Distyl AI and T-Mobile on Enterprise AI Transformation at Scale

At the AI Agent Conference in New York, one of the most concrete case studies of the two-day event came from Derek Ho, Co-Founder and COO of Distyl AI, and Julianne Roberson, Director of AI Engineering at T-Mobile. Where most sessions dealt in frameworks and principles, this one dealt in what actually happened when a top-five wireless carrier decided to move fast on AI — and what it took to pull it off.

The session's central argument: successful enterprise AI transformation requires two simultaneous shifts — a business model change and an operating model change. Technology alone delivers neither.

Everything Starts With the Customer, Not the Cost

The framing Julianne Roberson established early and returned to throughout the session is worth leading with because it's the thing most enterprise AI programs get backwards. "I don't care about reducing headcount. I care about improving customer experience."

T-Mobile's AI program wasn't designed as a cost-cutting exercise. It was designed as a customer experience transformation. That distinction shapes every downstream decision — what to build, how to measure success, what counts as a win. The best enterprise AI deployments, the session argued, are not the ones that minimize labor costs. They're the ones that create experiences customers didn't expect and couldn't get before.

"T-Mobile is extremely customer centric. Everything starts with the customer. We always prioritize the customer above ourselves."

The Goal Is to Solve Problems Before Customers Call

The most operationally specific concept from the session was what Roberson and Ho called the "breadcrumbs" model of proactive support. Customers constantly generate behavioral signals — frustration indicators, service usage patterns, early friction points — before they ever pick up the phone. T-Mobile's AI systems are designed to read those signals and intervene before a support call is necessary.

"Customers are constantly leaving breadcrumbs. How do we solve the problem before they ever have to call us?"

This reframes the entire customer service function. Reactive support waits for a problem to surface. Proactive service detects and resolves problems that the customer hasn't yet articulated. The ambition: "The future is customers not needing to call us. Customers don't want to spend 45 minutes on the phone."

That future state isn't hypothetical for T-Mobile. "AI is answering more than half of our phone calls now." That's not a pilot metric. That's operational scale.

August to January: Speed as a Competitive Advantage

One of the most striking facts of the session was the implementation timeline. T-Mobile and Distyl AI went from partnership kickoff in August to production deployment by January — at enterprise scale, across voice, chat, app, and multimodal interfaces.

"We started in August and were live by January. We didn't have time for shiny demos. You better move faster."

The organizations winning in enterprise AI right now are the ones willing to move before everything is perfect, iterate aggressively, and treat deployment as the beginning of the learning process rather than its conclusion. "No one has done this at this scale before. We're all learning. We're unafraid to move forward."

Roberson put the risk tolerance in memorable terms: "If we're not showing up on Reddit, we're probably not shipping enough." That's a cultural statement about what fast iteration actually requires — tolerance for public criticism, willingness to fail in visible ways, and the organizational confidence to keep moving.

Why T-Mobile Chose an AI-Native Partner Over Traditional Consulting

Roberson was direct about why Distyl AI won the partnership over traditional consulting firms. The distinction wasn't price or relationship. It was fundamental operating model.

"We were looking for someone AI-native. Not just throwing more people at the problem. They really understood the technology."

Traditional consulting scales delivery with headcount. AI-native partners scale with architecture. T-Mobile needed a partner that could move at startup velocity inside an enterprise environment, embed technically, and build systems rather than reports. That's a different kind of engagement than most large enterprises have historically procured — and it's becoming the standard for the organizations moving fastest.

Building a Startup Inside a Corporation

The organizational model Roberson built inside T-Mobile is worth examining closely. She described growing her AI engineering team from three people to twenty-five, deliberately mixing experienced enterprise operators with startup-minded builders.

"I think of my team as a startup inside a huge enterprise. We move at high velocity. I look for people who push past roadblocks."

The hiring philosophy was specific: people who know how to navigate large organizations without being slowed down by them, who treat constraints as problems to solve rather than reasons to stop, and who maintain urgency in an environment designed for process and caution. That's a distinct talent profile, and building it deliberately rather than accidentally is one of the things that separated T-Mobile's execution from the typical enterprise AI rollout.

Leadership Alignment Is Non-Negotiable

Both speakers returned to this consistently. Enterprise AI transformation at speed requires executive commitment that goes beyond budget approval. "This comes from the board all the way down. Our leadership is all in. Every time there's a roadblock, leadership is in the room."

The roadblocks in enterprise AI transformation are organizational as often as they are technical. Turf, process, legacy system dependencies, competing priorities — these don't get resolved at the working level. When leadership is genuinely in the room rather than nominally sponsoring the initiative, the roadblocks get cleared. When they're not, the project slows to the pace of the most resistant stakeholder.

An Unexpected Accessibility Win

One of the more surprising findings Roberson shared: older customers adopted T-Mobile's voice AI interfaces faster than expected, and for a reason the team hadn't anticipated. Traditional app navigation — menus, hierarchies, small touch targets — was genuinely difficult for many older users. Conversational voice interfaces bypassed that entirely.

"We thought younger users would adopt it first. We're seeing many older people use it. It's unlocking experiences for customers we didn't expect."

Conversational AI as an accessibility layer is an underexplored implication of this technology. When the interface is natural language, the barriers that trip up users in traditional UI — visual complexity, navigation logic, motor precision — largely disappear.

AI Transformation Is Organizational Redesign

Derek Ho closed with the framing that gave the session its title. Drawing on the Distyl AI experience across F500 customers, his argument was that successful AI transformation always comes down to two shifts: business model and operating model. Technology is the enabler. Organization is the work.

"Amazon isn't a bookstore." Companies undergoing AI transformation have to be willing to question what they fundamentally are and what they're organized to do — not just add AI capabilities to existing structures. Legacy software architectures were built for humans writing code. AI-native operating models require fundamentally different infrastructure, team structures, and decision-making processes.

"The architecture needs to change. The previous generation of software was built for engineers writing code." What comes next has to be built for something different.

This session was presented at the AI Agent Conference 2026 in New York. Speakers represented Distyl AI and T-Mobile.