From Pilot to Production: Delivering Reliable Voice Experience - Voice AI Space Conference NYC

Overview of Speakers and Organizations

Diego Sandoval, Head of Business Development at Bespoken AI, and Melissa Rojo, Global Product Leader at LTM, discuss the current landscape of voice AI and the transition from pilot projects to full production.

LTM's AI Footprint and the Evolving Landscape

Melissa Rojo manages LTM's AI solutions, which span from small agents to multi-agent orchestrated workflows. LTM's "Blueverse AI Group" generates approximately $1 billion in business, utilizing a playground called "Foundry" to experiment with agent orchestration, model farms, and small language models before client deployment. Rojo notes that while the past two years focused heavily on AI experimentation and pilots, the current trend is moving toward proprietary, specialized solutions. This evolution progresses from simple chatbots to digital assistants, multi-agent workflows, and ultimately, digital employees.

Case Study: Voice AI in Retail Banking

Rojo shares an example of a voice AI implementation for a large international bank. Moving beyond standard customer service, LTM developed a voice agent capable of making proactive "hard sales" during inbound calls. In a pilot with U.S. students, the voice agent resolves a student's initial request (such as locking a card) and then analyzes their account history. If the agent detects a sustained increase in their checking balance, it dynamically suggests opening a high-yield savings account. This process relies on real-time data integration and reasoning rather than rigid, pre-programmed conversation flows.

The Rise of "Digital Employees"

The future of the workforce includes digital employees, AI agents with defined roles, authority, and responsibilities modeled after human jobs. Rojo explains how these digital employees can be integrated directly into communication platforms like Slack or Teams. Human employees can bring a digital employee into a call to assign tasks, and the digital employee can proactively suggest process improvements. Organizations can manage these AI agents through lifecycle stages similar to human HR processes, including onboarding, performance evaluation, promotion, and offboarding. Rojo emphasizes that voice capabilities, including empathy and reasoning, are essential for the existence of digital employees.

Key Advice for Enterprise Voice AI Projects

Rojo concludes with advice for organizations implementing voice AI:

• Voice is expanding beyond the contact center to intersect with every part of the business.

• AI projects rarely fail due to the technology itself; instead, they fail due to a lack of governance, undefined escalation matrices, misalignment with existing business processes, or a lack of internal champions.

• Before beginning technical implementation, enterprises must thoroughly map out their business processes and organizational structures to ensure a blueprint for success.