More AI Tools = More Context Switching ft. Saumyo Mukherjee (Braze)

Saumyo Mukherjee, VP of Business Systems at Braze, joined host Josh Schachter for this week’s [Un]Churned episode, “202. Vibe Coding Is a Band-Aid on a Band-Aid.”

He talks about how to manage AI adoption in a 2,000-person go-to-market team, why vibe coding often causes more coordination issues than it fixes, and shares a simple rule for making AI decisions for Customer Success (CS) and beyond.

---

Listen on YouTube, Spotify, Apple Podcasts, and Gainsight.com.

Subscribe now and get the [Un]Churned take straight to your inbox.

---

🎯 The [Un]Churned Take: The Answer to AI Chaos Is Hiding in Plain Sight

Ever since AI tools entered the chat, we’ve seen a reinvigoration of debates around build vs. buy, native intelligence vs. Model Context Protocol (MCP), and point solutions vs. platform consolidation. In this week’s UnChurned episode, Saumyo Mukherjee, VP of Business Systems at Braze, shared his thoughts on balancing AI innovation with your existing tech stack.

Saumyo’s team manages the entire application stack for all external-facing teams: marketing, sales, BDRs, partnerships, support, CS, and services. He refers to his 24-person team as “the point of convergence.” Every tool decision, overlap, and gap between what a team needs and what the stack offers starts and ends with Business Systems. This bird’s-eye view into the entire GTM tech stack and how everything works together offers a unique perspective that a single team usually doesn’t get.

“I very carefully call it [Gainsight] a system of interaction, not a system of record. Because I want the user group, which is CSM in this example, to basically spend their day within that platform.”

A system of record just stores data. A system of interaction is where people actually do their daily work. For a Customer Success manager (CSM), Saumyo believes these should be in the same place.

What Context-Switching Actually Costs

When it comes to Customer Success, the debate over AI tools is often framed as a capability question. Does this tool surface the right signals? Does it write a better QBR prep? Can it predict churn more accurately than the one we have?

What tends to get overlooked is the fatigue that comes with context switching. When your CSM gets an alert, opens an account, prepares for a call, and logs notes, how many apps do they use? If it’s more than one or two, you have a context-switching problem that AI alone can’t solve.

“If I’m a CSM, I would rather like to wake up in the morning and log into a platform, let’s say Gainsight. And then, I should be able to log off at 5 p.m. from Gainsight without having to ‘swivel chair’ anywhere else. That is my ideal state of affairs for an end user like a CSM.”

Outside tools are useful, but every time someone switches between tools, it adds a small burden that slowly adds up. By the end of the day, a CSM who spends forty minutes switching between four tools limits the amount of impactful work they could’ve done in a single, complete environment.

Real progress comes from building within, not chasing every shiny new tool. The real challenge is harnessing this power without drowning in complexity, and that’s where smart governance comes in.

You Wouldn’t Put a Band-Aid Over a Band-Aid

When Braze rolled out access to Claude and Gemini, Saumyo’s team found themselves faced with what he referred to as “good chaos.” Teammates were inspired to be creative and explore the capabilities of their new AI tools, but without governance, nobody knew what was being built, where, and by whom. Without clear direction, CSMs can be the first to feel this chaos.

Let’s say your CS, RevOps, and product teams are all creating their own AI workflows across the customer journey. While each team may be solving their own part, without management for the gaps between workflows or handoffs, it can fall on CSMs to fill them.

“Oftentimes [vibe-coded, point solutions] become a band-aid over another band-aid... get rid of all that, build from scratch, a fresh solution that doesn’t require any other band-aids.”

Saumyo’s solution was to hold a weekly AI Idea Box meeting every Thursday. During these sessions, a core team reviews ideas based on business impact and effort before they spread. It’s not flashy, but it turns scattershot innovation into holistic additions to teams’ day-to-days.

Good governance doesn’t restrict innovation. It creates the guardrails and clarity that let teams use AI and existing features to their fullest potential, without overwhelming themselves in the process. Sometimes, the smartest move isn’t adding more, but making the most of what’s already there.

Finding Focus Within the Chaos

New AI tools are constantly launching, and this means the internal workflows being built with them will only compound. What separates the “chaotic” teams from the impactful ones is governance that filters out what actually improves workflows, rather than adding technical debt.

What updates have your existing platforms launched that can be utilized the same way as that shiny new AI platform? Are you really making building QBR decks and meeting preparation easier, or does it feel like you need to add a new monitor for every new program you have to use? Nobody wants to feel like the hacker in an action movie juggling a hundred screens, typing into a dozen systems at once, scrambling to keep up. When AI tools force CSMs to jump between platforms and piece together information, we’re not empowering them—we’re overwhelming them.

Real impact happens when everything a CSM needs is right in front of them in a single environment, so they can focus on customers.

---

🎧 Listen for These Moments

Saumyo and Josh talk about how to manage AI adoption in a complex GTM stack, and what CS leaders can learn by looking at how operations teams handle these challenges. Listen for these moments:

• Token budgets by department — Saumyo shares a simple way to show CFOs that AI investment is working. It includes giving every team an AI spending limit and adjusting from there.

• The quarterly hackathon — Braze runs a two-day event where anyone can pitch and build an AI idea. Saumyo explains why the person with the best idea usually isn’t who you’d expect.

• The handoff problem — Saumyo shares where Braze discovered customer relationships were quietly breaking down and what they did to fix it.

---

🔎 Where to Find the Speakers

• Saumyo Mukherjee — LinkedIn: https://www.linkedin.com/in/saumyo/

• Josh Schachter — LinkedIn: https://www.linkedin.com/in/jschachter/

📎 Referenced in This Episode

• Braze — the customer engagement platform where Saumyo leads Business Systems across all GTM-facing applications

• Gainsight MCP — referenced by Josh as part of Gainsight’s current AI architecture, connecting Gainsight, Staircase AI, and Skilljar data into your LLM of choice.

---

Wrapping Up

If your CSM had to describe their experience with AI in one word, would it be the same word as yours?

See you next week 🧠