If one thing didn’t need saying, it’s that AI is no longer a futuristic concept. Some say it’s already reshaping how organizations operate, compete, and serve customers. Others claim it already has. But as many leaders are discovering their AI journey, adoption is not simply a matter of deploying new tools. 

It requires thinking down to the digital core itself.  

At a recent executive roundtable in Chicago, hosted by Navisite, Part of Accenture, in partnership with SAP , technology and business leaders from financial services, healthcare, logistics, retail, and software came together to discuss what’s working, what’s not, and what’s next when it comes to AI-driven transformation. The conversation moved beyond buzzwords, focusing instead on practical lessons learned from real-world deployments and regulated environments. 

The message was clear: organizations that are seeing results are grounding AI innovation in strong governance, business-led use cases, and a modern, cloud-based digital core. 

The Reality Check: AI Projects Require Operational and Cultural Transformation 

Despite unprecedented interest and investment, AI initiatives are not ‘plug and play.’. Panelists across industry acknowledged that early enthusiasm has led to fragmented experimentation, duplicated efforts, and unexpected risk. 

Governance gaps were a recurring theme. Several leaders described an initial flood of proofs of concept (POCs) created in silos where teams experimented independently with little enterprise-level visibility. Without a centralized intake and review process, organizations struggled to prioritize the right use cases or eliminate redundancy.  

Cost overruns quickly followed. In one example, teams replaced existing microservices with AI-driven alternatives that delivered no additional business value while driving cloud costs up by three-to-five-fold. The lesson: AI adoption without disciplined evaluation can increase complexity rather than reduce it.  

Compliance challenges were especially acute in regulated industries. Financial services, healthcare, and payroll leaders emphasized that even minor inaccuracies or hallucinations can create legal exposure, erode trust, or violate regulatory requirements. As one panelist noted, in certain environments, “accuracy matters more than speed or latency.”  

What’s Actually Working: Real AI Use Cases from the Field 

While challenges were openly discussed, the panel also shared compelling examples of AI delivering tangible value when applied thoughtfully. 

One of the most widely cited successes was intelligent document processing (IDP). Organizations described using AI to analyze hundreds or even thousands of pages at a time, dramatically reducing manual review effort while keeping humans in the loop for final validation. This approach significantly improved efficiency without compromising accuracy or compliance. 

Customer-facing and agent-assist chatbots are also gaining traction. In call center and collections environments, AI is being used to summarize conversations, surface relevant policies, and reduce average handling time. Importantly, these systems are often paired with retrieval-augmented generation (RAG) models to ensure responses are grounded in approved enterprise data rather than open-ended generation. 

On the internal side, developer productivity tools such as code assistants and testing automation are delivering measurable gains. Several organizations reported productivity improvements of 20–30% once teams were properly trained, allowing engineers to focus on creative problem-solving rather than menial tasks. 

Another standout example came from logistics and hospitality organizations using AI to support multilingual workflows and intent-based search. This allowed users to interact naturally across languages and dramatically improving responsiveness in high-volume environments. 

Parameters, Not Handcuffs: How Leaders Are Governing AI 

A defining theme of the discussion was the shift from ad hoc experimentation to structured AI governance. 

Leading organizations described establishing centralized AI governance committees to bring together IT, security, data science, legal, and business leaders. These groups then evaluate proposed use cases, assess risk, ensure alignment with enterprise priorities, and prevent duplicate efforts across business units. 

Crucially, governance was framed not as a barrier to innovation but as an enabler. By creating clear parameters, organizations are empowering teams to experiment safely and then scale only those solutions that demonstrate clear business value and compliance readiness. 

Human oversight remains non-negotiable. Panelists described phased rollouts, frequent audits, and deterministic testing to ensure AI outputs remain reliable over time. In sensitive scenarios, models are explicitly designed to say, “I don’t know,” and escalate to a human when needed rather than risk an inaccurate response. 

People, Process, and Technology: The Winning Formula 

Despite the sophistication of today’s tools, one of the most consistent takeaways was that the fundamentals of digital transformation haven’t changed. 

Successful initiatives begin with business-driven use cases, not technology deployments for technology’s sake. When IT and business teams work in lockstep – defining problems together, aligning outcomes, and co-owning success – AI adoption accelerates and delivers broader impact. 

Training emerged as a critical differentiator. Several organizations admitted early missteps where AI tools were rolled out widely, only to see limited adoption. The turning point came when training was expanded beyond technologists to include business users, product leaders, and frontline teams, creating a shared understanding of what AI can and can’t do effectively. 

This cross-functional enablement is helping organizations move faster while making smarter decisions about where AI truly belongs. 

Preparing for What’s Next: Agentic AI, Audits, and the Cloud Core 

Looking ahead, panelists pointed to agentic AI: multi-agent workflows capable of handling complex, end-to-end processes. However, most agreed the technology is still maturing and must be introduced carefully, particularly in regulated environments. 

Today, it’s clear that scaling AI requires a modern, cloud-first digital core. Legacy, siloed systems limit integration, orchestration, and governance. By contrast, cloud-based platforms enable consistent security controls, faster innovation cycles, and the ability to consume AI capabilities as they evolve without constant reinvention. 

Final Takeaway: How Organizations Can Move Forward with Confidence 

This roundtable reinforced a powerful truth: AI success is not about the organization chasing the latest model. Rather, they need to focus on building the right foundation. 

Organizations that are winning with AI are aligning people, processes, and technology; governing innovation with intention; and modernizing their digital cores to support continuous change. With the right partners, guardrails, and strategy in place, AI becomes not a risk to manage but a catalyst for smarter and more resilient growth. 

Looking to explore how AI can level-up your organization? Contact us to start charting your path.