Many AI initiatives stall because enterprises rely on backward-looking BI, fragmented data, and human workarounds that mask weak foundations and limit real-world performance.
Bil Ahmed, Chief Cognitive Product and Technology Officer at LogicEra, explains how decades of enterprise systems built for reporting now block AI from supporting real decisions.
He defines the path forward as rebuilding data around shared meaning and decision logic, enabling enterprise-wide intelligence and reducing dependence on manual fixes.
For all the excitement surrounding AI, many deployments are stalled by a simple, fundamental issue: a lack of mastery over basic business intelligence. Ambition is running into the reality of disorganized data, inflexible reporting systems, and the manual processes employees use to bridge the gaps, resulting in a growing disconnect between the promised value of AI and its actual on-the-ground performance.
Bil Ahmed, Chief Cognitive Product and Technology Officer at LogicEra, brings a rare blend of commercial and technical leadership to today’s AI conversation. Over decades in senior roles at institutions such as Investec, WorldPay, and Barclays, he has worked at the intersection of data, decision-making, and enterprise scale. That experience shapes a pragmatic view of the current AI moment, grounded less in hype and more in the structural realities organizations must confront before intelligence can actually take hold.
"BI reporting is looking at the rear-view mirror of the car. Companies need to be aware of how they make their data AI-ready, and that is a moon jump," Ahmed says. The AI era has instantly made prior technology playbooks obsolete, demanding a complete foundational shift. Ahmed explains that the urgency is fueled by a decades-long reliance on unsustainable human workarounds.
All this tech, but still human: "With siloed data and a limited semantic layer, people are forced to become the human clusters that plaster the organization together. BI reports shuffle around, and it's the people who must act as the Band-Aids to make decision-making work at all." Moving past this manual dependence requires upgrading the system that defines data meaning beyond mere data routing to capture a rich, contextual understanding, including key business logic and human-level domain language.
Just add AI: Many organizations still operate under what Ahmed describes as a plug-and-play illusion, expecting enterprise AI to work like a consumer ChatGPT experience. The assumption is that centralizing data and layering an LLM with retrieval is enough to produce intelligence. "You can’t just centralize data, add a model, and expect an intelligent system to emerge," Ahmed says. Without structured, semantically rich data, these deployments produce unreliable outputs and quickly fail when exposed to real-world complexity.
This foundational weakness helps explain the widening gap between AI hype and technical reality. Ahmed points to a persistent plug-and-play mindset, where organizations expect advanced AI outcomes without first rebuilding the data foundations beneath them. That approach consistently fails because it attempts to layer intelligence onto data that lacks semantic structure, business logic, and contextual meaning.
Rather than a quick fix, Ahmed argues that AI success depends on a more disciplined, methodical reset. Treating data readiness as a core business capability, not a technical afterthought, reduces the need for human workarounds and reshapes how decisions are made across the organization. The payoff is not just stronger AI performance, but a more sustainable model for work and organizational design in an AI-driven economy.
Honoring intelligence: So what does the hard work of AI readiness actually look like? Ahmed’s advice to leaders looking to cut through the hype is to adopt a new philosophy he calls "honoring intelligence." Rather than relying on brittle systems patched together with human workarounds, the aim is to design intelligence into the enterprise from the start. The shift requires moving beyond legacy data structures and toward architectures that allow intelligence to emerge across the organization. As Ahmed explains, "The goal is to achieve intelligence that is enterprise-wide. That points toward designing a cognitive architecture, where a knowledge graph mirrors the neural structure of the human brain. If you carry forward the limitations of legacy data structures, you risk creating an immense opportunity cost. That opportunity cost is intelligence itself."
Ahmed’s guidance ultimately comes back to discipline. Start with a concrete business problem, work backward to define the ontology required to solve it, and resist the temptation to let AI strategy drift into abstraction. That rigor is what makes the moon jump possible in the first place.
But for Ahmed, even that leap is only a means to an end. AI, he argues, is not the destination but the foundation. The real transformation begins with Decision Intelligence, a new era focused on systematically improving judgment itself. "AI is the stepping stone, but it is not the final destination," he concludes. "The next cycle is Decision Intelligence. Everything we are doing is to build toward that future."
