Big companies and fast-growing firms in India and Asia-Pacific are heading into 2026 betting that autonomous AI agents will move from pilots into core operations, while warning that identity controls, open infrastructure and governance will decide who scales safely and who stumbles, executives across software, cybersecurity and cloud infrastructure told TechObserver.in.
Agentic AI is already shifting from novelty to workload engine, with firms automating routine business tasks and beginning to rewire how software is built, they said.
Vivek Ganesh, regional vice president, OutSystems India, said the discussion was moving ‘from experimentation to execution’ as organisations look for measurable impact rather than ‘proofs of concept’, adding that AI is evolving from general-purpose tools into specialised, industry-specific systems and cost-efficient hybrid architectures combining large and small language models.
For software developers, the shift is expected to be deeper than faster code writing. Rajeev Ranjan, Atlassian’s chief technology officer, said AI would become embedded across an AI-native software development lifecycle, from planning and design to code review, production and incident response.
Every developer to have an AI agent
‘Soon, every developer will have an AI agent embedded at every stage of their workflow,’ he said. Ranjan pointed to early gains at Atlassian, saying the company’s AI tool Rovo helped engineers resolve an incident in 14 minutes, an example of the ‘intelligent problem-resolution’ that he expects to become commonplace.
But wider adoption is colliding with the reality that the same systems that boost productivity also amplify risk, the executives said. Ganesh said that as AI systems become more autonomous and interact with real-time data, ‘governance, security, compliance and auditability become just as important as innovation’, with boards increasingly focused on responsible deployment rather than speed alone.
Identity security is emerging as a central control plane as AI agents proliferate, according to Rohan Vaidya, area vice president for India and SAARC, CyberArk. ‘Identity will become the main control point,’ he said, describing it as the only reliable ‘kill switch’ if an AI agent behaves unpredictably or is compromised.
Vaidya warned that ‘a single leaked API key or a malicious prompt will be enough to cause system-wide problems’ as organisations plug autonomous agents into business workflows.
Vaidya also pointed to an operational headache that could become a recurring source of disruption. Starting in March 2026, the maximum validity period for digital certificates will be cut from 398 days to 200 days, he said, adding that many Indian businesses still manage certificate lifecycles manually. Without automation, that could translate into outages caused by expired certificates and unmanaged machine identities, he said.
Even as companies race to deploy agents, some executives cautioned that the economics of AI could widen the gap between the largest technology groups and everyone else.
Vaidya said limited access to advanced chips, rising cloud costs and data centre constraints would reinforce the dominance of big tech, making large-scale AI ‘a privilege, not a standard capability’ for many Indian organisations.
Infrastructure choices are also coming under scrutiny as enterprises try to balance speed, cost and sovereignty. Peter Lees, head of solution architecture in Asia-Pacific, SUSE, said 2026 was ‘the year to act’, warning that dependence on single-vendor, proprietary technology could pose existential risks.
He said the push for digital sovereignty was pushing enterprises to take greater control of data and platforms, arguing that ‘choosing flexible, open options’ could help avoid ‘enormous migration costs’ and ‘arbitrary price hikes’.
Lees said edge computing would accelerate as factories, retail sites and remote healthcare generate more data outside central data centres, making containerisation a key strategy to manage sprawling deployments securely.
He also argued that perimeter-based security was becoming untenable and that enterprises needed to adopt a zero-trust posture — ‘never trust, always verify’ — and focus less on eliminating every flaw and more on preventing the exploitation of weaknesses at execution time.
Across these themes, executives converged on a common view: the next phase of AI adoption will reward companies that can operationalise autonomous systems while tightening control of identities, data and platforms.
‘Developers won’t just build software; they’ll orchestrate intelligent systems,’ Ganesh said, while Ranjan argued that the payoff would be ‘more room for creativity and complex problem-solving’ as routine work is delegated to embedded agents.
