Avaya CTO David Funck on AI's Persistent Memory with Model Context Protocol (MCP)

The Challenge of AI Sprawl

• 🤯 The current AI landscape is dizzying and rapidly changing, making it difficult for companies to keep up with emerging leaders and prepare for the future.
• 🧩 A major challenge is AI sprawl, where numerous AI features and agentic components exist in silos, leading to fragmented customer experiences.
• ⚠️ Customers often suffer the most from these fragmented journeys, experiencing interactions that feel like starting from scratch each time.

Understanding Model Context Protocol (MCP)

• 💡 MCP (Model Context Protocol) is a protocol established by Anthropic that standardizes how large language models (LLMs) interact with the real world.
• 🧠 LLMs are trained on vast amounts of text but can only predict the next token; MCP allows them to reach out and ask questions or take action in the real world.
• 🚀 This capability is critical for agentic AI, enabling AI to not only suggest but also act based on instructions, taking LLM capabilities to the next level.
• 🌐 MCP standardizes these interactions, democratizing AI by reducing the need for extensive technology input and allowing LLMs to connect with enterprise APIs and internet resources.

MCP's Impact on Customer Experience

• 🎯 MCP empowers agentic AI to adapt to specific customer needs and tailor responses, moving beyond prescriptive paths to create a true dialogue.
• 🤝 Avaya's approach, termed "Tandem Care," envisions AI and humans working together to improve customer care, handling routine tasks while augmenting human agents for higher-order thinking.
• 📈 In contact centers, AI is well-suited for repetitive tasks, freeing human agents to focus on empathy and complex problem-solving for a better customer experience.

Measuring AI Success and ROI

• 📊 Contact centers provide a constrained environment with sophisticated measurement tools, ideal for evaluating AI effectiveness using metrics like Average Handle Time and First Contact Resolution.
• 💰 MCP enables measurement of cost-effectiveness by understanding the cost of AI and comparing it to human agent costs.
• 🔬 Enterprises can create focused LLMs trained on specific domains, making them more cost-effective to run than large, generic models, allowing for clear ROI calculations.

Future of MCP and Enterprise Impact

• 🚀 Avaya is currently demoing MCP, with pilots available at the end of Q1 and production availability in calendar Q2 of 2026.
• 🛠️ Beyond contact centers, MCP has wide-ranging potential across the enterprise, helping product organizations document APIs and enabling LLMs to interact with disparate back-office tools like code repositories and ticketing systems.
• ✨ The increased productivity unlocked by MCP can lead to better experiences and outcomes for everyone, fostering a collaborative partnership between humans and AI.
• 🎯 In the next year, the focus will be on more fit-specific and focused AI models that excel at defined tasks cost-effectively, rather than solely on general intelligence.