Scaling AI in the Real World: A TechCrunch Disrupt Conversation
At TechCrunch Disrupt 2025, one of the standout sessions on the AI Stage was “AI That Scales: Lessons From the Frontlines.” The panel featured Sanjay Dhawan, CEO of SymphonyAI; Harmeet Bala, SVP and Head of Product Development, Commercial and Money Movement Solutions at Visa; and Tamara Pattison, SVP and Chief Digital Officer at The Save Mart Companies.
Together, these leaders explored what it truly means to scale AI across the enterprise—from global payment networks and financial institutions to grocery supply chains. The conversation focused on how Vertical AI delivers measurable ROI, sustainable competitive advantage, and real-world transformation far beyond the prototype stage.
1. Scaling AI for Impact: Depth Over Breadth
SymphonyAI CEO Sanjay Dhawan emphasized that scaling enterprise AI requires depth of domain expertise, not breadth of generic capability. SymphonyAI’s approach—Vertical AI—focuses on solving specific, high-value problems within industries such as retail, financial services, manufacturing, and IT operations.
Through the Eureka AI Platform, SymphonyAI delivers pre-trained, domain-specific solutions like CINDE for connected retail and CPG, Sensa for financial crime prevention, IRIS Foundry for industrial optimization, and Apex for enterprise IT.
“Scaling AI isn’t about building one-size-fits-all tools,” Dhawan noted. “It’s about applying deep domain context to real business challenges and delivering durable, measurable outcomes.”
Both Visa and Save Mart shared this philosophy: start by solving one mission-critical problem at scale, then expand AI’s role across the organization.
2. Proven Results: How AI Delivers ROI in Financial Services and Retail
Visa was an early pioneer in enterprise AI adoption. Harmeet Bala revealed that the company’s first AI model went live in 1993, long before AI was a mainstream concept. Today, Visa runs more than 300 AI models across over 300 billion annual transactions, preventing an estimated $14 billion in fraud last year alone.
“AI is not just a tool—it’s a strategic capability that underpins our entire payment network,” Bala said.
At Save Mart, Tamara Pattison described how AI in retail has revolutionized promotion optimization. “We spend hundreds of millions on promotions each year,” she explained. “With predictive AI, we can identify the right offers for the right customers, improving engagement, margin, and customer value.”
This transition from intuition-driven to AI-driven decision-making demonstrates how Vertical AI in retail delivers measurable financial outcomes—both for consumers and for the business.
3. Scaling Enterprise AI: The Trifecta of Data, Technology, and Culture
Both panelists agreed that the hardest part of scaling AI isn’t the algorithms—it’s the organizational readiness. Bala described success as a “trifecta” requiring:
-
High-quality, connected data
-
Robust AI infrastructure and expertise
-
A culture that embraces learning and iteration
“You can’t separate culture from capability,” she said. “Teams must be open to experimentation and comfortable with technology that’s evolving quickly.”
In retail, Pattison noted, data integrity and trust are critical. “You have to make sure your data is complete and current. Otherwise, AI can lead you to the wrong conclusions. Education and transparency are key to adoption.”
For SymphonyAI’s customers, this is exactly where Vertical AI makes the difference—by combining industry-specific data models with pre-built AI workflows that minimize risk and accelerate value realization.
4. AI That Empowers People, Not Replaces Them
Across sectors, AI at scale is augmenting—not replacing—human expertise. Bala described how Visa’s developers now use AI copilots throughout the software lifecycle, improving productivity in design, development, and testing while freeing engineers to focus on innovation.
At Save Mart, AI is improving workforce productivity and training, reducing onboarding time and enabling better, faster decisions. “AI helps our teams focus on what matters most—serving our customers and driving the business forward,” Pattison said.
Dhawan reinforced this theme: AI at scale enhances human creativity and strategic focus, rather than automating it away.
5. The Future of AI: Autonomous Agents and Sustainable Impact
Looking ahead, Bala predicted a future where AI agents autonomously transact on behalf of consumers and enterprises, simplifying procurement, negotiation, and payment. Visa estimates that by 2030, AI agents could drive up to $3.5 trillion in global commerce, supported by secure authentication and payment infrastructure already in progress.
Pattison envisioned a different kind of transformation: AI for sustainability. “Food waste is a massive issue in grocery,” she said. “AI can help us forecast demand more precisely, buy smarter, and ensure products are consumed rather than discarded.”
Both perspectives point to a single truth: AI at scale drives both profitability and purpose—a hallmark of sustainable digital transformation.
The session underscored a shared truth across industries: AI at scale is less about hype, and more about durable, domain-driven impact. As Sanjay Dhawan summarized, “This is AI in the wild—practical, measurable, and reshaping industries one workflow at a time.”
