When the Tools Don't Fit, Build Your Own

Aditi Kapur has spent over a decade on the front lines of customer success, starting as a software test engineer at Infosys before moving through technical support and into strategic CS roles at companies like Airship, Veracode, and Solace. Today she's a Staff Technical Customer Success Manager at Kong, covering the APAC region from Singapore, where she holds an MBA from the National University of Singapore alongside her B.Tech from JNTU.

What makes Aditi's story stand out isn't just her career arc from reactive support to strategic CS — it's what she did when the tools she needed didn't exist. Her journey with AI in customer success is the kind of scrappy, builder-driven story that doesn't show up in vendor case studies.

From Reactive to Proactive — Where It All Started

Before AI was a buzzword, Aditi was already wrestling with the problem it would eventually help solve: how do you get faster, better answers to customers without making them wait?

Working in technical support, her team ran what she now recognizes as an early AI experiment — a case deflection system that surfaced relevant knowledge base articles as customers typed their support queries. The team built a discipline: every time a case came in, if a matching KB article existed, attach it. If not, write one. Over time, the system got smarter. Cases that never got created became the signal that the approach was working.

The foundation underneath all of it, though, wasn't technology. It was trust. As Aditi puts it, whether you're doing reactive or proactive customer engagement, the real work is building relationships with your champions — following up when you don't have the answer, being consistent, showing up. That belief carried through every role and shaped how she would eventually approach AI tooling: not as a replacement for relationships, but as a way to protect the time you spend on them.

Building Client Pulse — Because Nothing Else Fit

When Aditi arrived at Kong, she ran into a problem familiar to CSMs at non-SaaS companies: the standard CS playbook doesn't quite apply. Kong's customers aren't all on the same version. Adoption timelines vary wildly — one customer might go live in a week, another on the last day before renewal. The consumption data that matters, like how many API calls are being made or which enterprise plugins are deployed, doesn't just show up in a dashboard.

Existing CRM tools, built for the SaaS world, couldn't capture the nuance. So Aditi built her own.

Using Base44 and an agentic IDE called Antigravity, she created Client Pulse — a local web app that pulls together everything a strategic CSM needs in one place: account health metrics, support tickets filtered to the three riskiest items, consumption data, renewal timelines, even a sticky-note feature for quick task tracking. She built a JSON parser over a weekend that analyzes license consumption files and projects API call volume.

"The AI helped me to build something that was missing. I'm not a coder — I've been a tester though, so that probably came to my advantage." — Aditi Kapur, Staff Technical Customer Success Manager at Kong

The response from her team was immediate. Other CSMs saw it and wanted the binary. But this is where the classic problem hit: a locally-built prototype, no matter how useful, doesn't give leadership visibility. It's not enterprise-grade. The next step — integrating with Salesforce so health metrics and summaries update in real time — would require a level of plumbing that goes beyond what one CSM can reasonably do on weekends.

The clearest signal that Client Pulse was working came from consumption tracking. At Kong, getting usage reports from customers is tricky because customers often don't care about the data the CSM needs. Aditi built a triggering mechanism into her dashboard that flags overdue reports every morning. That persistent nudge led to catching customers in overages and identifying expansion opportunities that would have otherwise gone unnoticed.

"The impact is I'm not running blind in these accounts that are coming up for renewal — because I'm literally chasing the data for it." — Aditi Kapur, Staff Technical Customer Success Manager at Kong

The Lessons That Stuck

The experiments that failed taught Aditi as much as the ones that worked. Her attempt to build a real-time integration between Client Pulse and her corporate folders crashed the entire tool. The plumbing — API keys, field mapping, multi-step integrations — was more than she could manage without engineering support. What she really needed, she realized, was a low-code system where she could just point at a data source, map the fields, and go.

The deeper lesson is about the gap between what individual CSMs can build and what an organization can actually adopt. A prototype that runs on your laptop is powerful for one person. But customer success at scale needs tools that feed into the systems leadership already watches — and that means the integration layer matters as much as the insight layer.

There's also the question of trust with AI itself. Aditi works in APAC, far from the corporate headquarters where approved AI tools tend to land first. She and her peers want the peace of mind that whatever they're feeding into AI — account research, raw notes, customer data — lives in a vetted, company-approved environment. The experimentation appetite is there. The guardrails aren't always keeping pace.

What's on the Horizon

Aditi's wishlist is specific. She wants AI agents that run continuous account research — scanning annual reports, tracking champion changes, flagging distress signals or expansion opportunities — and feed those insights directly into a dashboard. She wants automated QBR prep: take the consumption data, the notes, the adoption metrics, the business objectives, and generate a presentation skeleton in minutes instead of hours.

None of this is science fiction. It's the natural next step from what she's already built by hand.

"If you want to be impactful, you can use this to your advantage and be impactful. Because you can never take away the human element from this." — Aditi Kapur, Staff Technical Customer Success Manager at Kong

That's the thread running through Aditi's entire story: AI doesn't replace the CSM. It gives the CSM back the time and clarity to do the work that actually matters — building trust, deepening relationships, and showing up for customers with real insight instead of guesswork.