AI Needs a Seat at the Business Table
Why Business-Led AI Drives ROI and Changes AI Governance
AI Needs a Seat at the Business Table
A year ago, LinkedIn posts and articles were filled with the promise of autonomous agents taking over key workflows and transforming the enterprise.
The promise of AI is still real, but the conversation has shifted. The posts now showing up in my LinkedIn feed and Substack are more cautious. They cite analyst reports, university studies, and growing concern about weak ROI, failed pilots, and organizations struggling to move from experimentation to real value.
The discussion has also shifted toward best practices around AI strategy, governance, and cybersecurity.
But there is a more fundamental issue underneath all of this.
Many organizations are still approaching AI like a technology initiative, when in reality AI is becoming a business-led capability.
That is the difference.
AI should not be introduced as a technology project looking for a use case. It should have a seat at the business table, tied to the goals, constraints, opportunities, and measurable outcomes that matter most to the organization.
When good AI solves the wrong problem
Take a simple example.
A competitor launches an AI customer support agent and gets strong headlines. The market notices. Leadership teams notice. Suddenly, there is pressure to show the organization is not falling behind.
In response, a company moves quickly. It brings together its AI and technology teams and launches an even more advanced support solution. The system can personalize responses, pull from the knowledge base, adapt to the customer, and present itself as a modern agentic service experience.
The technology works. It gets good press. Internally, the launch is viewed as a success.
But six months later, the business sees little measurable improvement in retention, service costs, customer satisfaction, or growth.
The problem was not the technology.
The problem was that customer support was not a strategic weakness or a key business priority for the company. The organization had responded to market perception, not to a business objective that actually needed solving.
That is the trap.
AI can be technically impressive, well marketed, and still strategically misplaced.
AI is not just another enterprise technology wave
Previous technology eras brought enormous change to business. Mainframes, client-server, the internet, cloud, and mobile all changed how organizations operated. They improved speed, scale, access, efficiency, and reach.
AI can do all of that too.
But AI is different in an important way. It does not just automate or accelerate work. It can increasingly participate in decisions, workflows, customer interactions, and business operations in ways that move it closer to the strategy itself.
This is why AI is business-led.
The question is no longer just, “How do we use technology to support the business?”
The question is increasingly, “How do we use AI to help the business achieve its goals, overcome its constraints, innovate faster, and create differentiated value?”
That is a very different starting point.
The stock market shows the difference
One reason I keep coming back to this idea is that you can see the contrast clearly when you look at the long arc of technology in business.
At the turn of the 20th century, “computers” were often humans. In stock markets and financial operations, people calculated prices, recorded trades, and handled the math and record keeping that kept markets moving.
Then came tabulators and early computing systems that improved speed and accuracy. Mainframes expanded the capacity of exchanges and institutions to process more activity more reliably. Networking connected more firms and markets. Client-server expanded reach. The internet, cloud, and mobile extended access and participation even further.
Those changes were massive, but for the most part they were still accelerating and supporting a human-led business process.
AI is beginning to change that equation. Today there are already market offerings that use AI to help build trading strategies, analyze signals, and automate execution. Some platforms now market end-to-end AI trading, AI trading agents, or strategy bots that can act with significant autonomy once goals and guardrails are defined. The important shift is not just speed. It is that AI is moving closer to the execution and adaptation of the work itself.
That is a different kind of technology shift.
It is one reason AI needs stronger governance and closer business alignment than many prior technology waves.
The organizations that move past pilots will start differently
Most AI strategies still start with the technology.
What model should we use?
Where can we deploy a copilot?
Which agent platform should we buy?
Those are important questions, but they are not the first questions.
The better place to start is with the business:
What objective matters most?
What constraint is slowing us down?
What opportunity are we failing to capture?
Where could AI improve results in a measurable way?
That is how organizations move beyond failed pilots and low ROI.
When AI is tied to a measurable business goal, the organization knows what success looks like. It knows why the initiative exists. It knows how to evaluate value over time. And the technology choices become aligned to the business instead of driving it.
A legacy clothing retailer is a good example.
Suppose the leadership team identifies a real business problem: they are not responding quickly enough to changes in customer demand. That weakness is hurting loyalty, margins, and inventory performance.
Now the organization has something concrete. It can define goals such as improving customer loyalty, reducing markdown pressure, and increasing margin through more dynamic merchandising.
In that case, AI is not being deployed because it is new. It is being used because the business has identified a measurable need.
An agentic AI solution might help the retailer adjust assortments faster, personalize offers more effectively, adapt pricing decisions, or better coordinate supply and demand signals. If weather patterns shift, demand changes, or regional preferences move, the AI can help the business respond more dynamically.
That is a business-led AI strategy.
The AI is not replacing the business. It is helping the business perform better against its own goals.
Why this changes AI governance
This also changes how we should think about AI governance.
Much of today’s governance discussion starts with the AI system itself: the right guardrails, the right policies, the right tools, the right controls for compliance, accuracy, and security.
Those things matter. They matter a great deal.
But if AI is a business-led capability, then AI governance has to begin at the business level too.
The first governance questions should be:
Is this AI system effective?
Is it aligned to business purpose?
Is it creating measurable value?
Is it continuing to do so over time?
That is the shift.
In the past, technology governance often began with control: is it secure, compliant, approved, and well managed?
With AI, that is not enough.
An AI system can be compliant, secure, and still ineffective. It can be well governed on paper and still fail to deliver meaningful value. It can launch successfully and then slowly drift away from the purpose it was introduced to serve.
That is why AI governance has to be more than a risk and control function.
It has to be a continuous validation function for business effectiveness.
Why governance matters more with AI
This matters because AI is not deterministic in the same way traditional software has been.
Traditional software could often be governed through testing, release management, access controls, and predictable change processes because behavior was relatively stable. If the code and inputs remained the same, the outputs were expected to remain the same.
AI is different.
Especially with generative AI and agentic workflows, outputs can vary. Behavior can shift. Context changes. Models are updated. Retrieved information changes. Prompts evolve. Agentic systems may choose different paths, different tools, or different actions over time.
Even when nothing appears broken, effectiveness can change.
That is why AI governance cannot be a one-time approval exercise. It has to measure whether the system remains effective, business aligned, and valuable over time. It also has to detect when performance improves, when it drifts, and when changing business conditions require a different approach.
This is also why cybersecurity and compliance remain so important. They are not secondary. They are essential to protecting trust, managing risk, and ensuring the AI system can operate safely at scale.
But with AI, governance should not begin and end there.
Governance must lead with whether the AI is fulfilling its business purpose, and then continuously validate that purpose over time with strong security, compliance, and operational discipline around it.
AI belongs at the strategy table
This is the real shift organizations need to make.
AI is not just another technology layer to be added to the stack.
It is a capability that can influence workflows, decisions, customer experience, operating models, and innovation itself.
That is why AI strategies have to be moored to business goals, business constraints, business opportunities, and measurable outcomes.
The organizations that continue to treat AI as a technical experiment will continue to produce pilots.
The organizations that treat AI as a business-led capability will be the ones that create real value.
Thomas Edison is remembered for the light bulb, but the broader lesson is that he understood the bulb alone was not enough. He recognized the larger set of conditions required to make it successful , the supporting infrastructure, the practical use, the adoption model, and the path to scale.
AI is similar.
The model alone is not the transformation. Organizations have to think through everything required to make AI successful against a business goal: the workflows, the data, the governance, the operating discipline, and the way value will actually be realized.
That is why AI needs a seat at the business table.