How to identify the best AI opportunities for your business

Many companies today feel the urge to develop an artificial intelligence (AI) strategy. However, they often face a dilemma: how to hire the talent and identify the opportunities?

From my experience, there’s no need to hire new, fully trained talent from scratch to test the AI ​​waters. It may be better to work with in-house experts or, if necessary, a specialized consultancy to seize the new opportunities and determine which projects are feasible and which new resources are needed. Identifying the potential areas for improvement within the company, i.e. the benefits that AI could bring, takes precedence over the entire hiring step.

Hiring is a complex process and hiring decisions often have long-term consequences. Before proceeding, the company should be familiar with the level of innovation being introduced.

Identify opportunities for AI

The first step in AI adoption is identifying the initial data science application to deploy: the starter project. Its success or failure and the effort required will give management the information it needs to decide if and how to proceed with the introduction of more AI-driven applications. This project will mark the turning point for the AI ​​adoption process.

So, how can we choose it? Of the many possible AI projects, how can we determine which use case will deliver the most benefits with the least investment of time and resources?

This can become an overwhelming step in the AI ​​adoption process. There could be ideas all over the business about what use cases could benefit from AI; however, a clear understanding of the difficulty of implementation and the resulting benefits is often lacking. An idea may sound brilliant, but the technical counterpart seems too difficult for non-technical eyes, enough to curb all enthusiasm. On the other hand, an idea may sound trivial to the engineers but only benefits a niche area of ​​the business. There are also ideas whose implementation would be prohibitively expensive for the size and budget of the company. And so on. It can be daunting trying to get your bearings in the jungle of potential AI use cases.

Fortunately, there are guidelines. Best practices in this area, as described in “How to identify the best AI opportunities for your business”, often contain two steps:

• Identification within the company of a number of possible AI use cases
• Selecting the highest priority use cases in terms of budget, technical skills and benefits

Giving a workshop “Dreaming and evaluating”.

The rules are there. What’s stopping us?

Even though we know what’s important when it comes to inventorying AI projects and prioritizing, that doesn’t translate into a clear set of steps. Often we need the help of our technical staff to properly assess the difficulty of each idea, as well as the help of business experts to quantify its benefits. Once again we are standing still.

When I find myself in such a situation, I decide to have a short brainstorming workshop, inviting people from different departments from both sides of the company: the technical staff and the business experts.

The workshop is usually divided into two sessions:

• Wild Dreams: The first session includes the creative part of the workshop; it’s there to stimulate ideas, no matter how crazy. Each participant must come with an open mind and contribute at least one AI project idea. No judgment is allowed. In this session we are free to dream. All objections such as “But we don’t have the talent in-house?” or “But this is too hard to do…” should be stopped, because they don’t belong here. During this session, the imagination can and must run free.

• Coming to reality: The second session brings us back to reality. Here we stop dreaming and evaluate all ideas from the first session for feasibility, budget and return on investment.

While the previous session was the place for creative skills to shine, this session is the kingdom of prudence. Usually, after reviewing each idea from the previous session, after listing its costs and benefits and feasibility, a very clear list of prioritized projects emerges.

Make the cut

At the top of the list are the low-hanging fruit: those projects that require little work, little budget and often only internal skills for a valuable return. Then if we go down the list, we will find more complex projects, which are likely to provide a higher return on investment and sometimes require external data skills or even new hires.

Where to draw the line for the next projects to be implemented depends on management’s expectations. For a proof of concept, we stop at the first one or two projects from the list. For a broader breadth of innovation, consider a few more projects further down the list. If the introduction of AI is already approved and strongly supported by senior management, we can even start one of the more complex projects, which may require new employees.

After this “Dreaming and Evaluating” workshop, goals can become clear and the roll-out of the starter projects can be set in motion.

Workshop duration varies depending on company size and AI ambitions; it can take a few days for larger companies or just a day for smaller ones. However, at the end of the workshop, a very clear vision of AI opportunities emerges. This excites all the participants, both the dreamers and the prudent souls. For them, the path to AI adoption becomes clear and its implementation can go quite smoothly. The time investment is worth the result!

Rosaria Silipo is head of data science evangelism at KNIME.