Speaker
Ishan Mukherjee
Appearances over time
1 episodes
Episodes
1
Podcasts
Quotes & moments
ROX was founded on Valentine's Day 2024, making it roughly 2 years old at the time of this recording.
Ishan was an early member of Kiva Systems, which was acquired by Amazon and became Amazon Robotics — the Roomba-like robots that pick boxes in Amazon warehouses.
Ishan's Kubernetes infrastructure company Pixie Labs was acquired by software monitoring company New Relic in 2020.
ROX is built as a three-layer system: a context/knowledge-graph layer at the bottom, an agent swarms orchestration layer in the middle, and consumer applications on top.
ROX initially used AWS DynamoDB as its document store due to the founders' Amazon background, but migrated to MongoDB late last year when scaling demands required its additional features.
ROX uses MongoDB as its core document store for all unstructured data sources including emails, transcripts, call logs, and notes.
Ishan argues AI agents are only as good as the context they have access to, and most customer context lives in data warehouses — not in the CRM.
Ishan predicts the best engineers, marketers, and salespeople will be 2 to 5 times more productive in the AI era, while others will need to reskill.
ROX uses its own platform as its CRM, customer-facing toolchain, and general workflow system — running no other piece of software.
Given how fast AI is moving, ROX operates on weekly plans and daily execution rather than monthly planning cycles.
ROX pioneered the concept of agent swarms — an orchestration system where thousands to millions of pre-built agents communicate with each other to complete tasks.
ROX isn't a CRM upgrade — it's a replacement. Its AI agents autonomously handle customer research, outreach, engagement, and contracts, doing everything a human would have done inside a CRM without human input.
GPT, Claude, Gemini — they're available to everyone. The actual competitive moat is who can aggregate, activate, and make their data agent-ready. That's the insight driving ROX's entire architecture, and it's why systems like MongoDB matter more than model selection.
Every major CRM shift has redefined who does the work. Siebel put it on-premise; Salesforce moved it to the cloud; ROX removes the human entirely. The agent era isn't incremental — it's a category reset.
Think of ROX as a three-layer cake: a knowledge graph and governance system at the base, an agent swarms orchestration layer in the middle managing potentially millions of agents, and consumer-facing apps (web, mobile, Slack, ChatGPT) on top.
ROX started on AWS DynamoDB — the founders came from Amazon, so the choice was instinctive. But as scale hit, MongoDB's feature set and security profile pulled them over. Today it manages all of ROX's unstructured data: emails, call logs, transcripts, notes.
Technology stack decisions aren't one-size-fits-all — they depend entirely on what phase you're in. Early stage needs developer-friendly APIs and fast onboarding. At scale, you need architecture that handles enterprise spikes and fast queries. ROX learned this the hard way.
At ROX, there's no AI tool approval process — engineers use Cursor, Cognition, Claude Code, whatever helps. The whole company runs on ROX itself, with no other software. Eating your own dog food isn't a slogan here; it's the entire operating model.
Predicting AI six months out is nearly impossible — but the long-term direction is clear. The best performers will be 2 to 5x more productive; everyone else faces a reskilling imperative. ROX's response is to stop planning monthly and start executing daily.
Build with strong convictions at the infrastructure layer and relentless flexibility at the product layer. That's Ishan's formula for navigating an industry moving faster than anyone can fully predict.
Ishan Mukherjee has been building companies since before 'AI founder' was a job title. His path runs from MIT through Amazon Robotics and Apple's knowledge graph to two exits — and now ROX, the company he started on Valentine's Day 2024.
Analysis
What they talk about
- Technology 55%
- Business 36%
- Society & Culture 9%
