Birlasoft Limited says one-third of code uses agentic AI

Birlasoft Limited said on October 28, 2026, that around one-third of its code was generated using agentic artificial intelligence tools, up from almost 25% a few months earlier. Birlasoft linked the increase to its AI-driven delivery life cycle, which has been applied to more than 20 leading service offerings with human review controls.

How much Birlasoft code is generated by agentic AI?

Birlasoft said around one-third of its code is now generated with agentic AI tools, compared with almost 25% a few months before the Q2 FY27 earnings call. The company presented the figure as an operational measure of AI use in its engineering delivery process, rather than as a disclosure of AI revenue, headcount or total code volume.

The move from almost one-quarter to around one-third represents a broader use of AI-assisted engineering in Birlasoft's delivery model. Management did not disclose the number of developers using agentic tools, the volume of code generated, or the share of client engagements included in the measure. It did say that more than 20 top service offerings have had their engineering workflows reimagined with an AI-first approach.

Birlasoft described its AI-driven delivery life cycle, or AI-DLC, as combining architectural work, AI-assisted engineering, automated testing, accelerated migration and disciplined programme execution. This places the code-generation disclosure within a wider delivery process that includes testing and platform modernisation. The company also said it is equipping employees to work with agentic AI through AI-DLC.

Which platforms support Birlasoft's AI-led delivery model?

Birlasoft said Cogito, Lynx and Sigma are the three platforms supporting its AI-led offerings. Cogito is its enterprise agentic AI platform, Lynx is its AI-powered test-engineering platform, and Sigma covers agentic AI operations and application management services, or AMS.

The three platforms address different stages of technology delivery. Cogito supports enterprise AI deployment, Lynx covers testing, and Sigma is focused on operational and managed-services work. Birlasoft did not disclose revenue, client numbers or booking values for any individual platform, but linked all three to its AI-DLC model across 20-plus service offerings.

Birlasoft said human-in-the-loop controls remain in place for security, governance, quality and reliability as it expands agentic AI usage. That means management is not describing the tools as a replacement for engineering review. The Q2 FY27 transcript did not provide figures for defect rates, productivity gains, security incidents or post-review changes to AI-generated code.

What AI credentials and client wins did Birlasoft report?

Birlasoft said it earned Amazon Web Services, or AWS, AI Services Competency status during Q2 FY27 and became an OpenAI select partner. Management said AWS awarded the competency after reviewing live programmes involving enterprise agentic AI, legacy-application modernisation, intelligent document processing and governed analytics.

Management said the AWS credential recognises Birlasoft's ability to deliver generative AI and agentic AI solutions at production scale on AWS. It said the status was already helping it win clients and business, although it did not attach a contract value to that assertion. Birlasoft specifically said its new multiyear AI engagement with a global logistics company was secured on the strength of the competency.

The logistics engagement involves large-scale platform modernisation, cloud adoption and business transformation for a European organisation. Birlasoft said it will use architectural capability, AI-assisted engineering, automated testing and accelerated migration to modernise mission-critical platforms. The company did not disclose the engagement's value, implementation timetable or the proportion of work expected to use agentic AI-generated code.

Birlasoft also reported a three-year AI-led managed-services engagement with a global industrial manufacturer to support enterprise resource planning, or ERP, systems and business innovation. Separately, it was selected by a North American medical-devices manufacturer for programme governance, an AI and data centre of excellence, and core digital transformation. These wins span modernisation, managed services and transformation rather than a standalone code-generation product.

How does AI delivery fit Birlasoft's Q2 FY27 performance?

Birlasoft reported Q2 FY27 revenue of Rs 1,416.2 crore, up 2.7% sequentially in rupee terms and 2.1% in constant-currency and dollar terms. Management said Manufacturing, banking, financial services and insurance, or BFSI, and Life Sciences & Services drove the quarter's growth, while Energy & Utilities, or E&U, remained soft after a European client's work ramp-down.

The growth pattern differed across verticals and service lines. Manufacturing grew 5.2% sequentially, BFSI rose 3.6% and Life Sciences & Services increased 2.1%, while E&U was affected by the client reduction. Digital & Data grew 3.4% and ERP rose 2.9% sequentially, whereas Infrastructure was affected by the E&U revenue decline. Management said both Life Sciences and ERP recorded a second consecutive quarter of sequential growth.

Birlasoft reported earnings before interest, tax, depreciation and amortisation, or EBITDA, of Rs 215.3 crore and an EBITDA margin of 15.2% in Q2 FY27. Profit after tax was Rs 159.6 crore and the effective tax rate was 29%. Management attributed the margin result to operational efficiencies, productivity improvements and revenue quality, partly offsetting the organisation-wide compensation revision effective from July 2026.

The company said Q2 deal signings were US$11.61 crore, up 9% year on year, while first-half signings were US$28.5 crore, up 15% year on year. Birlasoft did not disclose how much of the signing value came from AI-led work. The transcript therefore does not establish a numerical link between the one-third code measure and quarterly revenue or bookings.

What does Birlasoft expect from AI delivery in H2 FY27?

Birlasoft said demand conditions remained unchanged and Q3 FY27 would reflect furloughs, fewer working days and holiday-season effects. Management nevertheless said it would continue investing in technology capabilities and its sales organisation, citing its pipeline, ongoing customer conversations and a seasonal uptick in renewals.

Management said it was confident that H2 FY27 financial performance would be better than H1 FY27, while stressing that it does not provide quarterly guidance. Its stated focus for the second half is sales execution and building a stronger order book. The company did not provide an AI-revenue target, an AI-productivity target or a forecast for the share of code generated by agentic tools.

On profitability, management said its endeavour is to continue operating at around a 15% EBITDA margin, while noting that seasonal effects could cause smaller movements. It cited the July 2026 salary increase and continuing investment in sales and service-line capabilities. Birlasoft also said reinvestment in technology infrastructure and office refurbishment contributed to higher depreciation and amortisation.

Conclusion

Birlasoft's Q2 FY27 disclosure shows that agentic AI is embedded in its engineering process, with usage moving from almost 25% of code to around one-third within a few months. The company is combining that delivery metric with AI-DLC workflows, the Cogito, Lynx and Sigma platforms, AWS competency status and AI-led engagements across logistics, industrial manufacturing and medical devices.

The next disclosed milestones are the expected Q3 FY27 furlough impact and management's aim for H2 FY27 financial performance to exceed H1 FY27. Investors can also watch whether the sales pipeline and seasonal renewals convert into further AI-led bookings, while Birlasoft continues technology and sales investment and seeks to operate around a 15% EBITDA margin.