Skip to main content

In Search of AI’s Next Killer CX App

No job function is under more top-down pressure to cut costs with AI than Customer Experience. 

CX teams have been put through the wringer over the last two years.  The tech downturn hit net retention sharply, with median NDR falling nearly 10 points, sending organizations, public and private, into a new mode of operating. Growth At All Costs became Path To Profitability as FCF margin raced upwards to offset declining revenue growth.  

Customer Support, Customer Success and other CX teams responsible for driving customer value realization and retention became the first frontiers for maximizing efficiency. Amongst startups, Customer Success teams had the highest rates of layoffs. Champions of CX like Jason Lemkin were calling it the End of Customer Success As We Knew It.

Just in time, GenAI has shot into our consciousness like a silver bullet. CEOs salivated over Klarna adding $40M to their bottom line by deploying a ChatGPT-powered chatbot to handle the work of 700 Customer Support reps. CX leaders are now racing to find ways to leverage AI to create efficiency in their orgs. It is clear that the Customer Support function is being massively disrupted, but outside of support case resolution, leaders are wondering – what are the other killer AI apps for CX?

I gathered some of the smartest CX leaders I know to address this question. Here are my learnings:

Real Customer Support transformation currently requires a DIY approach. The Buy vs Build question loomed large in my conversations. There are two types of Customer Support orgs right now.

  • One has built their support case resolution tooling in-house, almost always on GPT, and is achieving massive improvement (50%+) in case resolution. They are now working on how to get the next ~10-20 points of case resolution improvement. They are thinking about how to best navigate R&D resourcing and how to optimize around bringing in a human rep on the right cases.
  • The other is leveraging third-party software and is seeing minor to moderate improvements in case resolution. There is a strong relationship between level of improvement and level of effort or cost to deploy / train the tool. They are wondering if it is possible to achieve better outcomes with third-party software and, if not, how the hell they are going to wrangle resources and expertise internally to build in-house.
    • The current feeling amongst CX leaders is that third-party offerings are iterative vs transformative, and if you want transformative then you have to Do It Yourself (DIY). But there is a ton of interest in vendors that can close this gap. Assembled is one such Emergence Capital portfolio company seeing promising results on this front, leveraging the  proprietary data from their workforce management platform to better optimize case resolution with AI.

We’ve barely begun to scratch the surface on using AI for other CX departments. While everyone is deploying technology and systems to improve Customer Support efficiency, other functions like Customer Success and Implementation are not.  There is a fair amount of hand-wringing as we collectively seek game-changing applications of AI in these job functions. But to some, it feels like “technology in search of a use case.” Leaders can’t agree on a single next killer app after support case resolution, but there were some common themes of what CX leaders want next from AI:

  • Customer research and summarization: Good reps spend a lot of time researching the customer externally and internally to prepare for customer conversations and pitches. Next, we want that work done by AI. Note though: we all feel like we have big data integrity problems. Our various internal systems conflict with one another and we are very skeptical of technology’s ability to address this problem with our source material. Guru, with its Enterprise AI search tool, is an Emergence Capital portfolio company addressing this need.
  • Customer onboarding: Many customer onboarding tasks are very repetitive, but we have tended to use human-led onboarding because great in–product onboarding is expensive. Next, we want AI agent-led digital onboarding.
  • Customer information logging: We cruelly ask AEs and CSMs to be archivists. Of customer contacts, of sentiment, of product feedback etc etc etc. Next, we want relevant information logged on their behalf.
  • Cross-sell enablement:  We inundate our customers with cross-sell marketing and discovery, often in what can be described as “spray and pray.” Next, we want more intelligent lifecycle marketing and ‘next best product’ recommendations and timing for CSMs.
  • We’ve collectively come to the realization that rather than looking for a solution that is going to transform the job function, we will likely be cobbling together incremental efficiencies and amplifiers. For the near term future, we are dealing with stairs, not an elevator. 

Long term, CX’s core competency of product knowledge becomes irrelevant, but its scope of control may expand. Much to our collective chagrin, customer-facing reps are all constantly pulled into very tactical product-related conversations. We train our CSMs to engage executives and uplevel conversations to strategic level, but they so often get pulled back down into the weeds, relying on their product knowledge to deliver value to the customer.  In the next 3-5 years, we expect AI agents will become the product functionality experts, entirely replacing much of the work done by CX teams in answering product questions and designing workflows and workarounds.

But CX is not the only function that is being substantially disrupted. The engineering function is being upended as new technology commodizes coding skills.  In that future state where transactional customer work is delivered by AI agents and technical product work is orchestrated by non-technical humans, we might find that a CX org’s deep understanding of customer need puts it in a strong position to direct the development of the product.

So what does this mean for the future of the CX job functions?

  • We likely have too many roles in CX. Some companies have customer support, customer success, professional services, sales/account management, and/or technical account management all interacting with the customer. We expect these roles to consolidate down to 1-2 customer-facing roles.
  • Customer relationship management and strategic consulting become even more critically important for our customer-facing employees. CSMs (or whatever we call these folks in the future) will spend the majority of their time in direct customer conversation, versus the current state of a couple of hours, max, per day.
  • If the Customer Support function continues to exist, it will likely become a far more technical function, taking over responsibilities currently held by engineering. Great Customer Support teams are already on this path today.

If you are reading this and want to weigh in, please reach out to me, on Linkedin.  I am looking to host a few more round table discussions on this topic in SF and potentially virtually as well. If you are a founder building something that fits the asks above, Emergence Capital would love to hear from you. Send a note anytime to or

Thank you to Dan Hayward, Jenny Sha, Jesse Dailey, Kelly Bray, Lauren Mullenholz, and Sarah Bernard for contributing to this discussion!