ZDNET’s key takeaways
- Companies with a growth mindset see 3x more ROI on AI investments.
- Only 7% of companies see value, workflow transformation, and organizational alignment.
- AI agent adoption is widespread, but only 17% of agents are fully autonomous.
A growth mindset distinguishes companies with strong positive returns from AI investments, but most companies continue to focus on cost savings and time saved as their primary goals for investing in AI.
Also: 90% of CFOs using AI report multiple business benefits
In fact, a strong one ROI for AI is nearly three times more likely when companies measure growth versus efficiency alone, according to a study by Wharton Human-AI Research (WHAIR) in collaboration with GBK Collective, a market research and analytics firm.
The fourth annual report, called “The Adaptive Advantage: Who’s Winning with AI and How,” found that 35% of companies that measure ROI against growth results have strong ROI versus 13% of companies that focus solely on efficiency.
The WHAIR study, based on a survey of 1,415 US enterprise leaders and mid-level employees, focused on three markers of AI success:
- Provide value (eg a department reports strong positive ROI from its AI investments to date)
- Transforming workflows
- Aligning the organization around AI
Only 7% of companies show above all three markers.
Workflow transformation requires three conditions: AI is fully embedded in core business processes, the department is currently using AI in some form, and workflows have changed significantly because of AI.
Also: Who Owns AI Risk? Business and tech leaders could disagree
Moving from AI pilots to production deployment
Now in its fourth year of reporting, WHAIR shows a shift from exploration with generative and agentic AI (2023-24) to responsible, accelerated use of AI (2025), and then to a shift from pilots to more disciplined, enterprise-level adoption of AI in 2026.
Adoption of AI is no longer the question; 80% of organizations report strong ROI on AI investments. The report notes that in 2027 there will be less work on AI than on the distance between organizations that have made significant progress in transforming workflows with AI, and those that have not.
Also: Give AI ‘Things No One Wants To Do’
A key feature of the report was around the autonomous AI agent. The report found that AI agents are widespread, but workflow transformation is not keeping pace. Sixty-seven percent of companies use AI agents to complete tasks or execute workflows, but only 28% say AI is fully embedded into core processes, and just 17% let agents act autonomously.
AI delivers business value
Among managers, 62% use AI every day, but only 28% report a strong positive ROI. When designing value creation with AI, organizations that prioritize growth (net new revenue, increased sales conversions, increased output volume) report stronger results.
The research found that 83% of managers expect AI to fundamentally transform or impact their industry within 2-5 years. However, most of these senior managers judge success based on time savings (40%) versus growth metrics such as new revenue (27%). Departments that judge AI on growth, not just time saved, report much stronger ROI: 35% (growth metrics) versus 13% (productivity metrics alone). The majority of senior leaders (52%) cite reducing costs as AI’s primary job. Only 22% of senior leaders see AI’s primary job as a growth engine for their business.
Also: Who is responsible for catching rogue AI agents? You are
Organizations with a dedicated chief AI officer (CAIO) see stronger returns on AI investments; 58% of companies with a designated CAIO report strong positive ROI. Two thirds of organizations now report a CAIO or equivalent.
Another key success factor for achieving positive ROI is standardization on a single approved AI platform. Fifty percent of organizations that standardize on a single approved platform report strong positive ROI, versus 30% that allow multiple tools under a formal policy, and only 19% that run a central hub or orchestration layer.
The separation of strong versus weak returns also points to other key success factors, including the cost of licensing and tools exceeding the cost of labor and skills, confidence in tracking AI spend versus tangible business value, and the inability of organizations to demonstrate the level of confidence in their spend measurements.
The report found that tech-forward industries (40%) and professional services industries (33%) show stronger ROI. Departmental ROI leaders include IT (41%), finance (37%), HR (35%), and legal and engineering (19%).
Workflow transformation changes how work is done
WHAIR reports that 73% of managers say workflows have changed significantly because of AI, but only 28% describe AI as fully embedded in their core business processes. The use of AI is changing management structures, with 37% reporting staff reductions, and 33% reporting slowed or frozen hiring.
However, the departments with strong returns did not cut staff more than others. Instead, they have changed their management structure and workflows, and are more likely to monitor AI usage, including performance reviews and AI skills training. Companies with strong AI ROI say AI helps new hires contribute faster.
Most departments have reported AI agents doing the work, but only a few have allowed them to act autonomously. Only 17% of companies allow AI agents to be fully autonomous: 10% review a sample of AI output after the fact, and 7% rely on real-time monitoring and systematic safeguards instead of human review.
Also: Claude sent a false murder tip to the police – and it’s part of a disturbing AI pattern
The report found that 31% of departments using agentic AI to its full capacity are achieving transformative success, and 25% of departments where AI also serves as a consultant.
The level of AI agent autonomy is based on risk tolerance and not agent capacity. The barriers are data security, regulations, mistrust of AI’s reasoning, and a lack of employee skills. Liability for AI agent errors is also a limitation. Less than 1% of departments report no formal process to monitor agent performance.
Organizational alignment scales AI adoption
Fewer than four in ten organizations are aligned to AI, according to managers. Success in this area takes into account five aspects of the work environment: whether employees receive adequate AI training, whether leadership has set a clear long-term vision, whether people have a meaningful choice about how to use AI, whether the department avoids being held back by people and processes rather than technology, and whether there is real resistance among colleagues.
The oldest business leaders are the most exposed and the most uneasy with AI. Some (7%) who describe themselves as resistant to AI are concerned about their roles, and those who have reduced staff due to AI are even more concerned about job security, skills gaps and future staff reductions.
Also: Regulators are trying to protect you from being fired by AI
Organizations that have shown strong alignment and strong ROI from AI investments empower their leaders to drive AI change, providing them with the opportunity to develop a clear long-term AI vision and responsible AI policy. Adequate employee training on AI is an area for improvement to improve organizational alignment.
The level of AI use has increased significantly by 2025, and yet the level of positive sentiment has not changed. The optimistic view of AI remains unchanged at 59%. The 2026 report is the first time that AI usage and sentiment have moved in opposite directions. This is a clear sign that organizational alignment is suboptimal. The key is the ability for organizations to be adaptable. The adoption of AI has increased across all company sizes. It’s no longer about the size of your company, but your ability to adapt and scale.
5 recommendations
The research concluded with the following recommendations:
- Complete the redesign. Fully embed AI into core workflows rather than screwing it up to a select few and moving more than half of AI pilots to production.
- Improve how work is managed rather than reducing the number of people. Monitor AI usage, incorporate it into performance reviews, and provide adequate employee training.
- Give AI agents meaningful authority. Assign AI agents consistency work and use automated monitoring instead of step-by-step human approval.
- Measure value over efficiency. Track growth results alongside hours saved and maintain visibility into expenses.
- Establish dedicated AI ownership and a standard platform. Appoint a CAIO and adopt either a single platform or a formal multi-tool policy instead of allowing open experimentation.
To learn more about the WHAIR report, You can visit here.
See full bio