Why the ROI of Enterprise AI Still Eludes Many Firms
Despite billions invested, many global enterprises struggle to see clear financial returns from AI initiatives. Over the past two years, spending on artificial intelligence has surged, yet the expected "return" remains hard to pin down. Executives face a moving target when trying to quantify AI’s impact on their bottom lines.
Companies have quickly adopted generative AI tools and developed AI-driven products, but measurable financial benefits often lag. Ryan Kane, CEO of Soaring Towers, notes that truly impactful ROI that could replace jobs is still years away. Today’s AI is advanced pattern recognition rather than genuine intelligence.
Why Are Returns So Elusive?
Consultants and AI experts point to several reasons returns often fail to materialize:
- Mismatched Expectations: Many expect immediate, dramatic results, but current AI’s impact is subtler and slower.
- Implementation Challenges: Data preparation, model retraining, and infrastructure updates often get overlooked.
- Technical Limits: AI tools can’t yet tackle complex decision-making that requires human collaboration and reasoning.
For example, tools like Copilot improve user experience by speeding up tasks like data search. However, quantifying this boost in financial terms is tough. Kane explains it’s difficult to measure if a user finding data in 3 seconds instead of 30 actually adds value to the business.
The Hidden Costs of AI
Cory McNeley from UHY Consulting highlights that many companies underestimate the effort behind AI readiness. Preparing clean, labeled data from siloed systems demands significant time and resources. Additionally, ongoing costs like preventing model drift, managing security risks, and complying with regulations add to the burden.
Kris Bondi, CEO of cybersecurity startup Mimoto, points out that organizations often launch AI pilots without proper oversight or strategic planning. They face obstacles such as shallow usage patterns and outdated decision-making structures. Without proper training and query design, AI won’t deliver strategic breakthroughs.
Shallow Adoption and Early Abandonment
Many businesses limit generative AI to automating simple, repetitive tasks. This shallow use often leads to early abandonment after users get mixed results. Kane notes that without structured onboarding and alignment with real business needs, AI features can go unused—even when embedded directly in workflows.
Smaller companies face particular challenges, often relying on off-the-shelf AI tools with little customization. This can force shortcuts that make achieving ROI even harder. Larger enterprises have an advantage with dedicated teams to map use cases and integrate AI effectively, but they're not immune to structural and technical barriers.
The Gap Between AI and Real Business Needs
Ram Bala, business analytics professor, explains that current AI struggles to match the collaborative nature of real business decisions. Large language models help with language tasks but can’t yet handle complex team dynamics. J Stephen Kowski, field CTO at SlashNext, warns that simply adding more data won’t fix fundamental issues like hallucination or faulty reasoning. He advocates for targeted, domain-specific AI systems that solve real, repeatable problems.
Thomas Atkinson from NCC Group adds that many AI rollouts are reactionary rather than purposeful. Companies often adopt AI because of external pressure, not because they have clear business goals tied to the technology. A shift toward deliberate alignment with business objectives is needed.
Experimentation Has Its Place—But ROI Requires More
Stuart King, CTO at AnzenOT, acknowledges that early AI projects can serve as valuable learning experiences. Familiarity with AI technology builds problem-solving skills, which may pay off later. Still, this alone won't deliver the kind of returns companies seek.
McNeley recommends defining success upfront and measuring it continuously with layered metrics. These should track everything from model uptime to employee innovation rates and new revenue driven by AI. Bala suggests a staged approach: start by measuring individual productivity gains, then team outputs, and finally business-level outcomes like revenue or customer satisfaction.
Focus on Solving the Right Problems
Experts agree that most current AI tools are imitators, addressing surface-level issues while missing deeper inefficiencies. Kowski stresses that AI implementations often tackle the wrong problems. The real opportunity lies in building tools that solve specific challenges exceptionally well.
Increasing AI budgets alone won’t guarantee returns. Kowski sums it up: no CEO wants to miss out on AI, but chasing appearances without real impact leaves the ROI question unanswered.
For businesses looking to deepen their AI knowledge and improve adoption strategies, exploring focused training can help. Consider checking out Complete AI Training for courses that cover AI tools, implementation, and strategic use cases.
Your membership also unlocks:
