Corporate AI Transformation: How to not throw mud against the wall

Organizations today are rushing to implement the buzz-infused AI and ML technologies everyone is talking about. These technologies have the ability to automate processes, recognize patterns, provide incomparable insights, and quite generally, bring innovation to almost any sector. But, as the AI frontier has grown, and become a massive buzzword, organizations are running to implement these technologies, too often without doing their homework. In fact, According to Gartner research, only 15% of AI solutions deployed by 2022 will be successful, let alone create ROI positive value.

While I am the last person who will tell you not to leverage advanced technologies for improvement, what I have been seeing recently is too many companies taking templates like engines and throwing them on top of their existing operations without taking the relevant steps beforehand. The outcome of this is usually one of two options. Either the technology will be successfully implemented, but won’t even remotely provide as much value as it could. Or, it will completely fail with companies often not even able to dissect what the root cause of this outcome is. This is especially relevant to existing organizations who are trying to now optimize, rather than new startups who are building their operation AI-Infused from the ground-up.

Rather than throwing mud (AI/ML engines) against the wall, and seeing what sticks, companies must take a more analytical approach. Rather than taking general-usage engines and throwing them on top of the existing tech stack and data set, companies should first set out on a journey of company measurement, analysis, and mapping.

Answer the “why” before the “how” — In other words, first, gather and analyze your company data. Use it to work out what’s not working well enough, or what you would like to work better, and then find the tech solution that is actually built for it (or better yet, build it yourself). Just like a startup needs to prove there is a “need” in order to define product-market fit, every aspect of a company’s AI Transformation must also be suited to the organization’s strengths, weaknesses, challenges, and bottlenecks. If you aren’t capable of identifying these, it’s time to take a step back, and first, shift your organization to one driven by data. In other words, BI (Business Intelligence) solutions must be implemented alongside protocol shifts, in order to measure performance and processes and allow management to recognize key pain points.

Thinking that you can just throw an engine in your company’s mixing bowl and move on with life is going to result in failure. While AI can save hundreds and thousands of work hours, it also requires a time (and often financial) investment. In order for the solutions to be as effective as possible, part of the core company functions must become understanding, training, maintaining, and improving these algorithms.

Moreover, in order to approach a transformation well-prepared, protocols and tools must be put in place in order to collect, gather, clean, and tag the relevant data that will be used in the process. For this to be done effectively, the company must first put together a long-term transformation strategy, enabling them to foresee what data they will need later, so they can collect it as part of day-to-day operations. Otherwise, they will be forced to implement retro-active and ineffective collections methods later on. Or even worse, not have the data when the time comes, and have no way of producing it.

AI transformation requires cultural, operational, and management shifts. Most existing companies will need to edify their staff on the subject. If you’re letting an algorithm perform certain business functions, the humans, who once performed or managed these functions, must shift to a role where they are the ones guiding the engines, nurturing them, and measuring their performance. Often, the biggest barrier to this is the language barrier. Companies must empower their employees by giving them the linguistic tools to discuss and analyze these issues in depth. This requires education, enrichment, and openness to change. If your company culture is one that already embraces these values — the process is going to be a lot easier.

More often than not, it will be senior management or an external advisor who will be pushing for this transformation. While that’s fine, in order to really optimize or contribute to functions that are being performed by humans, the company must recognize and harness the immense expertise these employees have developed in their field, giving the AI the best chance to succeed, and opening the door to bottom-up innovative initiatives.

Finally, In order for existing organizations to undergo a successful AI transformation, companies must remember that they are creating human-machine teams. This means the humans must become experts in this craft, and the company must build a long-term strategy with this in mind. While we all assume AI transformation will require a technological shift, we need to remember that it also must include a human-cultural shift.

Moreover, if companies want to lead a successful transformation, they must approach it from a place of knowledge and data superiority. In other words, the companies that will lead the best AI transformations are the ones that previously led the best BI transformations. BI’s role in the age of AI is not just to map where AI is needed, and measure performance, it’s also the safety net we have to make sure we are on the right track and are building the best future versions of our companies.