Executives and workforces have long hoped that if they buy the right technology, that technology will solve their business problems. This hope has manifested itself differently during three different phases of evolution:

• 1950s to 1980s: This (mainframe/mini/personal/networked) computer is going to solve my company’s problem.
• 1990s to 2010s: This (mainframe/web/desktop/mobile) app is going to solve my company’s problem.
• 2020s to present: This AI (narrowly defined as deep learning, generative AI, etc.) is going to solve my company’s problem on its own.

The hopeful buying cycles continue.

Of course automation has helped, in terms of the immediate goals of the automation. But the issue with automation has always been that we automate and improve in some ways, but then we suffer the consequences of automation. 

For example, we are a more mobile society than we’ve ever been. That’s empowering. But the consequences of that mobility have led to global warming, climate volatility, and the shrinking diversity of the biosphere. And they don’t end there.

Digitization and the need for executives to understand data and architecture

With digitization, each business entity can articulate and adaptively control its digitized environment. Consistently applied control, however, requires an understanding of architecture, who owns and controls what elements of that architecture, and how that architecture changes. 

In the AI era, data is the means of control. Data, as I’ve pointed out before in these pages, plays the following role in an AI architecture that incorporates a statistical black box:

• The Nature of the Black Box: Statistically oriented machine learning operates as a neural network that does not inherently explain its internal reasoning—it simply maps an input to an output. This black box is probabilistic in nature; on its own, it’s never precise.
• Why Input Dictates Intelligence (Garbage In, Garbage Out): Because the model’s inner workings are probabilistic and opaque, the quality and accuracy of the output are fundamentally constrained and determined by the input data fed into it.
• The Role of Machine-Readable Knowledge in the Form of Meaningful Context: Knowledge graphs address this core data quality problem by providing structured, semantic context (a digital map of entities and relationships). By improving and grounding the input data layer with shared rules and semantic description logic (deterministic and fact-based in its nature), graphs that link scalably and contextually will directly enhance the reliability and performance of the AI system.

For this reason, business executives need to be hands-on when building and managing the knowledge and data foundation for business AI. That’s when therapy comes in.

Why executives need data therapy

Therapy is work, and a lot of patients find the work itself too difficult. So the therapy may be very effective, but is the person in the state where they can actually do the work involved in the therapy?

— Anthony Rostain, Chair and Chief of the Department of Psychiatry and Behavioral Health, Cooper Medical School of Rowan University

When I say business executives need data therapy, I mean they need to reevaluate how they think about data. They need to think through their flawed and incomplete assumptions about data. 

To help with this, they need to sit down with a therapist who can understand data, because data in all its varieties is critical to business success. The therapist can act as a third party, a more objective sort who will identify the misplaced beliefs and supplant those beliefs with grounded realities.

I don’t mean that a data therapist needs to be a licensed provider such as a psychiatrist. Who do you really turn to?

Psychiatry and behavioral health professor Anthony Rostain pointed out in a New York Times interview in September 2026 that “…the field of psychiatry and mental health needs a lot of allies, OK? We need everybody — teachers, parents, people in primary care — to really ask the fundamental question: What’s going to make a difference for this individual, to get them out of the crisis they’re in, to help them function better, to help them feel more hopeful about the future?”

Who’s going to make a difference for you when it comes to your own personal success as an AI-savvy busienss leader? As Rostain notes, it depends. But one thing is clear: if that “therapist” or other advisor doesn’t understand your company’s entire data landscape and how to manage it as one cohesive whole, including the contextualizing knowledge that brings it all together, they won’t be giving you the advice you need.

Without a holistic view of your entire data footprint and a better means of managing it, your company won’t be able to control the agents it needs for its next phase of automation.

Taking control of your disparate business data

When it comes to data, business leaders have indeed reached the crisis stage because of AI. Agents need the best input, guidance and guardrails they can get. 

Control and ownership of both unstructured and structured data have never been more important. Understanding the data layer will help you feel the need for that ownership and control. 

Executives need to broaden their data thinking. Data isn’t just grist for the mill. It isn’t just tabular, numerical data. It’s every kind of digitized content relevant to the business.

It isn’t just instance data, or application exhaust, or a generated log file that has seen its day. It’s digitized bits that can contain shared meaning in the form of metadata.  Your data landscape should be scalably and precisely accessible, which is why semantic knowledge graphs matter.

For More Information:

Morrison, Alan. 2026. “How Knowledge Graphs Improve AI.” The GraphRAG Curator (blog), March 18, 2026. https://graphrag.info/2026/03/18/how-knowledge-graphs-improve-ai/.

Wallace-Wells, David, Rachael Bedard, and Anthony Rostain. “Why We Can’t Stop Arguing About S.S.R.I.s: What’s Missing in the National Conversation on Mental Health Medication.” The New York Times, September 16, 2026. https://www.nytimes.com/2026/09/16/opinion/ssris-mental-health-depression.html.