Built to last.
Not only to impress.
Every major computing shift creates a control layer. Cloud created security infrastructure. APIs created gateways. AI is the next shift — and the operational control layer does not yet exist as a standard category. We are building it.
The origin story.
Not a pivot. Not a market opportunity spotted in a report. A specific operational failure — experienced firsthand — that turned out to be universal.
It was 2am during a major sale at ShopClues. We were processing a volume of orders we'd never seen before. Customer service queues were exploding. Agents were overwhelmed. And I remember thinking: we have no system for this. Not for the hard cases. The failed deliveries, the payment disputes, the genuinely angry customers who needed a real decision — not a scripted response.
We managed it the way most operations teams do in those moments: heroics, tribal knowledge, people working through the night. And when it was over, I knew we'd solved nothing. We'd just survived it.
That pattern — AI handling the easy cases, humans drowning in the hard ones, nothing connecting them — is what we kept seeing as enterprises began deploying AI in customer operations. Not because the AI wasn't capable enough. Because there was no operational layer around it. No system for deciding what AI resolves, what it escalates, how context survives handoffs, or how the whole thing improves from real production experience.
CygnusAlpha is the company we needed then. Built specifically for the operational gap that every serious AI deployment eventually hits — and that nobody had built infrastructure to solve.
Four convictions that shape everything we build.
These are not values on a wall. They are the intellectual positions that determine how we design our products, structure our engagements, and make decisions when things get hard.
AI capability is not the bottleneck.
The models are good enough. GPT-4, Claude, Gemini — they are all remarkably capable at answering questions. The constraint is not computational. It is operational. Organisations lack the infrastructure to define what AI decides, what it escalates, and how human judgment gets incorporated back into the system. That is what we build.
Pilots succeed. Productions fail. Those are different problems.
Most AI deployments look promising in a controlled pilot environment. The failure mode is not at go-live — it is three months later, when volume increases, edge cases multiply, and the operational scaffolding that was never built starts to show its absence. We design for production from day one, not for demo success.
Vendor dependency is not a business model.
We are structured to earn the right to stay — not to make leaving hard. The Build-Operate-Transfer model is how we build trust: we transfer knowledge and capability openly, so clients remain with us because the platform keeps delivering value, not because switching is painful. Aligned incentives produce better operations. That is what we build for.
The companies that win with AI will have the best operations, not the best models.
Competitive advantage in the AI era will not come from which model you use — it will come from how well you have designed the operational system around it. Escalation logic, autonomy boundaries, learning loops, supervisor workflows — these are the moat. They compound over time. They are hard to copy. And they require someone to build them deliberately.
Operators and builders.
Not researchers.
Everyone on the founding team has built and scaled something hard in production. That is not a coincidence — it is a hiring principle.
Sanjay Sethi
Sanjay co-founded ShopClues, one of India's first e-commerce unicorns, where he built and scaled customer operations through the complexity of high-volume, high-stakes commerce — processing millions of orders and managing customer operations at a scale where every failure had an immediate revenue consequence.
He also served as Global head of Product for Shipping & Logistics at eBay and as Global CTO for Wish.com, giving him deep exposure to how AI and operational reality collide across global markets. CygnusAlpha is built on what he learned — not from research, but from running operations where the cost of getting it wrong was real and immediate.
LinkedIn →Divye Tela
Divye founded Aurigait, an enterprise software company, where he built and delivered complex technical platforms for clients across global markets. His background is in the hard, unglamorous work of enterprise software delivery — the kind where requirements are ambiguous, integrations are messy, and production reliability is non-negotiable.
At CygnusAlpha, Divye leads the business and delivery function — ensuring that what gets promised in the sales conversation is what gets built and operated in production. His operating principle: nothing ships that the team cannot stand behind in a client review.
LinkedIn →Vivek Chopra
Former CEO — Wipro & L&T InfotechVivek brings one of the most distinctive enterprise scaling credentials available to an early-stage company. As CEO of Wipro and L&T Infotech, Group President at DXC, SVP at HPE, and EVP at IBM Daksh — and as Chair of the Boards at Knolskape and Mastek Inc. — he has built and led enterprise technology organisations through every stage of growth. His advisory role at CygnusAlpha is focused on enterprise go-to-market, scaling the delivery model, and ensuring the company is built with the governance and operating discipline required for long-term enterprise trust. An alumnus of IIT Delhi and Kellogg School of Management.
LinkedIn →We are in the hardest phase of the transition.
Every major technology transition has a moment where the technology is capable but the operational infrastructure around it hasn't caught up. That gap — between capability and operational reliability — is where companies either build enduring advantage or quietly fail.
We are in that moment for AI in customer operations. The models are capable. Enterprises are under pressure to deploy. But the operational layer — the infrastructure that makes AI reliable in live production — does not yet exist as a category. It is being built, badly, inside every enterprise that is serious about deploying AI.
CygnusAlpha is building that control layer. Every major computing shift creates infrastructure that governs the new technology at scale: cloud created security, APIs created gateways, data created observability tooling. AI is next — and the company that embeds first, at the moment enterprises most need it, builds a position that compounds with every deployment.
Human-only operations
Large customer ops teams. Expensive. Difficult to scale. But controllable and accountable. Enterprises understood how to manage it.
Hybrid operations ← We are here
AI handles the simple volume. Humans manage the complex cases. Nothing connects them properly. This is the hardest phase — and the one that requires the operational layer CygnusAlpha builds.
AI-native operations
Operations designed from the ground up for AI-first execution. The companies that navigate hybrid well will own this phase. The ones that don't will be starting over.
The commitments we hold ourselves to.
Not aspirational values. Actual operating principles that shape how we behave with clients, prospects, and each other — especially when it's inconvenient.
We tell you when we're not the right fit
If your problem doesn't match what we build, we say so in the first conversation. We would rather disqualify quickly than start an engagement that isn't right for either side.
We measure success by outcomes, not activity
Reports, documents, and go-live celebrations are not outcomes. Resolution rates going up, escalation quality improving, handle time coming down — those are outcomes. That is what we track.
We earn the right to stay
Clients remain with CygnusAlpha because the platform keeps delivering value — not because they are locked in. We transfer knowledge and capability openly, so the decision to stay is always theirs. That shapes every decision we make.
We treat production accountability as non-negotiable
If outcomes are not tracking to agreed targets, we go back into Operate — at our cost — until they are. We do not hand over an operation that isn't working. No exceptions.
We work in the open with clients
No black boxes. Every design decision is explained. Every metric is shared. Your team understands what was built and why — so they can own it when we step back.
We focus narrow and go deep
We do not try to be the AI platform for everything. We are the operational layer for hybrid customer operations. The narrower the focus, the deeper the expertise — and the more defensible the outcome.
If you're building for the long term,
so are we.
We take on a small number of clients at a time. If the problem fits, we'll tell you quickly — and if it doesn't, we'll tell you that too.