Automation, from robotic process automation to artificial intelligence, is transforming every function of every business in every industry. In fact, according to research from PWC, AI’s impact on business will be greater than the internet. The potential applications are limitless, from individualized customer marketing, to employee screening and selection, to smarter products that collect data, to automated customer support. AI has begun to change organizational processes on a scale that the re-engineering movement of thirty years ago could only imagine. Leaders of businesses that don’t move quickly to capitalize on the power of AI will be left behind.
Despite the many indicators of a transforming marketplace, almost all legacy leaders and board members still hesitate to apply artificial intelligence to corporate strategy. Perhaps wondering whether machines are beginning to complete with high priced-strategy consultants. The answer is yes. In fact, no consulting team, no matter how big, how skilled or how expensive, gather data, analyze it, and create recommendations with the speed and scale of machines. Board members and leaders who don’t believe this can simply look to see the evolution of AI powered marketing, sales and customer support. Adopting an AI powered strategy is the natural next step. No matter the application, the process is similar. The four steps of AI powered strategy:
1. Data
Creating an AI powered strategy is all about using machines and data science to chart a better and more valuable course, as opposed to using people and spreadsheets. The key ingredient is obviously data, and in this case we’re talking about data relevant to corporate strategy. That includes traditional data like financial reports and stock performance, and also alternative data, which can take many forms. Key topics for alternative data include customer sentiment, employee satisfaction, leadership capabilities, digital readiness and many more.
It’s important to recognize that to get the most out of an AI powered strategy initiative, you need to look beyond your industry peer group to consider at all top performers. Some of the most innovative strategies are best found among today’s unicorn startups that are applying modern business model principles such as AI powered platforms and multi-sided revenue models. Given that companies are crossing industry boundaries more frequently, an industry approach is far too narrow.
Some of this data is publicly available, some is created and owned by the firms themselves, and some can be purchased—data brokers are popping up all over the place. The key questions leaders should ask are:
- What metrics are more important for our success?
- What investments do we believe make a difference in our trajectory?
- What are the unmeasured, intangible items we want to understand?
2. Analysis
Once you have the data relevant to your strategic aims and your hypotheses about what really matters, you can start your machine learning journey. Unfortunately, machines aren’t self-starters yet. This means you need some smart humans, to teach the smart machines how to think about strategy problems. The competition for top machine learning talent is stiff, but remember that you don’t really need a Ph.D.-level scientist for most machine learning applications. There are a plethora of off-the-shelf tools that a good developer with some relevant experience can apply to your data and problems.
The goal is to begin uncovering the relationship between the data you have, and the outcomes you wish to track. Remember that it’s essential to bring a point of view to your artificial intelligence projects. You don’t want the team to be looking under every rock in hopes of finding insight, but instead to be validating and supporting what you believe to be true. The key questions you should ask are:
- How will I position a machine learning team organizationally?
- What are the key beliefs we would like to validate?
3. Prediction
Once your team has begun creating algorithms that reflect your strategy beliefs (or if proven wrong, algorithms that reflect your updated understanding), you will have a new understanding of what is really driving success. Perhaps you might uncover a relationship between employee and customer satisfaction, or between research and development and revenue growth. Whatever it is, before you can act on this insight, you will want to make sure it is not just descriptive, but also predictive. That is, you want to make sure you are doing more than just describing how things stand now—you want to be sure that your insight can actually help your organization chart its future.
A good way to do this is to examine historical data and see if it does a good job of explaining “what happened next.” For example, at AIMatters we examine how organization’s investment in business models affects their stock performance over future years. This proof point can help you push your machine learning past “interesting,” and into “useful.” The key questions to ask are:
- What does our algorithm tell us is important for strategy?
- Do our new insights help us predict the future?
- Does this insight apply to other companies than our own?
4. Recommendation
Once you are convinced of the predictive power of your machine learning, you can begin to derive recommendations. Transforming products, services or processes is never going to be an easy, overnight task but it does help to have some direction. The best machine learning applications for strategy will indicate clear recommendations based on their algorithms. What changes lead to what results and in what timeframes. Often there are some quick wins—short-term priorities—that will help demonstrate value and gain buy-in for bigger AI projects. Further, unlike consultants, AI powered strategy should be able to predict the quantifiable impact of recommendations based on thousands of data points. To ensure that your autonomous AI strategy agent is doing her job, ask the following questions:
- Are we identifying processes that can be optimized, relatively inexpensively?
- Which projects that offer great returns but require more investment?
- Does our AI quantify the impact of changes we could make?
Once you evaluate all the alternative moves/recommendations that are available to you, and you have weighed the cost/benefit of each, it is time to move onto execution – that’s right, getting done what the machines recommend. Think of it like a GPS – the machines can only recommend routes, but for the time being, you have to do the driving!
Adopting AI is all about people
For all those companies that aren’t Apple, Amazon, Uber or Airbnb, already AI and data powerhouses, adopting AI to power strategy is likely to be real challenge. Therefore, leaders and board members need to consider their own roles in its success. Will the leadership team commit to understanding the technology? To supporting a transformative team in the face of resistance? To funding a machine powered strategy?
Research shows that most leaders are still wary of AI, while simultaneously being afraid of its impact. Now is the time to get started and adapt to these realities. Waiting much longer might leave your company looking a lot like the yellow cab companies—too far behind to ever catch up.
