How To Talk To Your Board About AI
Jaime Meritt is Chief Product Officer at Verint, overseeing product management for the company’s market-leading CX Automation solutions.
gettyCX and technology leaders face a clear AI mandate today: Move quickly without sacrificing discipline. But the gap between AI activity and business outcomes shows a problem that many leaders aren’t acknowledging. From what I’ve seen in the industry, it may be easy to see productivity gains from AI, but it’s much harder to see how that translates into revenue gains.
Technology executives must be candid arbiters of AI, even when that means tempering boardroom enthusiasm. Their role is to provide the guidance and rigor needed to ensure AI materially contributes to enterprise value.
According to industry analyst Mark Smith, the fear of missing out that many leaders face in relation to AI needs to be refactored into AI discipline. In a recent LinkedIn post, he notes, “AI is advancing quickly across enterprise applications, technology platforms, data infrastructure and productivity tools, and the potential economic impact is significant. But urgency does not eliminate the need for discipline. In fact, the faster AI evolves, the more important it becomes for an organization to determine when to act, where to focus, who should be accountable and why an investment deserves priority.”
When I brief my own board, I focus on the key considerations required to generate a return from AI investments. The discussion often centers on outcomes, cost controls, governance and commercial safeguards.
AI must move from experimentation and pilot projects to enterprise-wide deployment with verifiable outcomes. Putting the focus on business value, IT leaders must explain how successful proofs of concept are being transformed into operational capabilities and corporate strategic plans with the necessary resources, governance, staffing and integration.
That dialogue should specify what concrete use cases are being targeted, what success will look like and the anticipated timeline for results. This can help build the organizational “muscle” of making AI a core business discipline, rather than an isolated innovation initiative. The goal is to show how each AI investment will affect revenue, costs or margins.
Every AI initiative must be tied to a quantified business case. This requires a clear, prioritized roadmap that starts with operational wins that are measurable, providing the runway to then level up to more ambitious use cases. To do this, you need defined metrics (cost to serve, NPS/CSAT, error rates, cycle times, revenue per interaction, etc.) with a clear payback period for return, ideally 12 to 24 months.
It’s essential to enact a stop-and-redirect discipline here, too. This means pilots that don’t show value should be either shut down or reshaped quickly.
This outcome-oriented discipline shifts the conversation. Stop saying, “We’re piloting multiple LLMs and experimenting with multi-agent frameworks.” Instead, tell the board, “We’re using AI to reduce average handle time by 15-20%, automate 30-40% of routine inquiries and free our employees to focus on more complex high‑value work. Here is the impact that can have on our operating margin over the next 12-24 months.”
Large, well-known companies have made headlines for outpacing their AI budgets or exhausting allocated funds for AI coding and tokens in a single quarter. Runaway usage has driven these costs upward, with one company reportedly spending $500 million on Claude licenses in one month.
The goal shouldn’t be token maximization or token minimization; it should be valuemaxxing, a term coined recently by Marc Boroditsky, the chief revenue officer of Nebius. Boroditsky says valuemaxxing focuses the conversation on where and how AI is creating business value in exchange for those tokens. It should center on a few questions:
• How much developer time was saved?
• How much modernization work was accelerated?
• How many vulnerabilities were resolved?
Under this organizing principle, IT leaders can then institute rate limits and quotas by use case or business unit. Then they can also provide guidance on who can deploy new AI-powered workflows within the organization.
IT leaders should explain to their board why model choice matters. Lower-cost models may be sufficient for simple tasks, while premium models should be reserved for use cases that require them. Boards also need to be looking at what workforce transformation is necessary to build the hybrid human-AI organization needed over the next three to five years.
The board-level AI economics discussion should answer two questions: What will this cost now and at scale? What is the expected payback?
A recent survey shows enterprise AI use is moving faster than oversight. While 87% of senior business decision-makers encourage AI agent use, only 47% have clear governance, oversight and controls in place.
Boards must balance AI enthusiasm with a disciplined and safe adoption path. Their goal should be to pursue transformative opportunities without turning the company into an uncontrolled AI experiment.
Boards are rightly worried about risk. Tech executives can build trust by staying ahead of those concerns. In doing so, these leaders must address data sovereignty and residency, security and privacy, regulatory alignment, model risk and outcome risk.
• We have governance and controls in place so we don’t violate data residency or sovereignty rules.
• We stay aligned with the regulatory regimes where we operate.
• We have a framework in place for model risk management and maintain human accountability for high‑impact decisions.
These are the elements of an outcome-driven AI strategy. By embracing these elements, leaders can turn AI enthusiasm into long-term enterprise value. That’s the reality check every board needs.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
