Using AI to Transform Customer Success

Karl Mosgofian doesn’t talk about AI as a buzzword or a quick fix. For the CIO of Gainsight, a company built around customer success, artificial intelligence represents both a reality and a responsibility. In his view, AI is reshaping how companies understand their customers, how internal teams collaborate on innovation, and how leaders balance expectation with practical application.

He frames the conversation by separating the hype from the real capability. That distinction matters not just for technologists, but for business operators in customer-centric organizations. As he says, “Some people are acting like AI was just invented, and it wasn’t.” Today’s enterprise leaders must navigate both the illusion and the substance of generative intelligence.

Why Separating Hype from Reality Matters

In many boardrooms and project plans, AI has become shorthand for “innovation.” Yet Mosgofian pushes back against that simplistic view. He notes that AI is not new; parts of it have existed in enterprise systems for years. “Gainsight has had AI in our products for a long time,” he explains.

But the conversation changed with the advent of large language models like ChatGPT. LLMs “are really different,” he says, and they do “some things that previous technologies couldn’t do,” yet they also expose how inflated expectations can become.

Mosgofian compares this phase of AI to past technology waves where early excitement eventually gave way to a more practical equilibrium. He argues that AI will deliver impact, but only when leaders understand both its possibilities and limitations.

From Summarization to Customer Insights

One of the most compelling parts of Mosgofian’s experience is how AI is already being put to work in customer success contexts.

He describes a simple but powerful generative use case: summarizing dense customer data into actionable insights. Enterprises often collect vast quantities of information about customers: usage metrics, support tickets, contract histories, feedback, but it can be too much for humans to synthesize quickly.

“A good example of that … is we could shove all of that [data] into an LLM and just have it summarize it and essentially give what we call a customer cheat sheet,” Mosgofian explains. With generative models, executives or account teams can get concise, high-value summaries when prepping for customer meetings, without wading through pages of raw data.

This use case is illustrative: the value of generative AI isn’t in replacing human judgment, but in augmenting it. IT teams don’t need data scientists to implement this workflow; any team member can leverage the summarization capability to accelerate understanding and decision-making.

Realistic Views on Chatbots and Productivity

While acknowledging the progress in AI-driven conversational agents, Mosgofian is cautious about overhyping their impact. He notes that chatbots, even powerful ones, aren’t a substitute for human support teams. Earlier chatbot generations didn’t eliminate support roles; they shifted the nature of engagement, and Mosgofian believes the same will be true for AI-enhanced agents.

“Is it going to be enough to make a significant impact? … Or is it going to be a nice productivity boost?” he asks, framing a balanced view that AI can help a lot without instantly transforming entire support paradigms.

That realism matters for customer success leaders evaluating AI investments. Mosgofian’s perspective is grounded in outcomes, like productivity and quality, not hype. His position reflects a broader theme: AI should improve human work, not pretend to replace it.

Data Preparedness

A recurring theme in Mosgofian’s remarks is that AI’s utility depends on the quality and structure of underlying data. AI models can’t extract insight from poor inputs, and many enterprises discover that their data isn’t ready for scaled AI use when they first try to deploy it. “We can have this chatbot that works super well, but we need to give it a base of knowledge base articles to work off of and we don’t have those,” Mosgofian says.

This emphasis on data is critical. It’s not about flashy demos; it’s about ensuring that enterprise data systems, classification schemes, and governance practices support meaningful AI outcomes. For customer success platforms like Gainsight, where contacts, usage histories, and behavioral signals are central to value delivery, data cleanliness and accessibility become prerequisites for AI success.

Governance, Coordination and Democratized Innovation

Another insight from Mosgofian is the organizational challenge of AI adoption. With tools like ChatGPT accessible to everyone, innovation risks becoming fragmented and ungoverned. That’s why Mosgofian established an internal AI focal point and broader cross-company working group. “The danger … is when anybody can go to ChatGPT, you’ve got all these different people … doing cool stuff, but nobody’s talking to each other,” he explains.

Rather than imposing rigid restrictions, Mosgofian’s team embraced coordination: regular meetings, shared Slack channels, and cross-functional conversations about AI use cases and roadmaps. The goal wasn’t to control every AI initiative, but to facilitate them responsibly while ensuring security, collaboration, and shared learning.

IT as Partner, Not Gatekeeper

What stands out in Mosgofian’s approach is his emphasis on IT as an internal customer success partner. He likens the role of IT teams working with business units to the way customer success teams partner with external customers: not as order takers, but as collaborators who help drive outcomes.

This requires two shifts. First, IT must understand business context deeply, including priorities and workflows, to apply technology that genuinely supports goals. Second, IT must build guardrails (like security and governance) without stifling innovation.

“We need to secure everything … but the interesting thing … is there’s a lot of just people in Slack saying, ‘Hey, I’m looking for a tool, does someone have something that does this?’” Mosgofian recounts, highlighting a collaborative culture rather than command-and-control.

The Future of AI in Customer Success

Karl Mosgofian strikes a balance between excitement and pragmatism. On one hand, he says, “ChatGPT is maybe the most magical technology I have ever seen,” and looks forward to future advances. On the other, he warns against overselling capabilities or expecting AI to instantly solve structural challenges.

His experience suggests that AI value emerges incrementally: through summarization features, productivity enhancements, support augmentation, improved collaboration, and better data insights, not by flipping a switch. Enterprise leaders who treat AI as a strategic tool, not a silver bullet, will see meaningful outcomes.

Lessons Learned

Across his conversation, Mosgofian’s insights reflect four key lessons for enterprise organizations embracing AI:

1. Manage expectations: understand the difference between hype and current capabilities. 
2. Build on solid data foundations: AI is only as good as the data it consumes.
3. Coordinate innovation: establish governance and collaboration without throttling creativity.
4. Focus on human augmentation: leverage AI to enhance, not replace, customer and internal teams.
   

For leaders steering customer success strategy, or indeed any business unit where relationships, insight, and interaction matter, AI is not a distraction. It is a capability that, when grounded in reality and integrated thoughtfully, can elevate how organizations serve customers and make decisions.