Written by: James Vincent, Head of Client Solutions, John Santias, Senior Site Reliability Engineer, and David Sharpe, Head of Managed Services, Mantel
Key takeaways for business leaders:
- In traditional IT operations and managed services, cost and capacity grow in step with the size and complexity of your estate. Agent co-workers break that link.
- A specialist team of agents handles the operational volume, including incident triage, cost control and compliance monitoring. Expert engineers supervise them and focus on higher-value work.
- Agents earn autonomy step by step on a Trust Ladder. Each agent gets more scope as it proves itself on each workload.
- Agents run inside your own cloud tenancy and existing tools such as Jira and ServiceNow. Every decision leaves a full audit trail, and your data is never used to train models.
- The code, documentation and architecture stay yours, so you don’t become dependent on Mantel.
IT operations and managed services have traditionally faced a structural ceiling: cost and capacity scale linearly with the size and complexity of your estate. But modern technology is allowing us to decouple operational scale from human effort. At Mantel, we’re helping clients lift their productivity with agent co-workers.
Our vision
Our approach to Modern Managed Services uses an AI-led operating model in which a specialist team of agents handles operational volume, supervised by expert human engineers. Agent co-workers take on incident triage, cost control, and compliance monitoring, so engineers can focus on higher-value work. This inversion of the traditional model creates significant productivity gains, ensuring that operational cost tracks business value rather than estate complexity.
These agents do not operate with unchecked freedom. Autonomy is earned, not assumed. Agents progress up a “Trust Ladder”, gradually gaining autonomy over time as they earn trust with their human operators. This allows human operators to become more consultative and tailor operational systems to better suit business objectives.
Looking further ahead, the potential of this approach applies to more than traditional IT operations. The underlying technology stack powering our agent co-workers is versatile. On new builds, agent co-workers can reduce delivery risk and improve observability during critical implementation phases. In the long term, we envisage deploying agentic assistants across a broad array of business processes, such as fulfilment, customer service, and complex systems analysis, using the same technical foundation to provide targeted benefits in business quality and throughput.
The impact
We’ve taken a proactive approach to ensuring that our agentic solutions move beyond theoretical promise and deliver proven results. Integrating agent co-workers into live, highly regulated environments is already delivering measurable, compounding value.
In our early deployments, the statistical evidence and benefits are profound:
What sets Mantel apart
Hyperscalers and traditional ITSM providers are releasing a wave of generalised AI tools. Ours is a more targeted approach, focusing on role-level agents, tuned specifically to each client workload. We work independently across clouds, with native support for AWS, GCP, and Azure, and our agents extend into data platforms such as Snowflake and Databricks.
Key differentiators of Mantel’s agent co-workers:
- Tenancy-native: Our agents deploy directly inside your own cloud tenancy, ensuring data remains within your managed environment and is not shared for model training purposes.
- Leverages existing knowledge: Agents tap directly into your existing knowledge bases, historical data, and operational playbooks to make contextually accurate decisions.
- Deep ITSM integration: Agents integrate seamlessly into your existing service management tools (such as Jira and ServiceNow) as well as your observability tools.
- Tunable autonomy: Agents can be tuned with differing levels of independence on a per-workload basis.
- Client capability growth: Code, documentation and architecture remain your assets; there is no new dependency created on Mantel
The result is a unified workflow where agents operate inside your existing tools, not around them, leaving a full audit trail of every decision and piece of evidence considered.
What’s next
Our roadmap for AI Operations focuses on expanding each agent’s scope in measured, evidence-based steps. As agents earn trust on each workload, we give them more scope and autonomy.
”Operational AI runs on the same proven technology and governance as your customer-facing AI. That means your IT operations team can move quickly and respond as business needs change.
James VincentHead of Client Solutions, Mantel
Getting started
When organisations scale agent trust in a deliberate way, they can safely use agent co-workers to run their operations more efficiently. The same technology stack can also take on higher-value business processes, so organisations can extend their agents’ scope as business needs change. If you’re ready to get your first agent co-worker up and running, talk to us about where to start.
