AI – So, What Now?

Católica-Lisbon SBE
Wednesday, July 1, 2026 - 17:00

Everyone is talking about Artificial Intelligence. At conferences, in newspapers, on television, and in hallway conversations. Most business leaders already understand that it is something important, perhaps the most important technological development of the past few decades. The challenge is knowing what comes next.

Many people have started using ChatGPT without first learning how to write an effective prompt. Unsurprisingly, the results often fall short of expectations. Disappointed, they feel as though they are missing the train without even knowing where the station is.

The right question is no longer, “Is AI important?” The real question, and a far more difficult one, is: Where, within my company, can AI generate real, measurable value?

The most common mistake is to start with the technology itself. Just as you should never start building a business by focusing on the solution rather than the problem, you should not begin here with the tool. Those who ask, “What can this

tool do?” often end up wasting time and money chasing promises. The right approach is to ask the opposite question: Where are we wasting time, spending too much money, or delivering lower quality than we would like? Only after identifying those pain points does it make sense to ask whether AI can help solve them.

To identify the best opportunities, I recommend a simple framework. Look for processes that meet three criteria: they are repetitive and high-volume; they depend on creating or transforming text, data, or images; and errors or delays are costly. Wherever these three conditions overlap, the greatest impact is usually found. But don’t conduct this assessment alone. The people who perform these tasks every day know better than anyone where inefficiencies exist.

In my experience, there are three areas where almost every company can begin.

The first is customer service and support. Answering frequently asked questions, triaging requests, and providing initial responses to emails. AI reduces customer waiting times while freeing up staff to focus on the cases that genuinely require human intervention. Success can be measured by faster response times and a higher number of requests resolved each day.

The second is marketing and content creation. Social media posts, email campaigns, and website copy are among the areas where these tools can provide the greatest assistance, dramatically reducing both the cost and the time required to produce content, while still relying on human review to ensure quality and preserve the brand’s voice.

The third is internal knowledge management. Preserving, organizing, and classifying company emails and documents to prevent valuable information from being lost or difficult to retrieve, preparing proposals, drafting reports, and summarizing meetings. Here, the benefits are measured in hours saved per employee each week and in the increased speed with which customer requests can be addressed.

There is, however, one important trap to avoid: impact that is not measured is not impact. It is merely opinion. And opinions are not enough to justify investment. The right way to move forward is the same approach I have advocated for years when validating any new business: conduct small, fast, and inexpensive experiments. Select one process. Measure its current performance: how long it takes, how much it costs, and the quality of the outcome. Run a pilot project lasting just a few weeks. Measure the results. Compare the before and after. Without these figures, you will never know whether it is worth moving forward.

Today, the real risk is not experimenting and discovering that a pilot project failed. Experimentation is inexpensive and highly educational. The true risk is standing still, waiting for complete certainty while competitors continue learning, improving, and building competitive advantages.

So, my invitation is a practical one. This week, choose one process within your organization that meets the three criteria. Map the sequence of tasks that are actually being performed, which are often quite different from those described in the Quality Manual, and collect real data on time and cost. Launch your first pilot project. That is how you begin turning curiosity into tangible results.

António Lucena de Faria, Professor at Católica-Lisbon SBE