AI observability gains ground in India as enterprises focus on cost control and security: Splunk

Indian enterprises are strengthening their ability to monitor artificial intelligence (AI) systems, with a growing focus on controlling costs, improving security, and ensuring reliable performance as agentic AI adoption expands, according to Cisco’s Splunk State of Observability 2026: Restoring Predictability in the AI Era report.

Based on a survey of 2,050 IT operations and engineering professionals across 10 countries, the report examines how organisations are adapting observability practices to manage AI workloads, assess agent performance and maintain visibility into security risks and operational spending.

Globally, 87% of respondents said monitoring AI was a priority for their organisation, compared with 85% in India. However, measuring AI performance remains a challenge. Worldwide, 56% of respondents reported difficulty measuring metrics such as large language model (LLM) response quality, compared with 46% in India.

Indian organisations also reported progress in identifying AI-generated inaccuracies. Although 87% of Indian respondents considered visibility into AI hallucinations important, 70% said their organisations were effective at detecting them, compared with 58% globally.

The findings underline the importance of AI observability as enterprises deploy autonomous agents across business operations. The report found that 83% of respondents globally and 81% in India were concerned about the unintended consequences of AI agents.

Jeetu Patel, President and Chief Product Officer, Cisco, said managing AI agents at scale requires closer integration between observability and security. He highlighted the importance of shared data platforms that help organisations identify and respond to issues in real time.

Enterprises focus on controlling AI costs

Managing unpredictable AI expenditure is another priority. As organisations deploy multiple agents and models, tracking consumption and attributing costs to specific workloads becomes increasingly important.

The report found that 55% of respondents globally struggled to control AI costs, compared with 43% in India. Unexpected billing spikes were a greater operational concern than baseline costs for 60% of respondents worldwide and 54% in India.

Telemetry expenditure is also emerging as a concern, with 57% of respondents saying it had exceeded the cost of the infrastructure being monitored. Meanwhile, 67% reported using a single observability platform to gain a consolidated view of operational performance and costs.

Kamal Hathi, SVP and General Manager, Splunk at Cisco, said observability can help enterprises make AI spending more predictable by introducing guardrails and linking AI usage and token costs to measurable business outcomes.

Observability strengthens security collaboration

As AI agents gain access to enterprise systems and data, collaboration between IT operations, engineering and security teams is becoming increasingly important.

According to the report, 89% of Indian respondents said monitoring AI workloads had increased collaboration between observability and security teams, compared with 81% globally.

The study also found that 81% of respondents in India could accurately trace performance issues to security-related root causes, against 73% globally. Worldwide, collaboration between observability and security teams on troubleshooting rose to 77% in 2026 from 62% in 2025, while data sharing increased from 74% to 84%.

These findings suggest that integrating operational and security data can help organisations investigate incidents more effectively as AI systems become embedded in business-critical workflows.

Stronger observability linked to business outcomes

The report identified organisations achieving better results through advanced observability practices. These organisations frequently use AI agents to prevent or resolve issues, closely monitor model quality and costs, and connect telemetry insights to business decisions.

Among these leaders, 91% said they could accurately attribute AI costs to specific agents, models and workloads, compared with 69% of other respondents. They were also 2.4 times more likely to report that their observability platforms delivered returns significantly above expectations.

Additionally, 54% of leaders said observability had generated at least $5 million in annual business revenue, compared with 28% of non-leaders. Collaboration with security teams was another differentiator: 94% of leaders reported increased collaboration through AI monitoring, compared with 80% of other respondents.

India reports progress in AI monitoring

Indian organisations demonstrated comparatively strong capabilities in monitoring AI model quality. Some 51% of Indian respondents described their practices as very effective, compared with 35% globally.

The findings point to the need for enterprises to build on these capabilities by improving AI cost attribution, strengthening security controls and linking monitoring practices to business outcomes.

As agentic AI adoption grows, the report highlights observability as an important foundation for helping Indian enterprises scale AI with greater visibility into performance, security and expenditure.