Cekura is an innovative technology startup focused on enhancing the reliability and performance of Conversational AI agents, specifically voice and chat AI. Founded in 2024 by a team of IIT Bombay alumni with advanced research backgrounds, Cekura operates out of San Francisco and has quickly established itself as a key player in the quality assurance space for AI-driven conversational interfaces. As enterprises increasingly deploy AI assistants in sectors like healthcare, finance, and retail, ensuring these systems operate flawlessly becomes critical. Cekura’s platform addresses this need by automating the testing and monitoring of AI voice and chat agents, helping companies reduce error rates, accelerate deployment, and improve user experience. This article unveils ten fascinating insights into Cekura’s origins, technology, team, market impact, and future outlook.
Cekura was founded in 2024 by three close-knit friends and IIT Bombay graduates—Sidhant Kabra, Tarush Agarwal, and Shashij Gupta. Combining their expertise in quantitative finance, NLP research, and consulting, they identified a common pain point: the difficulty of reliably testing conversational AI agents. Frustrated by slow, manual QA that missed critical edge cases, they set out to build an automated system capable of simulating thousands of conversation scenarios. Their diverse backgrounds, including research experience at ETH Zurich and Google, empowered the team to innovate confidently in the growing voice AI space, culminating in Cekura's launch and participation in Y Combinator's Fall 2024 batch.
Conversational AI agents are prone to unexpected failures due to subtle changes in prompts, user interruptions, and complex edge-case interactions. Traditional quality assurance methods—manual testing, call reviews, and reactive monitoring—are too slow and error-prone to keep pace with rapid AI development cycles. This unreliability is especially critical for high-stakes industries like healthcare or financial services where errors can lead to compliance risks or customer distrust. Cekura solves this by offering a proactive testing and observability platform that automatically generates diverse test scenarios, simulates production calls, catches failures early, and continuously monitors real production conversations.
At the heart of Cekura’s platform is an AI-powered scenario generation engine that creates thousands of diverse conversational test cases from agent descriptions. It supports custom personas with varied accents, background noise, and conversational styles to replicate real-world users. Cekura measures agent performance using customized metrics—focusing on instruction adherence, tool calls, latency, interruptions, and sentiment analysis. The platform also integrates with Continuous Integration/Continuous Deployment (CI/CD) pipelines to ensure smooth development cycles. Its observability features provide deep analytics such as call volume trends, drop-off tracking, and alerting on anomalies, enabling teams to swiftly identify and fix issues.
Since its launch, Cekura has garnered more than 75 customers spanning multiple industries including healthcare, BFSI (Banking, Financial Services, and Insurance), retail, logistics, and recruitment. Its clients range from fast-growing startups to large enterprises seeking to deploy reliable conversational AI at scale. Notable customers include AI mortgage servicing companies like Kastle and AI receptionists like Sandra for automotive dealerships. The platform’s ability to reduce testing times from weeks to minutes has been a key driver for adoption, along with the critical need for compliance and high reliability in regulated sectors.
In mid-2025, Cekura successfully closed a $2.4 million seed funding round, led by renowned accelerator Y Combinator and supported by Flex Capital, Hike Ventures, Pioneer Fund, and prominent angel investors such as Kulveer Taggar, Ooshma Garg, and Austen Allred. This funding has enabled the company to expand its engineering team, accelerate product development, and scale its enterprise customer base. It also signaled strong investor confidence in Cekura’s vision to build the “reliability layer” for conversational AI, underscoring the urgency and market potential for trustworthy AI voice and chat agents.
Cekura has formed key partnerships, including with companies like Retell AI, a conversational AI platform offering voice and chatbots powered by leading large language models (LLMs) such as GPT and Claude. This integration allows Retell’s customers to seamlessly connect their agents with Cekura’s testing and observability tools, enabling automated simulation, transcript analysis, and real-time production call monitoring. Such partnerships expand Cekura’s ecosystem reach, enabling it to serve a broader audience while offering unified tools for AI agent development, testing, and deployment.
The founders share a passion for leveraging AI to solve complex real-world problems. Tarush Agarwal brings expertise from ultra low-latency quantitative finance systems; Shashij Gupta has conducted influential NLP research at Google and ETH Zurich; and Sidhant Kabra comes from a strategic consulting and scaling background with extensive experience in customer experience management. Their combined vision is to make AI conversational agents as dependable as human employees, thereby transforming how businesses interact with customers. Their journey from identifying testing pain points to building a market-leading QA platform epitomizes their dedication.
Unlike typical testing tools that focus on generic metrics, Cekura deeply understands conversational context and mission-critical requirements. Its automated scenario generation covers thousands of edge cases not feasible through manual tests. The product supports evaluation against user-defined success criteria, compliance checks, and even simulates aggressive or interruptive users to test agent robustness. Furthermore, Cekura offers extensive analytics for production monitoring with proactive alerting, allowing enterprises to catch failures before users do. This comprehensive approach positions Cekura as a pioneer in the conversational AI reliability space.
Headquartered in San Francisco, Cekura plans to expand its footprint internationally, notably opening an office in Bangalore, India, tapping into its founders' home market and developer talent pool. The company already serves Indian clients including Prodigal, Nurix, and Skit. With its global and diverse team, Cekura is poised to capture a significant share of the booming voice AI market, which is becoming ubiquitous in customer service, healthcare, and other sectors worldwide.
Looking ahead, Cekura aims to refine its testing engine with more advanced scenario generation and enhanced coverage metrics. As conversational AI adoption accelerates, maintaining security, reliability, and compliance will be paramount. Cekura’s vision is to make quality assurance for AI agents a seamless, scalable process, allowing organizations to confidently deploy voice and chat applications at scale. Its growth and investor backing highlight an industry shift towards operational excellence in AI, positioning Cekura to play a crucial role in shaping the future of trustworthy, resilient conversational AI systems.
Cekura represents a critical innovation in the Conversational AI landscape by addressing one of its most urgent challenges—reliability. Founded by passionate IIT Bombay alumni with deep technical and domain expertise, Cekura offers a comprehensive platform that transforms manual, slow QA processes into rapid, automated testing and monitoring workflows. Its swift market traction, robust funding, and strategic partnerships underscore growing demand for dependable AI voice and chat agents, especially in regulated industries. As conversational AI becomes integral to customer engagement, Cekura’s tools are poised to become indispensable, ensuring AI interactions are secure, accurate, and user-friendly. The question remains: as AI agents become the frontline of business communication, how will Cekura’s innovations continue to safeguard trust and quality in this evolving space?