Most of today's focus for customer service operations is on automation: using artificial intelligence agents to augment and replace customer service reps. But there is a bigger story at hand: how AI reshapes the greater customer service technology landscape that is now forcing every vendor to offer scalable and affordable AI within its products.
There's a new AI tech stack forming. This is what it looks like:
- The experience layer becomes a smart, adaptive interface for work. Customer service reps use AI assist tools in their flow of work to summarize customer and ticket details, understand customer sentiment, draft service replies, and understand trending issues. Tomorrow, AI will take over an increasing amount of the work that human agents do and will eventually impact the jobs to be done in the contact center The future user experience for customer service technologies will be purpose-built to support the new jobs< and will dynamically adjust to the task at hand. Few vendors at this time think differently about their products' UX. Push your vendors to do so.
- The agentic fabric orchestrates and optimizes AI agents. This is where AI agents from vendors or partners or custom developed are coordinated, orchestrated, and optimized to deliver real business outcomes. This will evolve. Your core customer service and contact center vendors all offer AI agents and are battling it out to own the orchestration of AI agents. Ask your vendors how their AI Agents will align with interoperability standards and which level of coordination and orchestration they offer.
- The application mesh interconnects components, workflows, and platforms. Monolithic customer service applications will get deconstructed into discrete components and workflows that can be individually monetized. The application mesh will evolve into a set of vendor-provided microservices, deterministic and compliance-based workflows, and custom components built via low-code platforms. AI agents will use capabilities from the application mesh within their processes. Ask your vendors about the future of their monolithic products.
- The intelligence layer. This is the new layer. The intelligence layer carries all the risks of a new technology category that changes every week. It also carries four major risks: 1) Hyperscaler/model vendor tie-ups (e.g., Microsoft/OpenAI, Amazon/Anthropic); 2) quality controls and guardrails are just emerging; 3) you might have sovereign AI requirements across hosting, models, and assurances; and 4) using dozens and possibly hundreds of models will make model operations and version control more of a nightmare than it already is. Ask your vendors for their strategy to support foundational models, including trust and governance.
- The data fabric with semantic intelligence, which powers AI decisions. A data fabric will unify, govern, and operationalize data across silos, clouds, and formats. It will include a semantic layer. This is the foundation of your differentiated knowledge and expertise. Vendors like Databricks and Snowflake are building cloud-agnostic data fabrics to support metadata catalogs, vector databases, and knowledge graphs. Ask your vendors about their data strategies and whether they involve copying data from one repository to another.
Customer service organizations will deploy more AI Agents in their processes, leading to agent sprawl, security risks due to ungoverned agents, and escalating AI costs. Before you invest in AI agents, take a step back and set up an AI center of excellence that will help you decide which vendors and technologies to invest in, your AI strategy, and use cases that yield a real return on investment and governance.
Kate Leggett is a vice president and principal analyst at Forrester Research.