November 25, 2024

The new Q&A: question and action

Customer service agents will spark a new UX for all software 

There are reportedly 17 million people worldwide who work in contact centers, three million in the US alone. It’s one of the most common jobs and a critical function for all businesses. Unsurprisingly, there is a long list of billion-dollar software companies addressing the category - from core contact center systems like Genesys, NICE, and Five9, to omnichannel and ticketing systems like Zendesk, Intercom, Freshworks, and many others. Yet, much of the process remains broken, with poor customer experiences and high employee turnover. 

‍

AI to the rescue

Enterprises recognize this and commonly cite customer support as the “most urgent” use of generative AI spending. Our recent survey showed that 80% of GTM leaders thought customer support was the role AI would impact the most. Anecdotally, CX leaders that recently adopted next-gen AI are seeing improvements across all parts of the process: knowledge gathering, ticket deflection, better AI conversations, and now complex agentic actions that are already impacting hiring plans. Given the clear business desire, market size, and poor customer & employee experience, you have the makings of a screamingly obvious and high ROI use case for AI. 

‍

‍

But there’s even more to customer support than the large market opportunity at hand. Whoever wins in customer support AI may also define the new UX paradigm for all software.

Did this pique your interest? It should. 

‍

Question & action will become the dominant UX

Customer support AI doesn’t just herald a better customer experience that finally transforms this field with better technology—it’s about building an engine that can do more complex tasks and much more than customer support. Once AI agents can effectively take accurate and complex actions across multiple systems, customer support is just the tip of the iceberg.

Most of our interactions with software are some version of 1) question and answer (looking for a data point) or 2) taking an action (entering or changing data). Now, imagine a future where AI agents seamlessly navigate multiple systems to answer questions and take actions based on natural language requests. Gone are the days of point and clicking, and instead we have a new UX for software.

There will certainly be a lot of competition, but the agents that see the most interactions, traverse the most data and integrations, and attract the most revenue will continuously improve and win adjacent AI agent markets over time. This is the opportunity and core functionality of customer support agents—question and answer/action—and it may very well be the most critical workflow that defines the agent economy. 

‍

"Illustration of the evolution of AI agents in a pyramid structure, labeled with three stages: 'Today,' 'Superagents,' and 'Future.' The bottom layer, 'Today,' includes professions impacted by AI such as SDR, copywriter, researcher, accountant, customer service, software engineer, and more. The middle layer, 'Superagents,' highlights capabilities like 'Prospecting & Outreach,' 'Question & Answer/Action,' 'Generative Assets,' and 'Search & Summarize.' The top layer, 'Future,' features advanced AI categories such as 'General Purpose Agents' and 'Vertical Full AI Stack,' culminating in 'Superintelligence' at the peak.

‍