The CS Function Will Still Exist. The Shape Will Look Different

Ishmeet Singh has spent over a decade building customer success at MuleSoft, a Salesforce company. He joined in 2014, back when CS was still a nascent function outside of Salesforce itself, and rose from an individual contributor working with large banks and regulated industries to Chief Success Officer, leading global customer success across a multi-billion dollar business. Before MuleSoft, he cut his teeth as a developer at IBM and spent nearly a decade in management consulting at A.T. Kearney. He holds an MBA from Kellogg at Northwestern.

Not many CS leaders can trace their career back to writing code at IBM, then through management consulting, and into building a customer success org from scratch at a company that grew into a multi-billion dollar Salesforce business. That path gives Ishmeet a rare vantage point on what AI can actually do for post-sale teams, and what it can't.

From Greenfield to Playbook

When Ishmeet joined MuleSoft, the CS function barely existed at the company, or at most of their competitors. It was a greenfield opportunity to shape the charter, the playbooks, the entire operating structure from scratch. That's what drew him in.

What shaped his thinking most was working directly with customers. He started as an IC responsible for the success of large enterprise accounts. His background as a developer at IBM meant he could speak the language of the architects and engineers he was working with, while also understanding what CIOs needed to hear about realizing and communicating value.

At the time, MuleSoft was disrupting traditional middleware with a lighter-weight approach. But customers still carried the muscle memory of heavy implementations: the tech debt, the slow timelines, the painful migrations. They kept asking: how are we going to do this differently? How do we avoid repeating what happened before?

That pressure pushed Ishmeet and his team to build something he's still proud of: a repeatable success methodology. Playbooks for developer adoption and reuse. Frameworks for measuring the value of what customers built. Constructs that could be programmatized, taught to the field, and scaled. That methodical, evidence-based instinct (build it, prove it, repeat it) is the same muscle he's now applying to AI.

What's Actually Working

When asked about his first real AI experiment, Ishmeet tells a story that will sound familiar to any CS leader who's been asked to pull a rabbit out of a hat. An executive came to him and said: summarize the full history of a strained customer relationship within the hour. Normally, that would mean organizing calls with multiple CSMs, account executives, and former employees. Days of work compressed into a scramble.

Instead, Ishmeet turned to an AI-powered tool integrated with their enterprise systems. Within 30 to 45 minutes, he had a complete, C-level-ready summary: the customer's history, what they bought, how they used the platform, and why they chose to go in a different direction. All generated by querying internal systems and using AI to synthesize and format the output with the guardrails he specified.

"These are the kinds of things that would require multiple meetings with CSMs, with account executives. And within 30 to 45 minutes, I could actually get a summary that I could deliver to a C-suite executive." — Ishmeet Singh, Chief Success Officer at MuleSoft (Salesforce)

But the bigger bet — and the one Ishmeet calls the most important implementation to date — is using AI to build a predictive, multi-signal view of customer health. His team is integrating product telemetry, customer sentiment from calls, and even publicly available data on executive turnover to generate propensity scores for churn and adoption velocity. No single signal tells the full story, but stitched together with AI, the picture becomes far clearer than anything a human could assemble manually.

And then there's what he sees as the real frontier: agentic AI inside the product itself. Instead of waiting for a CSM to reach out about onboarding, an agent embedded in the application can prompt a new customer the moment they log in — guiding them to resources, connecting them with the right people, surfacing methodology that already exists but was previously buried in content repositories.

"Before I get to a point where a customer's adoption is off track, I can today, with agentic AI, help a customer stay on track by prompting them proactively within the application." — Ishmeet Singh, Chief Success Officer at MuleSoft (Salesforce)

His team has also started using AI to analyze customers' existing API deployments and show them how those same assets could power their agentic transformation initiatives. For many large enterprises, APIs built a decade ago are now the building blocks of their AI strategy — but they don't always know it. When Ishmeet's team surfaces those insights, the conversations change entirely.

The Hard-Won Lessons

At the highest level, Ishmeet says the impact comes down to a few things: reducing attrition, driving deeper product adoption, and showing up to customers with insights they can't generate themselves. That last piece is what's changing the nature of engagement. Customers who previously didn't want to engage — either because they were satisfied or distracted by other priorities — are now re-engaging because the CS team can bring them something genuinely new.

The lesson isn't just about the technology. It's about what the technology makes possible: changing the conversation from "how are you using us?" to "here's how you could use us even more — and here's the proof."

Ishmeet is also clear-eyed about timing. He's not cutting any experiments yet. This, he believes, is the moment to play and experiment — not to prematurely optimize. What felt new six months ago is now part of the daily workflow. The tools that seemed advanced are becoming table stakes.

What's Next

Ishmeet's advice to CS leaders navigating this shift is simple and direct: use the technology every single day. Use it for every customer problem. If you're not using it by default, ask yourself why. The fears about the function becoming obsolete miss the point. As long as you're learning the tech, you'll find where you fit.

"The function will still exist — the shape of it will look different. Which is no different for any other function today in the industry. You have to understand where those gaps are and be able to address them so you can deliver value to your customers." — Ishmeet Singh, Chief Success Officer at MuleSoft (Salesforce)

There will always be a gap between how a product works and how a customer understands how to use it. That gap may be changing — but it isn't going away. The CS leaders who thrive will be the ones who find the new gaps and close them.