Nabugu - stock.adobe.com
Why enterprises shouldn't accept "good enough" AI ROI
Many companies see only modest AI gains while far greater value sits untouched. The real gap isn't technology, it's strategy.
Finance market leaders report AI returns of approximately 10%, which is barely above the cost of capital, according to a BCG study. That's not success. That's survival. The real story isn't about failed AI pilots. It's the billions left on the table: why settle for breadcrumbs when much higher ROIs are proven and possible?
Leaders must stop applauding small wins and recognize this opportunity gap. Deloitte's study shows only 20% of firms report GenAI returns above 30%, just 6% exceed 50%, leaving the vast majority stuck in single digits. The cost of complacency is steep: stagnation, loss of competitive edge and pilot projects that never scale beyond isolated use cases. Those missed opportunities block the path to real revenue growth and operational transformation.
The safety of being "fair to good"
Too many business leaders treat AI as an experiment, starting with proofs of concept that lead nowhere, chasing "let's see what we can do with this" results instead of asking, "what needle are we trying to move?" The result is predictable. Fragmented data, unclear reporting and growing frustration across teams who blame the tech rather than themselves. MIT recently reported that 95% of enterprise generative AI pilot projects fail to deliver measurable results. That's often not a failure of technology; it's a failure of the enterprise.
Current missteps
The common response has been to throw more people at the problem. The rise of the "forward-deployed engineer" model, embedding technical staff inside client organizations for months or years, sounds impressive but often benefits the vendor more than the customer. It's slow and costly. Meanwhile, the customer is left waiting for insight while competitors move faster with less work and get results in weeks.
Too often, teams feel the pressure of being behind and launch AI initiatives carelessly without clear KPIs or a definition of success. If you want AI projects to succeed, you need to be explicit about what metrics matter and how impact will be measured from day one. No structure is what produces modest ROI when exponential value is happening on a daily basis.
The formula
We've seen what happens when companies design AI with structure, not spontaneity. Businesses that contextualize and tailor their AI strategies typically achieve better returns. The key is starting with clear visibility into where value is hiding and building momentum one use case at a time. Leaders should focus on three areas: observability and reporting, unstructured data use, and workflow and automation. It starts with creating transparency around how data moves through the business and what impact those flows have on decisions. For example, many enterprises are drowning in unstructured data, such as spreadsheets, statements, reports and contracts, that sit idle in PDFs and shared drives.
Without observability, reporting and a defined workflow strategy, organizations cannot extract value from what they already own. Once businesses uncover these areas, the compounding effect becomes clear: one success funds and informs the next. The building blocks at our company are designed so that customers feel the impact from the first use case, return for the second and third, and so on. Being able to meet with a company, understand its specific needs, workshop use cases, and deliver a production-ready resolution within a week, built on a platform tailored to that challenge, is exactly the advantage most enterprises are missing.
Elevating past experimenting
We've passed the time when enterprises treat AI in an experimental capacity. It is proven tech that fundamentally changes the way your organization operates if the effort is put in to provide clear and obtainable results. It will be the enterprises and their leaders who understand this and demand more than "good enough" that will unlock three-digit exponential ROI.
Shay Levi is the co-founder and CEO of Unframe, where he focuses on building transformative enterprise AI solutions. Before founding Unframe, he co-founded Noname Security and helped lead the company to $40 million in ARR and a $500 million acquisition by Akamai in just four years. Throughout his career, Shay has tackled complex technical challenges at scale across cybersecurity, infrastructure and product development, and now applies that experience to redefining what's possible for enterprises through AI.
