AI’s Slam Dunk: Xebia’s Innovation Xchange Meets March Madness

Insights from Xebia’s Innovation Xchange Atlanta.

College basketball wasn’t the only excitement in Atlanta in late March. Xebia’s Innovation Xchange event gathered business and tech leaders for a panel titled “The AI-Enabled Enterprise: From Vision to Reality.” The atmosphere blended the buzz of the tournament with the thrill of enterprise innovation. On stage, three experts from distinct fields – cloud computing, banking, and legal services – examined how artificial intelligence (AI) is transforming their industries. Raj Sachde of AWS, Kevin Hearn of Axos Bank, and Kyle Dumont of the law firm Morgan Lewis & Bockius shared insights on leveraging generative AI for growth, modernizing legacy systems, overcoming adoption hurdles, and fostering a culture of curiosity. The evening felt as charged as a championship game, but here the true game changer was AI.

AI as a Growth Enabler, Not Just for Efficiency

To kick off, moderator and Xebia’s Global Managing Director, Preetpal Singh posed a big-picture question: How is AI pushing industry boundaries? Historically, many companies viewed AI primarily as a tool for efficiency or cost-cutting. Raj Sachde, who leads enterprise AI partnerships at AWS, challenged that mindset. He urged businesses to focus on “AI as a growth enabler” rather than a mere productivity booster. At AWS, Sachde explained, they use a “working backwards” approach: “We start with what is the objective the customer is trying to achieve, work with partners in academia and come up with... co-creation, growth, [a] co-innovation approach”​

In other words, begin with the end value in mind. By identifying new revenue streams or business models first, enterprises can build AI solutions that unlock those opportunities. This emphasis on value creation echoed a broader trend – top companies adopting AI today are aiming beyond cost savings. In fact, industry research shows AI leaders anticipate 60% higher AI-driven revenue growth by 2027 (along with significant cost reductions) compared to their peers​ (bcg.com). AI isn’t just about doing the same for less; it’s about doing more and doing it differently.

Kevin Hearn, a technology executive at Axos Bank, underscored that point from a financial industry perspective. Sure, AI can automate routine processes in banking, but Axos is using it to enhance customer trust and safety, which ultimately fuels growth. Hearn described how AI helps the bank proactively detect fraud and smooth out onboarding for new customers. For example, when a customer opens an account online, AI-driven ID verification kicks in behind the scenes. “You’re typically loading a picture of your ID… but in parallel, [our system is] live scanning using the device’s camera… in real time, comparing it to the uploaded photo to make sure it’s a real person,” he said​.

The bank protects itself and its customers by catching identity fraud or even “fake employees” before they slip through. That kind of proactive risk management isn’t just about cutting losses – it builds a safer, smoother customer experience that can drive growth through trust. In Kevin’s words, AI in banking means looking at every step of the customer experience process to both “protect the bank but also protect the customer.”​

Over in the legal realm, Kyle Dumont is helping Morgan Lewis – a 150-year-old global law firm – leverage AI not just to speed up legal research, but to create entirely new client offerings. Law is famously conservative, but Dumont’s digital transformation team focuses on client value and “co-creating” solutions with clients. He noted the “low-hanging fruit” for AI in legal has long been tasks like reviewing documents or drafting basic contracts. Those areas see immediate efficiency gains since legal work is so document-driven. However, Kyle is pushing further. The real promise of AI in legal services lies in augmenting human judgment to address unmet needs. “We look for opportunities of unmet customer need, where we can augment human intelligence with computer intelligence in novel ways that no one else is doing yet,” he explained.

In practice, that could mean AI tools that predict case outcomes or advanced analytics that guide legal strategy – applications that create new client value, not just faster paperwork. As Dumont put it, the key is decoupling a lawyer’s value from the physical act of writing documents. If “mechanical generation of text” costs next-to-nothing now, lawyers must refocus on higher-value advisory roles​. All three panelists agreed: AI’s real power is in opening new frontiers for growth and innovation, not simply trimming fat.

Automation for Augmentation, Not Elimination

When the discussion turned to automation, the panelists quickly dispelled the “robots coming for our jobs” anxiety. As the Preetpal noted, “automation has moved far beyond simple task elimination – it’s now a strategic decision-making tool​.”

Raj Sachde illustrated this with Amazon’s concept of “agentic AI.” He described how AWS envisions AI agents that can manage multiple linked tasks across a business process rather than just solving one problem in isolation. “It’s not just a point solution,” Raj said. Instead of one AI doing one thing, you might have a constellation of models and tools working in concert: “How do you have AI manage multiple tasks across the business process? ... For a first task, it may be talking to one LLM at the back end. For a second task, it may be talking to another LLM. And for a third task it might be doing something on-prem in a customized way,” he explained​.

In essence, AI can act like an orchestrator – delegating subtasks to different specialist models or systems – to carry out a complex workflow. This shifts the conversation from automation vs. jobs to automation as augmentation. The goal is to have AI handle the grunt work (and even coordinate other AIs) so that humans can focus on supervision, creativity, and decision-making.

Kevin backed this up by sharing how Axos Bank approaches automation in software development. Rather than searching for a single silver-bullet AI tool, his team is “assessing… nine different AI tool sets to augment” various roles across the SDLC (Software Development Life Cycle)​.

“It’s not one tool to fill every need,” he noted – you might use one AI to generate test cases, another to optimize code, and yet another to monitor customer experience. This toolbox approach ensures that automation truly supports each team member instead of replacing them. Kevin emphasized that the key is finding the right tool for each job and preparing the organization to adopt it: “The key is finding the right tool, but it’s also [about] adoption and readiness,” he said, stressing the importance of change management as they introduce AI helpers​. In practice, Axos developers use AI coding assistants (like GitHub Copilot for “high-code” work) alongside low-code platforms enhanced with genAI to accelerate development​. The result is faster delivery of features without sacrificing human oversight – a competitive edge in a fast-moving digital market.

Even in the legal industry, automation is about working smarter, not making attorneys obsolete. Dumont pointed out that law firms have already lived through waves of automation (e-discovery software, document review tools, etc.), and lawyers have continued to thrive by moving up the value chain. Now, generative AI can draft a decent legal brief in seconds. Instead of fearing that, leading firms ask how attorneys can partner with AI to deliver better outcomes. One lesser-known application Kyle highlighted is using AI to analyze and learn from past cases in ways a human could never do alone. For example, an AI system might comb through thousands of case files to identify patterns that inform litigation strategy – something beyond mere “repetitive task” automation. By offloading drudgery to machines, lawyers can spend more time on client counsel, complex reasoning, and courtroom strategy. Automation, in this view, isn’t subtracting humans from the equation; it’s adding AI to amplify human expertise. As all the panelists agreed, the narrative has shifted from AI versus people to AI and people. The enterprise of the future uses automation to free employees for higher-impact work.

Modernizing the Core: From Legacy Systems to an AI-Ready Foundation

One theme that came up repeatedly was the importance of modernizing a company’s digital core – its data architecture and legacy systems – to fully capitalize on AI. You can’t become an “AI-enabled enterprise” when your data is locked in silos or running on 40-year-old technology. This is a challenge in every sector, but perhaps nowhere more stark than in banking. Kevin Hearn gave a frank account of the hurdles banks face with old core systems. “We have some really, really antiquated technologies… 40-plus years old, and we’re still running [them],” he revealed​. This technical debt can impede AI projects for a born-digital bank like Axos, which has grown through acquisitions. Rather than accept that fate, Kevin’s team tried a bold solution: using AI itself to modernize legacy code. He described taking an AI-driven “reverse engineering” approach to migrate old systems into modern tech stacks​.

In pilot projects, they pointed to generative AI at end-of-life code and had it convert outdated languages and frameworks into updated ones. For instance, Axos used AI tools to turn old .NET code into Angular 15 web applications and even to translate ancient BASIC code into C#​.

In one example, they successfully transformed a PL/1 (an old mainframe language) program into a modern cloud-compatible service. This kind of AI-assisted refactoring could be a game changer for banks: it shrinks multi-year core modernization efforts down to months by automating the understanding and rewriting of legacy code. Hearn called it out as a “game changer… for companies that inherit tech [through acquisitions]”, because it removes what was once a costly, slow barrier to innovation​.

Given that 92 of the world’s top 100 banks still rely on mainframes for their core systems​ the potential impact of AI in core modernization is huge (Planetmainframe.com).

Raj Sachde agreed that data and infrastructure modernization are often the make-or-break factor for AI initiatives. From the AWS perspective, he sees many enterprises stumble by trying to layer fancy AI tools on top of broken data foundations. One of the biggest mistakes is not addressing data quality, silos, and governance upfront. Sachde emphasized basics like establishing a unified data lake or cloud data warehouse so AI models can access consistent, comprehensive data. He mentioned that AWS works with customers to avoid these pitfalls by demonstrating clear ROI at the pilot stage and scaling out. “If you want to do a POC to show the benefits... there is a way to do that,” Raj noted, explaining that AWS’s GenAI Center of Excellence and partner network help companies measure value (in ROI or TCO terms) early. But beyond technology, true AI readiness requires alignment across security and compliance as well. Raj reminded the audience that for AWS, “security is job zero… CEO down will not compromise on security”, especially in the cloud​

He pointed out that moving to cloud infrastructure can improve an enterprise’s security posture and compliance as long as proper guardrails are in place. In regulated industries like finance and healthcare, modern cloud architectures now offer solutions (virtual private clouds, encryption, audit trails) to satisfy regulators while enabling advanced AI analytics.

Modernizing the core isn’t just an IT project – it’s a strategic imperative. As the panelists made clear, legacy systems and siloed data can be major barriers to AI, but new approaches (sometimes using AI itself) are making it easier to break through these barriers and lay an AI-ready foundation. Companies that invest in updating their “digital core” today will lead the pack with AI solutions tomorrow.

Overcoming Barriers: Ethics, Risk, and Organizational Readiness

Of course, transforming into an AI-enabled enterprise is not without its challenges. The panel candidly discussed the ethical, practical, and organizational hurdles they’ve encountered. Dumont noted that one big barrier to AI adoption in the legal sector has been the profession’s cautious culture and legitimate concerns about confidentiality. Lawyers are trained to be risk-averse – after all, client secrets and case outcomes are on the line. “There are duties of confidentiality, of course,” Kyle said, meaning any AI tools must rigorously protect sensitive data​. He gave the example of attorneys experimenting naively with ChatGPT and accidentally exposing client information or getting “made-up citations” in return​. “Headlines around attorneys using Gen AI are not super great because they’re confusing a deterministic computing process with a stochastic one,” he joked, referring to well-publicized incidents of chatbots fabricating legal citations.

This gap in understanding is both a technical and educational hurdle – many legal professionals don’t yet grasp how generative AI works under the hood, leading to misuse. However, Kyle believes this barrier is starting to fall as awareness grows. Law firms are now actively mitigating risks by implementing review processes for AI outputs and choosing vetted, secure AI platforms. Notably, a recent survey found 30% of law firms are now using AI, up from just 10% a year prior​ signaling rapidly growing comfort. (Lawnext.com)

Morgan Lewis itself has strict guidelines: attorneys must “know your server” just as they “know your customer,” meaning they should understand where their data goes when using an AI tool​.

By combining policy, training, and technology safeguards, legal organizations are gradually overcoming ethical concerns and unlocking AI’s benefits in a responsible way.

In banking and other regulated industries, governance and risk management are equally paramount. Kevin Hearn mentioned that any AI deployment at Axos goes through compliance and security reviews to ensure it meets regulatory standards. For instance, AI models that make credit decisions or flag fraud must be auditable and explainable to satisfy examiners. Kevin’s earlier example of AI-based ID verification shows how these tools are rolled out with privacy in mind – customers consent to biometric scans, and data is handled per strict protocols​. He acknowledged that balancing innovation with compliance is a delicate dance: moving fast with AI is great, but not at the expense of customer privacy or fairness. This is a common tension many enterprises face when adopting AI. According to a 2024 Deloitte survey, managing AI-related risks (like bias, privacy, and security) has become a top priority for large organizations​.

The panelists noted that one encouraging sign is how internal audit, legal, and IT teams are now partnering to set guardrails for AI use rather than blocking it outright. Raj added that the cloud can provide a secure environment for experimentation – for example, using synthetic data or sandboxed datasets to develop AI models before exposing real customer information. The message was clear: ethical AI and successful AI are two sides of the same coin. Companies must invest in the “boring” stuff (privacy policies, model validation, bias checks, staff training) to confidently scale the exciting stuff (personalized products, autonomous processes, etc.). Those who navigate these challenges will reap AI’s rewards, while those who ignore them risk setbacks or public trust debacles.

Culture Matters: Curiosity, Upskilling, and Cross-Disciplinary Talent

Toward the end of the panel, the human element of AI adoption was discussed. All three experts agreed that technology is only half the story; the other half is people and culture. How do organizations prepare their workforce to thrive in an AI-enabled enterprise? One theme was the need to foster curiosity and continuous learning. “Celebrate knowledge acquisition and teaching, and have that be a really strong, intensive motivation to learn within the team,”Kyle Dumont advised​.

At Morgan Lewis, he encourages a culture where technologists and lawyers alike share insights and experiment collaboratively. Dumont even asks his attorneys to become more “T-shaped professionals,” combining deep legal expertise with a broad understanding of tech and data (and yes, many have learned basic coding or data analysis). This cross-disciplinary skillset helps legal teams identify AI opportunities and work effectively with data scientists. Importantly, Kyle leads by example – readily admitting to his team when he uses ChatGPT or other tools to augment his own knowledge, thereby normalizing these aids as part of the job​.

Kevin offered a perspective from banking: in a fast-evolving field like AI, adaptability trumps experience. He looks for employees who are eager to pick up new tools and methodologies since the AI tools of today might be outdated in a year. Upskilling is a continuous process at Axos. “We’re looking at every persona… and what are the things that accelerate the work,” he noted, implying that every role, from product managers to engineers, is being reimagined with AI in mind​.

The bank has launched internal training for AI literacy, ensuring even non-technical business leaders understand concepts like machine learning, data science, and model risk. Hearn mentioned overcoming internal resistance by showing results from pilot projects – success breeds appetite. When people see AI taking tedious tasks off their plate or enabling them to achieve things they couldn’t, they become more curious and open to learning. Raj Sachde added that democratizing access to AI is crucial. At AWS, they not only train their own staff on AI, but also work with clients and partners to spread knowledge. Sachde highlighted a case where an insurance company “upskilled their staff with a training partner, and [then] built their own insurance claim processing platform” using AI​.

Empowering domain experts (in this case, insurance professionals) with AI know-how led to a ground-up innovation. Raj emphasized that AWS’s mission is “to democratize AI” in the enterprise – making it accessible to all departments, not just a centralized R&D team​. This often involves providing easy-to-use tools, sandbox environments, and forums for sharing AI success stories across the company.

The panel concluded with an optimistic outlook. Yes, building an AI-enabled enterprise is a journey filled with challenges: legacy technology, ethical dilemmas, skills shortages, and cultural inertia, to name a few. But none of these are insurmountable. The general advice? Start with a clear vision of value, modernize your core systems gradually, involve your people through learning and participation, and prioritize ethics. Organizations that follow this approach will enhance efficiency and unlock new levels of growth and innovation. As the crowd prepared to shift their focus back to the nail-biting basketball games, one thing was clear: in business at least, AI may be the ultimate slam dunk.