Nalin Gupta has been through Y Combinator before. His last company, an autonomous shuttle venture called Auro Robotics, made it through the accelerator in 2015 and was acquired by Ridecell in 2017. This time around, he's aiming at a bigger target: the sprawling, expensive, notoriously frustrating world of enterprise customer support.
His vehicle is Cignara, a four-person outfit that emerged from stealth in January under the name Bujo before quietly rebranding. The pitch is familiar enough—AI voice and chat agents that handle customer inquiries without humans—but the customer list, at least according to Gupta, is not. He claims Fortune 500 companies across retail, banking, travel, and telecom are already paying customers. And if a recent LinkedIn post is to be believed, the startup hit what Gupta reported as "multiple six figures in revenue" and profitability around January 2026, though these figures remain unverified.
Whether those metrics hold up under scrutiny is anyone's guess. But in a contact center market suddenly crowded with well-funded competitors, Cignara's lean structure and claimed early traction suggest the company has found something worth paying for.
What's Under the Hood
The platform breaks down into three products, each addressing a different slice of the customer service workflow.
The voice agents field inbound calls—routing them, deciphering intent, executing tasks. No hold music, no "press one for billing." The chat agents work the same logic across text channels, troubleshooting and completing workflows with what the company says is performance comparable to strong human agents. Then there's the Copilot, a real-time assistant that sits beside actual customer service reps and surfaces policy details, suggested responses, and next steps in multiple languages.
It's designed to plug into existing enterprise software stacks. Voices and personalities are configurable. Analytics cover the usual suspects: intent recognition, resolution rates, containment metrics, sentiment tracking. The goal, predictably, is lower costs, faster response times, higher satisfaction scores.
The Enterprise Angle
Where Cignara tries to separate itself—at least in its messaging—is around privacy and reliability, two areas where large companies have hesitated to embrace AI agents wholesale.
The company says it doesn't route personally identifiable information directly to underlying language models. SOC 2 compliance is part of the sales pitch. To address hallucination risks (the AI equivalent of making things up), Cignara anchors responses in what it calls an "enterprise knowledge graph," built from approved internal data sources: product catalogs, service documentation, operating procedures, call scripts.
The platform ingests information from disparate systems and structures it to support complex workflows. Importantly, it's designed to work with multiple LLM providers rather than tying customers to a single foundation model—a hedge against vendor lock-in and shifting performance benchmarks.
Templates cover standard use cases: account access, incident management, appointment scheduling, complaint handling, retention calls, booking inquiries. Policy guardrails are baked in. Reporting and audit trails come standard, the kind of table stakes features procurement teams expect.
A Very Crowded Field
The problem for Cignara—or perhaps the opportunity, depending on how you view markets in flux—is that it's hardly alone.
Sierra, the high-profile startup co-founded by Bret Taylor and Clay Bavor, hit a $10 billion valuation in 2025 and announced a Japan expansion backed by SoftBank Vision Fund 2. PolyAI pulled in an $86 million Series D last December and rolled out its Agent Studio product. Observe.AI introduced VoiceAI Agents in March. LivePerson, Replicant, Uniphore—all occupy different segments with established customer bases and multi-year head starts.
A CB Insights report from March, based on buyer interviews, flagged reliability, integration headaches, and lack of meaningful differentiation as recurring pain points when enterprises evaluate AI agent vendors. Four people competing against well-capitalized teams is a tough position, even with Y Combinator's imprimatur.
Still, Gupta seems to be following the classic playbook: find revenue early, prove the concept works, then build. The company is currently hiring a founding frontend engineer for a remote role in India, offering equity between 0.20% and 1.00% and compensation in the ₹4 million to ₹6 million range.
From Bujo to Cignara, Quietly
The rebrand itself was handled without fanfare. Both domain names remain live, but the company's website now reflects Cignara across the board. No press release, no announcement beyond Gupta's earlier LinkedIn post about Bujo's launch, which positioned the startup as among the first publicly visible companies from Y Combinator's latest batch.
It's the kind of low-key pivot that happens frequently in early-stage startups—a name change prompted by trademark issues, branding considerations, or simply a better story to tell customers. What matters more, perhaps, is whether the underlying business can scale beyond a handful of early enterprise deals.
The contact center industry is enormous, fragmented, and ripe for automation. But it's also filled with vendors promising transformation and buyers who've seen enough demos to know the gap between pitch and production. Cignara's edge, if it has one, will likely come down to execution: how well the product actually works when the phones start ringing.
