Kyndryl has unveiled a new service model aimed at helping large companies shift artificial intelligence from experimentation into day-to-day IT operations, betting that enterprises now want more than pilots and proofs of concept. The offering, called Agentic Service Management, was launched on 2 April and is being pitched as a framework for assessing readiness, setting governance rules and deploying autonomous workflows with human oversight across complex IT estates.
The New York-based technology services group said the model combines a maturity assessment, implementation blueprints and a governance structure designed to support “agentic” systems, in which AI software can take decisions and perform tasks with limited human intervention. Kyndryl is offering the service through Kyndryl Consult, while linking it to its existing Kyndryl Bridge platform, which already provides AI-driven operational insights to customers managing hybrid and multi-cloud environments.
Kyndryl’s pitch addresses a clear market problem. Companies have poured money into AI over the past two years, but many have struggled to convert trials into measurable operating gains. Deloitte said in its 2026 enterprise AI research that organisations are moving from pilot programmes towards production, yet only a minority have transferred a large share of their pilots into live use. Gartner has also warned that at least half of generative AI projects had been abandoned after proof of concept by the end of 2025 because of weak data quality, insufficient controls, rising costs or unclear business value. McKinsey, meanwhile, has said bigger companies are leading the move to scale, but governance and operating-model discipline remain central to capturing value.
That wider backdrop helps explain why Kyndryl is emphasising governance as much as automation. Its maturity assessment reviews existing service management, AI governance, security and operations against standards and frameworks including ISO 42001, then produces a phased roadmap for deployment. Kyndryl is also pairing the service with what it calls Agentic AI Digital Trust, a security-first framework aimed at regulated sectors where data classification, compliance and risk controls are especially sensitive. The company says autonomous capability must sit inside guardrails and remain subject to human supervision rather than be turned loose as a stand-alone efficiency tool.
This framing reflects a broader shift in enterprise technology. Early enthusiasm around generative AI focused on chatbots, coding assistants and productivity gains. The next phase is centred on workflows, orchestration and digital agents that can carry out repeatable tasks across service desks, infrastructure operations and incident response. Deloitte has described agentic AI as one of the areas executives expect to have high impact, while also stressing that success depends on redesigning work, governance and accountability around human-machine collaboration.
For Kyndryl, the launch also fits its own corporate strategy. Since its separation from IBM, the company has tried to improve margins and reposition itself away from legacy infrastructure management alone and towards higher-value consulting, cloud, security and AI services. Its third-quarter fiscal 2026 results showed Kyndryl Consult revenue rising 24 per cent year on year, with the business generating $3.6 billion over the previous 12 months. Company filings also show management highlighting agentic AI capabilities, Kyndryl Bridge and innovation-led delivery as part of its growth plan.
There are reasons for caution alongside the opportunity. Agentic systems promise faster resolution of IT issues, better use of operational data and less dependence on manual intervention, but they also raise harder questions about accountability, access rights, auditability and error management. Even advocates of agentic AI have warned that adoption is outrunning governance in many organisations. Boards and senior management teams are under pressure to prove they understand the risks as well as the productivity gains, especially where AI agents may interact with sensitive systems or regulated data.
Kyndryl’s challenge will be to show that its framework produces concrete business outcomes rather than another layer of advisory language around AI transformation. The company says several capabilities are already being applied in its own service delivery operations and exposed to customers through Kyndryl Bridge, which now serves more than 1,400 customers and generates millions of AI-driven insights each month. That installed base may give Kyndryl a practical advantage as enterprises look for suppliers able to connect AI ambition with operational discipline.
The New York-based technology services group said the model combines a maturity assessment, implementation blueprints and a governance structure designed to support “agentic” systems, in which AI software can take decisions and perform tasks with limited human intervention. Kyndryl is offering the service through Kyndryl Consult, while linking it to its existing Kyndryl Bridge platform, which already provides AI-driven operational insights to customers managing hybrid and multi-cloud environments.
Kyndryl’s pitch addresses a clear market problem. Companies have poured money into AI over the past two years, but many have struggled to convert trials into measurable operating gains. Deloitte said in its 2026 enterprise AI research that organisations are moving from pilot programmes towards production, yet only a minority have transferred a large share of their pilots into live use. Gartner has also warned that at least half of generative AI projects had been abandoned after proof of concept by the end of 2025 because of weak data quality, insufficient controls, rising costs or unclear business value. McKinsey, meanwhile, has said bigger companies are leading the move to scale, but governance and operating-model discipline remain central to capturing value.
That wider backdrop helps explain why Kyndryl is emphasising governance as much as automation. Its maturity assessment reviews existing service management, AI governance, security and operations against standards and frameworks including ISO 42001, then produces a phased roadmap for deployment. Kyndryl is also pairing the service with what it calls Agentic AI Digital Trust, a security-first framework aimed at regulated sectors where data classification, compliance and risk controls are especially sensitive. The company says autonomous capability must sit inside guardrails and remain subject to human supervision rather than be turned loose as a stand-alone efficiency tool.
This framing reflects a broader shift in enterprise technology. Early enthusiasm around generative AI focused on chatbots, coding assistants and productivity gains. The next phase is centred on workflows, orchestration and digital agents that can carry out repeatable tasks across service desks, infrastructure operations and incident response. Deloitte has described agentic AI as one of the areas executives expect to have high impact, while also stressing that success depends on redesigning work, governance and accountability around human-machine collaboration.
For Kyndryl, the launch also fits its own corporate strategy. Since its separation from IBM, the company has tried to improve margins and reposition itself away from legacy infrastructure management alone and towards higher-value consulting, cloud, security and AI services. Its third-quarter fiscal 2026 results showed Kyndryl Consult revenue rising 24 per cent year on year, with the business generating $3.6 billion over the previous 12 months. Company filings also show management highlighting agentic AI capabilities, Kyndryl Bridge and innovation-led delivery as part of its growth plan.
There are reasons for caution alongside the opportunity. Agentic systems promise faster resolution of IT issues, better use of operational data and less dependence on manual intervention, but they also raise harder questions about accountability, access rights, auditability and error management. Even advocates of agentic AI have warned that adoption is outrunning governance in many organisations. Boards and senior management teams are under pressure to prove they understand the risks as well as the productivity gains, especially where AI agents may interact with sensitive systems or regulated data.
Kyndryl’s challenge will be to show that its framework produces concrete business outcomes rather than another layer of advisory language around AI transformation. The company says several capabilities are already being applied in its own service delivery operations and exposed to customers through Kyndryl Bridge, which now serves more than 1,400 customers and generates millions of AI-driven insights each month. That installed base may give Kyndryl a practical advantage as enterprises look for suppliers able to connect AI ambition with operational discipline.
Topics
Technology