies get caught up in hype. Under pressure to innovate, they over-invest without being pragmatic and rigorous. Instead, companies should step and test their way into a new domain based on clear expectations about the business value the new technology is expected to create.
I will not talk about data sets, language models, or AI tools. Instead, I will offer an approach for how to pragmatically cut through the noise and assess different AI solutions against your goals, and quantify their value, in order to make a decision that’s best for your company.
Below is the framework I use to cut through the noise, identify opportunities, and quantify their business value. This approach helps select the right AI solutions that will generate incremental revenue or increase efficiency, resulting in cost savings. The framework consists of two parts. The first part is to identify the problem and quantify the value, and the second part is to select the right AI solution to deliver that value.
The first set of steps identifies problems, assesses their value, and prioritizes them accordingly.
Identifying use cases
We always must start with the problem we are trying to solve, in this case, in the form of use cases. Think of various tasks or jobs in your business that can be automated using an AI solution. The use cases we identify could be customer-facing or a part of internal operations. For example, we could use an AI agent to answer customer calls for a specific set of inquiries. Alternatively, we could use AI to identify customers likely to return purchases.
Assessing the value (quantifying the opportunity)
Now that we’ve identified the use cases we are solving for, we can assess the size of the opportunity for each use case (or set of use cases if they should be solved together). Solving our use case should result in either incremental revenue or efficiency gain (cost savings). Once we’ve quantified the value of our use cases, we can then prioritize them in order of decreasing value. This will enable us to address the highest value use cases at the top of the list.
Assessing usability
In this step, we review demos and prototypes in order to understand as clearly as possible how effectively a given AI solution solves the target use case. This means that we should understand how effectively the solution delivers value to its users – is it clear, intuitive, easy to use, and does it make the user experience better? If possible, we should try to understand how users feel about it.
Evaluating technology maturity
This is the only part of the framework that is technical. Here we need a level of technical expertise to understand the arc of the technology used in a solution. We ask questions like: How well does it solve our use case (is it accurate)? How reliable is it (does it work all the time)? How does it scale (can it serve our number of customers, transactions, etc.)? Last, we need to consider where this technology is on its maturity curve and whether it’s ready for commercial application.
Understanding the cost model
The last step in the framework focuses on the cost model. We need to understand the ROI but also how the cost of the solution changes over time as the usage grows. Ideally, we want the value (the difference between the benefits and costs) to grow over time, or in the worst case, remain constant.
In order to navigate this noisy phase of AI evolution, evaluate promising solutions, and adopt the ones that will benefit our business, we need to focus on rigorously defining problems worth solving, quantifying and prioritizing the opportunities, evaluating AI solutions for their maturity, and selecting the ones that will work for us. And, as we all know, the adoption of a solution is only the beginning, as we need to continuously measure the performance, adjust as needed to maximize the value for our customers and the company.
Danijel Stankovic
Danijel Stankovic is currently Chief Digital Officer at QualDerm Partners. A digital omni-channel retail executive with a career that started in tech and later evolved to helping build and scale digital products and services, and transform organizations at some of the largest brands in retail, media and QSR in the US and Europe.
