KPIs are industry-specific and should be aligned carefully with your AI strategy. My course at UC Berkeley drills down heavily on how to define success when implementing your AI strategy, and measurement, like anything else, is the top priority.

One technique that can potentially be used in your organization as you embark on your AI strategy is to use the SMART method for KPIs. SMART stands for Specific, Measurable, Attainable, Relevant, and Time-Sensitive. It’s essential to understand the differences between a goal and a KPI, which measures the success and milestones and not the outcome or goal attained. It’s critical to rank the KPIs in terms of importance and be focus on a small subset of KPIs that will align to your most important outcomes. If your focus is too broad, then chances are your AI strategy portfolio can get out of balance. I’ll share what I mean about the portfolio shortly.

Let’s take a SaaS company that wants to reduce churn, for example. The goal is to implement AI in a way to reduce the churn rate by identifying a plan. Initially, let’s be “specific” and say we want to reduce churn by 10% over the next QTR. Now that the specific goal has been identified, we can focus on measurement. One KPI, for example, in the initial month of the quarter, is to maximize retention rate. What can the company do to minimize churn and maximize retention in the first quarter? A brainstorming session ensues, and you focus on three KPIs for the 1st quarter. Perhaps one KPI will focus on the success of a promotion intended to a subset of subscribers initially. Based on that success, you might want to extend the same offer to an increasing number of targeted visitors, offering an annual subscription at a reduced rate 50% off. If AI is used correctly, it can help you discover and target the promotion rate, 50% or 60%, the specific price point that your subscribers will happily renew, based on the velocity of subscribers renewed in a given hour, day, week, and of course predict the percent of subscribers that will restart/renew in a given month.

The third aspect of measuring your success for this program is the fact that it was attainable. This is critical for the maturity of the data set. AI will build upon the success of your current retention rate to discover more ways to keep your clients subscribed. If you reach too far with AI and, for example, introduce 3rd party data sets, your specific goal might be threatened by various factors. Be careful of 3rd party datasets as you present them to your internal datasets before implementing the AI strategy. Another threat that could derail your AI strategy is the portfolio model. Your AI strategy is your asset basket. It’s essential to have a balance on which AI strategy is most important in your organization and rebalance after each successful implementation to build on that success. For example, it’s not prudent to implement an AI strategy across all departments of your organization until you have had many successes. So, once again, focus on small wins before taking on new implementations.

KPIs that are relevant is critical to the success of your AI strategy. Churn rate and retention rate are applicable. What was the critical success factor? Was it the offer of 50% an annual subscription? Or was it the timeliness of the message? Perhaps it was the timeliness of the message? Or was it the type of message? Did the prospect engage in chat commerce, or was it through the old-fashioned email promotion? These are all relevant KPIs that are measuring sticks for how well the AI succeeded. It will learn from and build upon for the next set of goals for your company.