Most organizations now use AI somewhere. Very few can point to what it has changed.
McKinsey's latest global survey found that 88% of organizations use AI regularly in at least one business function. Yet nearly two thirds have not started scaling it across the enterprise, and only 39% report any impact on enterprise-level earnings. Most of those say AI accounts for less than 5% of their EBIT.
That gap between use and value is where most AI budgets disappear. In our work with governments, institutions and growing companies, we see the same pattern again and again. The organization buys a tool, runs a pilot, celebrates a demo, and then the project quietly stops. Nobody decides to stop it. It simply never becomes part of how work gets done.
The cause is almost never the technology. It is the order in which things were done.
The Gap Between AI Use and AI Value
Why AI Programs Stall
When Gartner predicted that at least 30% of generative AI projects would be abandoned after proof of concept, it named four reasons: poor data quality, inadequate risk controls, escalating costs, and unclear business value.
Look closely at that list. None of the four is a technology problem. Each one is a question about the organization that should have been answered before a tool was chosen.
Poor data quality means nobody mapped what data exists, where it lives, and whether it can be trusted. Gartner's own survey found that 63% of organizations either do not have, or are not sure they have, the right data management practices for AI. It predicts that through 2026, organizations will abandon 60% of AI projects that are not supported by AI-ready data.
Inadequate risk controls means governance was treated as paperwork for later. In Deloitte's survey of 2,770 business and technology leaders, the most common barriers to scaling generative AI were regulatory compliance, risk management, and the lack of a governance framework.
Escalating costs usually follow from building before the scope was clear. Unclear business value means the use case was chosen because it was exciting, not because it mattered.
S&P Global found that the share of companies abandoning most of their AI initiatives before production jumped from 17% to 42% in a single year. On average, organizations reported scrapping 46% of projects between proof of concept and broad adoption. IBM's 2025 CEO study adds a telling detail: half of CEOs say the pace of recent investment has left their organization with disconnected, piecemeal technology.
That is what tool-first AI produces. A collection of pilots that do not connect to each other, to the data, or to the way the business actually runs.
What the Organizations Getting Results Do Differently
The research on organizations that do see returns is remarkably consistent.
McKinsey identifies a small group of AI high performers, about 6% of respondents. They are nearly three times as likely as others to have fundamentally redesigned individual workflows around AI, and three times as likely to say their senior leaders show clear ownership of and commitment to AI. Out of 25 attributes McKinsey tested in an earlier edition of the survey, workflow redesign had the biggest effect on whether an organization saw an impact on earnings. Only 21% of organizations had done it.
The same survey found that CEO oversight of AI governance is one of the factors most closely linked to bottom-line impact. Only 28% of organizations say their CEO is responsible for it. And the single practice with the most impact on the bottom line was tracking well-defined KPIs for AI solutions. Fewer than one in five organizations do.
BCG's research points the same way. Companies leading on AI pursue, on average, about half as many opportunities as their peers, and expect more than twice the return. They focus. They also follow what BCG calls the 10-20-70 rule: 10% of the effort goes into algorithms, 20% into technology and data, and 70% into people and processes.
Leaders pursue half as many AI opportunities as their peers and expect twice the return. Focus is the advantage.
Read together, these findings describe organizations that understood themselves before they bought anything. They knew which workflows mattered, who owned the outcome, what data they had, and how they would measure success.
Five Questions to Answer Before Choosing a Tool
Before any AI tool is selected, leadership should be able to answer five questions with confidence.
1. Where does value actually sit in the organization? Not where AI looks impressive, but where time, cost, risk or revenue is concentrated. This requires a view of the business model, the strategy, and the functions that drive results.
2. What data do we have, and can we trust it? Every AI use case depends on data. If it is scattered across systems, incomplete, or owned by nobody, that is the first project, not an afterthought.
3. Who owns this? An AI initiative owned by "everyone" is owned by no one. The research is clear that visible senior ownership is one of the strongest predictors of results.
4. In what order should we do things? Some use cases are quick wins that build confidence and clean up data. Others are strategic anchors that depend on those foundations. A few are innovation bets. Doing them in the wrong order wastes the good ones.
5. How will we know it worked? Define the KPIs before the build, not after the demo. If success cannot be measured, it cannot be defended at the next budget review.
Sequencing Is the Strategy
Most AI roadmaps are ranked lists of use cases. The best ones are sequences.
The right order of ordinary use cases will outperform the wrong order of great ones. A document automation project that cleans and structures your data makes the next analytics project possible. A small internal assistant that earns the trust of a team makes the larger change easier to accept. Each step should create the conditions for the next one.
This is why we score every opportunity on more than its potential value. Data readiness, dependencies, risk, change effort, and time to impact all decide where it belongs in the sequence.
What This Looks Like in Practice
When the UAE government set out to move from digitization to AI-enabled services, the work did not start with a tool. Innavera began with a baseline of digital maturity across government entities and benchmarked it against leading digital governments, including Singapore, Estonia, and South Korea. Only then did we map more than 90 high-impact AI use cases across citizen services, prioritize them, and design a 3 to 5 year roadmap with governance for responsible AI and data ethics. You can read the full UAE government AI roadmap case study.
The same principle applies outside AI. For Dubai SME, we designed the full startup journey, from awareness through learning, mentorship, evaluation and funding, before building the platform that runs it. One integrated system with structured scoring and review replaced what would otherwise have been a set of disconnected tools, and it attracted more than 3,000 entrepreneur registrations. Mapping the process first is what made a single platform possible. It is also exactly the kind of foundation that automation and AI depend on later.
How Innavera Approaches AI Integration
The Innavera AI Framework is built on this sequence. It has five phases: Illuminate, Identify, Architect, Activate and Accelerate.
The first two phases answer the five questions above. Illuminate produces an Organization Intelligence Brief, a strategic baseline across eight assessment domains, including strategy, data, talent, culture and regulation. Identify produces an AI Opportunity Map: the top five to ten use cases, scored across eight dimensions and sequenced into quick wins, strategic anchors and innovation bets, each with a business case.
Only then do we design the architecture, build or integrate the solutions, and run the adoption program that makes them stick.
For most organizations, the right first step is our AI Compass Assessment. In three to four weeks it delivers the Organization Intelligence Brief, the prioritized AI Opportunity Map, peer benchmarking, and a leadership workshop that ends with clear next steps. It gives you a defensible plan before you commit significant budget to any tool.
Start with a clear picture
If your organization is running AI pilots that are not adding up to results, or you are about to make your first serious investment, we would be glad to help you map the ground first.
Explore the Innavera AI Framework →
References
- McKinsey & Company (2025). "The State of AI in 2025: Agents, Innovation, and Transformation." mckinsey.com
- McKinsey & Company (March 2025). "The State of AI: How Organizations Are Rewiring to Capture Value." mckinsey.com
- Boston Consulting Group (2024). "AI Adoption in 2024: 74% of Companies Struggle to Achieve and Scale Value." bcg.com
- Gartner (2024). "Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025." gartner.com
- Gartner (2025). "Lack of AI-Ready Data Puts AI Projects at Risk." gartner.com
- S&P Global Market Intelligence (2025). "AI Experiences Rapid Adoption, but With Mixed Outcomes." spglobal.com
- IBM Institute for Business Value (2025). "CEO Study: CEOs Double Down on AI While Navigating Enterprise Hurdles." ibm.com
- Deloitte (2024). "State of Generative AI in the Enterprise, Q3 Report." deloitte.com
