Zohar Bronfman is the CEO and co-founder of Pecan.aia predictive analytics platform designed to solve business problems.
Have you heard this recently from an executive in a meeting? “We really need to start using AI.”
With new AI innovations revealed every day, the pressure is on for businesses to “just use AI” to keep up and compete.
However, implementing AI solutions like predictive analytics without an intentional strategy can be disastrous. Without a thoughtful, ROI-driven approach underlying your AI strategy, your organization will waste resources and watch initiatives wither.
To select the right first steps toward using AI, I suggest applying a framework that evaluates both the feasibility and value of potential predictive AI use cases. This approach will help you focus squarely on areas where you can generate measurable value in your highest priority strategic areas.
There are three critical areas you will need to address using this approach. Let’s go.
Your main strategic priorities
One of the most crucial aspects to consider is your company’s top strategic priorities this quarter or next quarter. These are the areas where your decisions and actions can have a significant impact.
You may be aware of some operational inefficiencies that need to be addressed. There may be key KPIs that support your business strategy and goals. You may be aware of strategic initiatives that could have a major impact on business results.
All of these could represent rich opportunities for predictive AI. Collect information about each of them.
All the opportunities may seem intriguing, but there is a deeper question that needs to be answered before continuing: If you could accurately predict future outcomes using AI, would this capability enable you to significantly realistic to better execute these priorities?
What I mean by this: If you could anticipate customer activity related to this critical KPI, or if you could forecast sales associated with this strategic initiative, would you be able to make better decisions or direct the strategy in order to improve results?
Not every inefficiency, KPI, or initiative lends itself easily to predictive AI. It is essential to be realistic about the potential and answering the remaining two questions will also guide your decisions.
Data Availability for Predictive AI
No data, no AI. Data is the fuel for any effective predictive AI initiative.
Once you have identified candidate areas for strategic AI implementation, take an honest inventory of your existing data assets surrounding them.
A clue that could reveal the strong potential for success of AI: if you already use business intelligence (BI) tools generating forecasts, rules or heuristics linked to the priorities you have selected, it can be relatively simple to ‘overlay predictive machine learning models on top of it.
Machine learning models require a lot of historical data to predict the future. They evaluate your past data and identify patterns that allow them to predict future results. It is therefore essential to have sufficient relevant, reliable and relatively good quality data over a sufficiently long period. Without this quantity and quality of data, your machine learning models will struggle to generate accurate predictions. (However, increasingly, quality issues can be addressed with automated data preparation and preprocessing tools.)
Did you narrow the list down a little further? Make sure you also have the third criterion for success.
Act on predictions
Gaining insight into the future through AI-driven predictions is only part of the equation.
It is not enough to obtain insights from AI-based predictions. You also need to have a clear plan for translating this information into meaningful actions that can change future outcomes in your business’s favor.
Predictions provide limited value if they don’t shape your decisions and help you execute better. You need to know exactly how the predictions will be used in your business to ensure you are on the right track to achieving ROI from your AI projects.
When considering potential use cases, make sure you can visualize the specific process changes and operational responses that would be affected by the predictive results.
Be innovative with AI, but set the bar high
Ultimately, the key is to pursue predictive AI initiatives that provide a clear, well-defined path to positive ROI in your most strategically crucial areas. They should also be projects where you have appropriate, readily available data, as well as the operational flexibility to act on predictive insights. This approach ensures that your AI projects are not only innovative, but also practical and impactful.
While this may seem like a more complex way to “just do AI,” this practical mindset will help you achieve the true competitive advantages and ROI of using AI. This reflects discipline, focus and an overarching commitment to driving quantifiable business impact. And that’s precisely what the start of using AI should be.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Am I eligible?
