CEO's AI Dashboard Series: Business Value (#2)
Welcome to a new edition of The Board: Distillation Aftershots (*).
This is an online copy of the weekly newsletter, written by Constellation's Chief Distiller and Board Advisor Esteban Kolsky, sharing curious and interesting insights and data points distilled from enterprise technology to identify what’s notable. To subscribe, click here.
In this issue, I want to start the series with the most abused term in the AI discussion: value. Many enterprises still talk about AI value as though it comes from replacing people or automating the easiest task first. The evidence is narrower than that, and more useful than that.
First, my take.
GenAI has already crossed its adoption inflection point: the open question is where it creates measurable business value. McKinsey’s 2025 global survey found that only 39% of organizations attribute any enterprise-level EBIT impact to AI, and most of those say the contribution is still below 5% of EBIT. Adoption is broad. Material economic impact is still uneven.
The organizations that get value from GenAI usually do so in bounded use cases where probabilistic output is acceptable, and the cost of being slightly wrong is manageable. Summarization, drafting, coding assistance, search enhancement, and narrow support tasks fit that pattern. McKinsey also shows that organizations that see stronger bottom-line impact are more likely to redesign workflows and assign explicit senior ownership, rather than add AI to existing work. BCG reaches a similar conclusion: only 5% of companies are achieving AI value at scale, while 60% report minimal gains in revenue and costs despite substantial investment.
Agentic AI is now entering the same phase of scrutiny. Value appears first in small, repetitive, and tightly bounded tasks. Deloitte reports that by 2027, 74% of organizations expect to use AI agents at least moderately, while only 21% say they have a mature governance model for them. Much of what is being called value today is still constrained execution under close supervision, not autonomous business performance at scale.
Scale is where the story gets harder. Governance has to hold. Talent has to adapt. The human linchpin still sits inside too many workflows. Technical debt grows when vendors use AI to extend lock-in rather than reduce complexity. That point sat underneath the linchpin argument I made in April: process economics improve when the connective layer is redesigned, not when another layer of software is dropped on top of fragmentation.
That is the business-value issue. AI creates limited value when it speeds up a task. It creates larger business value when it changes the economics of the process.
More advanced organizations are already moving away from the assumption that value comes from the largest public models. The direction is toward smaller, more specialized systems that fit the task, the process, and the context. Public models may work for narrow use cases and fast-return tasks. They are weaker as a long-term operating strategy when token economics, vendor dependence, and limited control begin to outweigh convenience.
This is why I would be careful with the claim that GenAI has broad enterprise value today. A more accurate statement is that its value is real in narrower categories than expected, and still uneven at scale. I would be equally careful with agentic AI: the current value is evident in constrained, supervised domains, while the leap from bounded execution to scalable autonomy still looks more like an ambition than a reality.
Business value needs a stricter test than adoption metrics and pilot counts. CEOs should separate usage from impact.
Usage tells you who touched the tool and how often. Value shows up when cycle time falls, handoffs shrink, decisions need less rework, cost per transaction drops, cost-to-serve improves, conversion rises, throughput per employee increases, or revenue capacity expands without equivalent labor growth.
A useful example is BBVA. Accenture reports that BBVA’s redesigned model reduced onboarding from days to minutes, improved its cost-to-income ratio to 41.7%, and helped it acquire more than 450,000 customers in Italy through its digital proposition. The point is the pattern: value appeared when the process and platform were redesigned together.
The CEO question is whether AI changes the economics of the process in a way the operating metrics can confirm. If it does not, then the enterprise is funding motion and calling it value.
Recommended CEO actions:
- Fund AI where it changes process economics, with a bias toward strategies that increase enterprise control over context, cost, and adaptation.
- Make every major AI investment prove one of four outcomes: faster cycle time, better decision quality, lower cost-to-serve, or new revenue capacity. Then track the operating evidence behind each claim: cycle-time compression, exception-rate reduction, human handoff reduction, decision rework, cost per transaction, cost-to-serve, conversion uplift, throughput per employee, and the percentage of value captured beyond the pilot.
- Focus on creating business value, not subsidizing experimentation.
Here are some reading resources:
- PwC’s 2026 AI Performance Study is useful because it says 74% of AI’s economic value is being captured by 20% of organizations, which supports your point that value is real but concentrated and uneven.
- PwC’s 29th Global CEO Survey is useful because it argues that isolated, tactical AI projects often fail to deliver measurable value and that tangible returns come from enterprise-scale deployment aligned with business strategy.
- Bain’s 2025 piece on AI in sales is useful because it says most companies still have not unlocked GenAI benefits at scale or seen meaningful gains in cost efficiency or revenue growth, which fits your distinction between activity and business value.
- Bain’s 2025 agentic AI transformation piece is useful because it reports that AI leaders are achieving 10% to 25% EBITDA gains by scaling AI across core workflows, which supports your argument that value comes from redesigning key workflows rather than layering AI onto isolated tasks.
- Harvard Business Review’s January 2025 article on process management and AI is useful because it ties AI’s value to process redesign and operating discipline, which aligns with the logic of your post.
- Harvard Business Review’s March 2026 article on the “last mile” problem is useful because it focuses on why large AI programs stall between pilots and real transformation, which supports your warning about funding motion and calling it value.
- IBM’s 2025 report on Chief AI Officers is useful because it frames AI ROI as an executive operating problem and focuses on how organizations create clearer paths from experimentation to business value.
- IBM’s 2025 report on debt-adjusted AI ROI is useful because it connects AI returns to technical debt, investment discipline, and downside risk, which strengthens your point about token economics and vendor dependence becoming a substitute for strategy.
What’s your take? We are fostering a community of executives who want to discuss these issues in depth. This newsletter is but a part of it. We welcome your feedback and look forward to engaging in these conversations.
If you are interested in exploring the full report, discussing the Board’s offering further, or have any additional questions, please contact me at [email protected], and I will be happy to connect with you.
(*) A normal distillation process produces byproducts: primary, simple ones called foreshots, and secondary, more complex and nuanced ones called aftershots. This newsletter highlights remnants from the distillation process, the “cutting room floor” elements, and shares insights to complement the monthly report.
Esteban Kolsky spent the first part of his professional career working on the customer service front line and in management roles, culminating in early implementations of first-generation CRM tools. In 2000, he became a research analyst at Gartner, focused on CRM, customer service, customer experience management, enterprise feedback management, and voice of the customer, among topics that later became the basis for the current field of CX. During his tenure at Gartner, Kolsky assisted vendors and Global 2000 organizations with their customer strategies. He left Gartner to focus his career on strategic advisory services and influenced all the major software vendors, as well as many smaller vendors, in the CX industry and advised executive leadership teams across industries and countries…...
Read morePublished
Author
