How realfast.ai is helping IT services industry cut delivery times
realfast.ai aims to enhance IT service delivery by developing hyper-scalable AI assistants designed to streamline software development, customisation, and maintenance processes.
Artificial Intelligence (AI) is driving transformation across sectors. From healthcare and finance to manufacturing and retail, AI is streamlining processes and enhancing efficiency.
Realising the growing challenges faced by the Indian IT services industry, Sidu Ponnappa, Aakash Dharmadhikari, and Steve Sule started realfast.ai in 2024.
The Singapore- and Bengaluru-based startup leverages AI to accelerate execution, potentially doubling revenue and halving delivery times for IT companies.
The 11-member team has developed AI agents by integrating AI solutions into software development and maintenance processes to enhance the entire lifecycle of enterprise software development.
AI agents speed up execution, which leads to faster client feedback. By reducing the manpower needed per unit of software, gross margins can increase to 60%, says Ponnappa.
“There is an impending industrial revolution for IT services. Building software is a craft. The industrialisation of this has not been possible because that would mean industrialising reasoning—the ability to think, judge, and make decisions… and the latest developments in AI have enabled just that,” Ponnappa tells YourStory.
Before starting realfast.ai, Ponnappa and Dharmadhikari co-founded C42 Engineering, a boutique software product engineering firm, in 2010. It was acquired by Gojek in September 2015.
Sule, a former partner at Singapore-based venture capital firm Edith Grove Capital, initially joined the company as an advisor before becoming a co-founder this year.
Role of AI agents
With recent advancements in AI, the team believes AI can revolutionise IT services by augmenting human capabilities across the entire service lifecycle.
realfast.ai focuses on deploying AI agents to support human workers across tasks like requirement gathering, design, development, testing, and deployment. This approach speeds up delivery times while maintaining, or even improving, code quality and service standards.
The company’s core offering, Vayu, integrates with the Salesforce platform, offering AI agents that can access developer tools and assist humans in their tasks.
Vayu's USP lies in helping Salesforce implementation teams work faster, while providing better service and support to its clients.
A Salesforce implementation team is a group of experts that works together to set up and customise Salesforce for a business.
Salesforce partners can implement these solutions at double the speed while achieving higher SaaS profit margins through the use of AI technology.
“We are already seeing early success with our design partners. Our unit-testing agent has helped them reduce the time it takes for writing unit tests by 50%—reducing a typical 10-day sprint cycle to just 8 days,” Ponnappa says.
Vayu, which is built on top of multiple AI models such as ChatGPT and Anthropic’s Claude, has improved software systems by enhancing traditional and outdated code, making them more efficient.
Vayu currently handles over 700-unit testing tasks by mimicking how developers complete them.
“All of our AI agents undergo the same automated assessments as human developers when assisting with tasks. Just like a human’s work is reviewed by a leader, the AI’s output is also reviewed and approved by a senior leader in the team. We never allow AI assistance without human oversight,” he says.
As the company expands its AI agents to manage more tasks, it believes that the combined benefits in speed and quality will improve project delivery times while reducing the risks and volatility typical of large IT service projects.
The founder adds that a major white space for AI innovation lies in the IT industry, as every role, from sales to QA, deals with text-based tasks—spanning from contracts, requirements, code, logs, and tickets. Despite being a prime AI use case, this area is often ignored due to its complexities, requiring deep expertise in services.
“We’re building an AI platform which allows AI agents to operate on top of deep, complex enterprise stacks and deliver real work and productivity. And the amazing thing about services is productivity translates to money. This is why this combination of taking an AI attack on services is something that is really gaining traction in the market,” says Ponnappa.
Funding and plans ahead
India's artificial intelligence sector is expected to reach $17 billion by 2027, with an annual growth rate of 25-35% between 2024 and 2027, according to a joint report from and consulting firm BCG (Boston Consulting Group).
Alltius, Ema, and Revrag.ai are some of the other players in the market focusing on creating intelligent agents that enhance enterprise productivity.
After being in stealth for the past year, realfast.ai has secured seed funding from PeakXV, RTP Global, and DeVC Dugout.
The startup aims to expand its Vayu model into new areas of IT services, enterprise resource planning (ERP), IT service management (ITSM), and human capital management (HCM) implementations, along with manual QA. This move is part of the company's strategy to boost efficiency across key service domains.
The platform is currently in early pilot programs with mid-sized IT companies in India.
Currently, the IT services sector’s gross margins are around 35%, EBITDA is between 15-30%, and net margins range from 7-15%. realfast.ai aims to bring the gross margins to ~60% and cut delivery time by half.
With several new solutions arising in the market, Ponnappa believes the future is in hybrid human-AI teams, where AI supports specialised tasks rather than functioning as general chatbots.
“One of the key trends that we foresee in the industry with AI adoption is that human beings will now spend a lot more time reviewing the work than generating. In the past, it was effectively human beings sort of doing a lot of this work and also then reviewing it, which is the senior people doing this. Now everybody will have an AI assistant,” he explains.
He also emphasises that raw LLMs trained on academic benchmarks lack the consistency required for enterprise work, stressing the need for task-specific agents that deliver better results.
In terms of challenges, Ponnappa notes that the main challenges come from concerns about IP (intellectual property) and privacy, though there has been progress in addressing these issues.
“We like to think of it as building an Iron Man suit—it changes everything, but the individual inside ultimately shapes the outcome. The goal is to give everyone in the industry superhuman capabilities, while ensuring that the final result reflects their unique skills and approach,” Ponnappa says.
Edited by Megha Reddy
