The first wave of enterprise AI answered an important question: Can it create business value? The answer, for most organizations, is yes. The next question is far more consequential: How do you make that value repeatable across the enterprise? Findings from MathCo’s latest research with HFS Research point to a clear shift in how leading organizations are approaching that challenge.
Based on a survey of more than 100 senior AI and data leaders across CPG, Pharma, Retail, Manufacturing, and High-Tech, the study finds that enterprises are increasingly prioritizing AI that augments human expertise rather than replacing it. Their investments point toward a future where AI amplifies judgment, accelerates decision-making, and improves outcomes, even if many organizations continue to frame their AI ambitions in terms of automation.
The Strategic Shift to Process Intelligence
The survey suggests that enterprise AI is entering a new phase, one where AI priorities are driven less by the breadth of use cases and more by their business impact. Process intelligence is emerging as the primary focus of that shift. Enterprise AI portfolios are relatively balanced, with 37% of investments leaning toward process enhancement, 35% maintaining an equal focus across priorities, and 29% emphasizing people’s productivity. The distribution reflects a period of experimentation, where enterprises have explored multiple paths without committing too heavily to one.
Looking ahead, however, that balance begins to disappear. When the same leaders were asked where they intended to move, the responses pointed to a clear shift. 57% said they plan to shift further toward process enhancement, while only 25% intend to double down on productivity. This is not a gentle rebalancing. It is a directional bet, and it is being made by the people who control enterprise AI budgets.
The rationale is straightforward and, on the surface, defensible. When asked which approach delivers better returns, 51% pointed to AI-augmented processes, compared to 27% for AI-enabled productivity. With AI investments increasingly scrutinized for business impact, process improvement presents a compelling case. Reduced costs, compressed cycle times, and greater throughput create tangible outcomes that are reflected directly in financial performance.
The Process-People Paradox
This is where the findings take an interesting turn. While earlier responses reflected where AI creates the most value, leaders expressed a very different perspective when asked where organizational confidence actually lies. Only 14% said trust is higher when AI operates on a process. 45% said trust is higher when AI supports a person. Balanced or undecided views made up the rest.
That is a three-to-one gap, and it lands squarely on the fault line between what enterprises are funding and what their organizations are prepared to believe in. Process-led AI may promise a stronger return, but it is people-supporting AI that earns the benefit of the doubt from the workforce and middle management. Ultimately, it earns that same confidence from the customers and regulators watching how these systems behave.
This is the process-people paradox. Capital and ambition are moving one direction, while trust, the actual precondition for AI to be adopted at scale rather than piloted indefinitely, is moving the other.
The Missing Link Between AI and Adoption
If ROI determines where enterprises invest, trust determines whether AI is ultimately adopted. The survey makes it clear that these are not the same thing. Process-led AI may offer the strongest business case, but technical performance alone has rarely been enough to earn organizational confidence. Trust is built when people understand how decisions are made, when those decisions can be explained, and when accountability remains visible. Without those elements, even the most effective AI systems can struggle to gain acceptance beyond pilot programs.
The challenge is compounded by a gap in enterprise knowledge itself. Two-thirds of respondents say their critical business logic, decision rules, and tribal knowledge remain fragmented or undocumented. Only 26% treat that institutional context as proprietary intellectual property, and just 38% have funded initiatives to capture it. As organizations accelerate investments in process intelligence without first codifying the knowledge that gives those processes meaning, they risk creating AI systems that are efficient but opaque. Trust cannot be engineered through automation alone; it is built on context, transparency, and confidence that the system’s decisions can be understood and justified.
The Strategic Choice Ahead
Building AI that people trust requires more than technical excellence. It demands a deliberate balance between human judgment and machine intelligence, ensuring that AI enhances decision-making without replacing accountability. Explore this idea further in our article, “The Human-Machine Architecture: And Why the Ends Must Stay Human,” which examines why the future of enterprise AI depends on keeping humans at the center of critical decisions.
The research does not diminish the value of process-led AI; its business impact is well established. However, business value alone is not enough to sustain enterprise-wide adoption. Organizations that embed transparency, explainability, and human accountability into their AI initiatives will be far better positioned to earn trust and scale with confidence. For enterprise leaders, the strategic imperative is clear, optimize processes, but design for trust from the outset.
The organizations that will scale AI most successfully are unlikely to be those investing the most in process automation alone. They will be the ones that pair process intelligence with human-centered AI, using augmentation to build trust while systematically capturing the business knowledge and decision logic that make AI transparent and explainable. That is the real strategic choice facing enterprise leaders today, not automation versus augmentation, but whether trust is treated as a foundational capability or left to emerge on its own.
The findings in this article are only part of the story. To explore the emerging patterns, strategic trade-offs, and enterprise priorities shaping the next phase of AI adoption, read the Take 5 Report, developed in partnership with HFS Research: Enterprise AI Context Gap Report: Why AI Fails to Scale
