From Hype to Habit: How AI Is Actually Transforming Business

AI is moving at a blistering pace, with new developments in models and their impact on the business world, it can be hard to keep up. Luckily, we have experts like Ritcha Ranjan, Senior Vice President of Product at Expedia Group and Advisory Board Member with the Wharton AI & Analytics Initiative, to help us make sense of it all. In a recent interview with Eric Bradlow, Vice Dean of AI & Analytics at Wharton, the pair helped make sense of the current state of AI. Here are the key takeaways from their conversation.

1. GenAI isn’t a Spreadsheet – Don’t Treat it Like One

“When we say automated, we think deterministic—A then B then C,” Ranjan explained. “With generative AI, the path isn’t deterministic. It can be A–C–B or something else entirely depending on the context.”

For financial firms, this means AI can move from executing transactions to interpreting context. For example: flagging unusual client behavior, dynamically adjusting credit risk parameters, or generating first-draft analyses.

Bradlow emphasized that this adaptability will change how organizations think about oversight: “You really have to decide where to have a human in the loop. Sometimes you’ll need it. Sometimes, when you’re confident in the output, you won’t.”

2. Success Will Come from Specialization, Not Scale

Ranjan noted a common AI talking point: “95% of AI projects are failing from a financial point of view.” She says that’s often because they aim too broadly. “If you make it too horizontal and you aren’t focused enough on the workflow – the steps, the personas – you start to fall down.”

Her advice to companies: move from horizontal experimentation to vertical execution. Financial institutions, for instance, should focus on well-defined domains like fraud detection, compliance reporting, or personalized wealth recommendations before scaling to enterprise-wide platforms.

3. Data Discipline Determines ROI

Even the most advanced AI models fail without robust data foundations. “I assumed all companies had their data together,” Bradlow remarked. “It’s not true. The data is messy, different systems don’t integrate, and that’s a big challenge.”

Ranjan agreed, stressing that firms need to begin with “data centralization and accessibility” before diving into generative workflows. Without that discipline, AI can amplify errors rather than insights.

For leaders at banks, asset managers, or fintechs, the implication is clear: data readiness is an investment-grade metric. Before calculating AI ROI, firms must evaluate whether their systems and teams are truly AI-ready.

4. Change Management is the Hidden Variable

“I’ve heard it said many times—AI dies at the front lines,” Ranjan noted. Her point: technology adoption often fails not for technical reasons, but because of human behavior.

To address this, she recommended making HR a strategic partner in AI deployment. “Some companies now have HR leading AI change management—and it makes a ton of sense,” she said. Hands-on training also matters. “You can’t just play a training video. Get people in a room with laptops. Let them try prompting. Let them see the benefit.”

Organizations that treat adoption as a human challenge, not a software rollout, will build more sustainable AI cultures and realize more measurable returns.

5. The Future: Quiet Integration and Industry-Specific Breakthroughs

Looking ahead, Ranjan predicted that by 2030, AI will become less of a headline and more of a habit. “No one talks about the web anymore—it’s just there. AI will be the same,” she said.

But she also sees sector-specific leaps coming: “Healthcare and financial services are going to see quantum jumps in capability, especially in personalization, fraud reduction, and financial health.”

As models converge and differences narrow, it’s how businesses apply them that will define competitive advantage.