Kusho AI creates artificial intelligence agents for software testing before failures occur

Kusho AI, a Bengaluru-based company founded in 2023 by Abhishek Saikie and Saurabh Gavande, develops an AI-powered platform that automates much of this work. Its AI agents generate, run, and continuously update tests for mobile applications, web applications, and APIs, helping engineering teams spend less time on software maintenance and more on creating new products.

The platform uses proprietary models layered onto foundational large language models (LLMs), functioning as an AI software engineer. It tests code both during development and after every release to detect issues before they reach the production environment.

The idea originated from Abhishek Saikie's experience working as a product manager at Flipkart. He holds degrees in economics and chemical engineering from BITS Pilani. He later teamed up with his classmate Saurabh Gavande, who is now the CTO and previously worked as a full-stack engineer at FalconX.

Consider an example of a food delivery application launching a new recommendation engine. This feature must work correctly for users with five previous orders as well as those who have placed thousands of orders, across different cities and usage scenarios. Manually testing every possible case requires designing test scenarios, writing code, running tests, and analyzing results—a process that can take hours or even days, yet still leaves edge cases unnoticed.

Kusho AI automates this process. First, the platform learns the company's software environment before simulating the approach of a human tester. It identifies critical scenarios, generates the necessary tests, executes them automatically, and flags issues requiring attention.

According to Saikie, this process runs 'almost 100 times more efficiently and much faster' than manual testing. The platform supports engineering teams throughout the entire software release cycle.

Before code merge, Kusho AI automatically generates and runs tests for early risk detection. During release, it acts as an additional check, confirming the readiness of the new code for production. Saikie calls this one of the company's main advantages for enterprise clients. He notes: 'Even if your developer is rushing and wants to quickly push the product to production, Kusho AI stands there almost like a watchman.'

If the release is not ready, the platform returns it with a report detailing what needs fixing. After deployment, Kusho AI continues monitoring for anomalies and changes that lead to breakdowns. As applications evolve, the platform automatically updates its tests to reflect these changes, which, according to Saikie, creates 'self-healing systems that identify changes and incorporate them accordingly,' thereby reducing the need for repeated manual updates.

Every release also receives a risk score based on changes in the base code and executed tests. In addition to automated testing, Kusho AI positions itself as an enterprise testing dashboard, offering release control, audit trails, domain-aware automation, and integration with existing software release workflows. The goal is to provide engineering leaders visibility into the entire testing process, not just individual test runs.

Instead of relying on a single AI model, Kusho AI operates through what Saikie calls an 'agent orchestration layer.' Different testing requests are directed to specialized models depending on the application type. For instance, a banking API is handled differently than an e-commerce platform.

In Saikie's view, the company's competitive advantage lies precisely in this orchestration layer and the proprietary data used to fine-tune it. He emphasizes: 'This is proprietary training that can only be gained once when thousands of people use your product.'

The startup is also selective about the clients it serves. Saikie says: 'Our entire ideal client profile is people who cannot afford even one minute of downtime,' referring to sectors such as banking and ride-hailing, where even brief outages can immediately impact the business.

Among over 5200 clients, Kusho AI lists Ultrahuman, Paytm, Ola Electric, HCL, Michelin, HP, Razorpay, Roche, Schneider Electric, and Paymentus. Kusho AI has largely followed a product-led growth strategy. The platform is available for free use without requiring payment details, allowing it to spread among developer and QA communities through technical content rather than paid marketing. Free users can consent to the platform being used for learning based on how they use it. Kusho AI tracks the tests users save or discard, using this anonymized data to improve future recommendations.

As adoption grows within engineering teams, conversations with large enterprises often follow. Currently, the platform has generated and executed nearly 10 million tests and serves about 35,000 developers and QA specialists on its free tier. According to the company, one Indian payment platform reduced API testing time by 95%, and a private bank lowered annual QA costs by 80–90%. User feedback also notes a reduction in manual UI testing due to continuous test generation.

Kusho AI operates in the growing market of AI-powered software testing and reliability alongside companies such as Postman, Qodex.ai, and Devzery. Saikie reports that the startup's Annual Recurring Revenue (ARR) is in the 'high six figures' (in dollars), growing by 10–20% monthly.

The company raised $600,000 in a seed round led by Antler India, with participation from Blume Founders Fund, UpSparks Capital, and angel investors, including Ultrahuman co-founders Mohit Kumar and Vatsala Singhal, as well as IDfy CEO Ashok Harikharan. Saikie stated that the company plans to conduct another round later this year or early next year.

Kusho AI is registered in the US, has Indian founders, and its official headquarters are in San Francisco. Their seven-member team currently operates from Bengaluru, with a physical office planned in San Francisco by late 2026 or early 2027 as their US client base expands, which already accounts for 60–70% of their customers. Over the next year, the startup intends to deepen its presence in banking and financial services, logistics, supply chains, and medical technology, targeting the top 10 companies in each sector. For Saikie, the long-term vision is simple: to make software maintenance so automated that engineers can focus entirely on creating new products.