The trouble with agentic AI services is that they become popular.
When an agentic (or indeed a predictive or generative) artificial intelligence service lands the right way around in terms of users’ expectations, it needs to solidified and moved out of any sandboxed prototyping zone, it needs to be galvanised and hardened for compliant production deployment, it needs to be managed for all sorts of considerations from observability to configuration to interconnectity… and it ultimately needs to be scaled for its wider usage if the market so dictates.
That’s a lot of platform capabilities in one mouthful.
Aiming to provide a fully governed end-to-end platform for operationalising agentic AI systems in this regard is Domino Data Lab. The company says it has now delivered new platform capabilities that unite experimentation, evaluation, deployment and monitoring in one governed workflow.
Domino’s latest release claims to equip organisations for what the industry is now calling the new agentic development lifecycle (ADLC) experience with the underlying LLM hosting capabilities needed.
Tracking, evaluation & monitoring
Nick Elprin, co-founder and CEO of Domino suggests that many agentic AI application teams lack the tracking, evaluation and monitoring capabilities commonplace in traditional machine learning workflows. He says that this creates challenges when moving agentic AI systems from prototype to production… it also erodes enterprise trust in these applications to execute real business workflows and automate complex, high-impact decisions.
“Building and deploying agents in production requires both rapid experimentation and robust governance,” said Elprin. “Domino’s Winter Release gives enterprises the agility and control they need to deliver agentic systems that drive real business impact. Teams building agentic AI in Domino can visualize, evaluate, and compare systems at both summary and trace-level detail using shared metrics and complete configuration lineage.”
Domino’s latest moves see it make sure the same integrated and governed platform teams rely on for traditional AI now supports the full agentic AI lifecycle. This foundation now supports customers building, deploying, and monitoring agentic applications with the integration and governance they expect.
Agentic instrumentation & evaluation
Through its new platform expansion, Domino says it delivers new dedicated agentic instrumentation and evaluation capabilities that connect all stages of the agentic AI lifecycle (i.e. build, evaluate, deploy and monitor) within a shared system of record and with the ability to iterate at scale across each stage. This is intended to ensure complete lineage, reproducibility and governance.
A built-in universal tracing software development kit also exists. Using any agentic orchestration framework, teams can trace every step of agentic AI creation—including prompts, tool calls, decisions, and output through each ADLC stage.
“Structured evaluation and side-by-side comparison also features here. Teams building agentic AI systems can visualise, evaluate and compare applications at both summary and trace-level detail using shared metrics and complete configuration lineage supporting consistent, repeatable evaluation,” confirmed Domino, in a product statement.
Metrics, custom evaluations & feedback
Teams can evaluate production performance of agents using metrics, custom evaluations and human feedback, re-visiting historical agent decisions and exploring detailed traces captured in production.
“Fragmented tools and ad-hoc processes are critical obstacles keeping agentic AI stuck in prototype,” said Shawn Rogers, CEO of analyst firm BARC US. “Enterprises need a single governed lifecycle and a unified platform that connects experimentation, evaluation, deployment, and monitoring of agents at scale. This approach gives teams the ability to iterate rapidly and move agents to production with confidence.”
Underpinning the ADLC experience, teams can now securely host, serve and manage LLMs in their own infrastructure for high-performance inference and reduced operational costs.
In summary, Domino Data Lab details its capabilities and says that they allow organisations to adopt LLMs at scale while maintaining control over data, costs and security within established regulatory boundaries.
