Customer Story
How an AI-native observability company built autonomous investigation agents on Akka — root cause analysis up to 90% faster, time-to-mitigation down 70%.
DeductiveAI built a multi-agent system that autonomously investigates production incidents across logs, metrics, and code changes, ranks probable causes, and hands engineers the evidence. Root cause analysis runs up to 90% faster and time-to-mitigation falls 70%, across knowledge graphs spanning millions of nodes and billions of time series.
Akka ROI Scorecard
faster root cause analysis. Investigation agents correlate telemetry and code changes autonomously, turning multi-hour manual triage into minutes.
lower cost to operate the investigation pipeline. A 10-person team runs enterprise-scale RCA on infrastructure that stays responsive under peak event volume.
of time series and petabytes of logs analyzed in real time, over knowledge graphs of millions of nodes and edges built from live telemetry.
Business outcome: time-to-mitigation reduced 70% on average, with actionable evidence delivered even when the exact cause is not found.
DeductiveAI serves engineering teams that run large, multi-region distributed systems. When those systems fail, root cause analysis means correlating billions of time series, petabytes of logs, and hundreds of millions of lines of code changes across sources like GitHub, DataDog, Prometheus, and New Relic. Done by hand, that investigation stretches across hours while an incident stays live.
Automating it demanded a system that could ingest millions of events in real time, hold the full state of an ongoing investigation, and coordinate many AI agents working a single incident at once — without losing progress when a step failed. Conventional request-response infrastructure could not carry that stateful, long-running, concurrent workload at production scale.
DeductiveAI built its multi-agent system on Akka, delivered and operated through Akka. Agents ingest telemetry into a knowledge graph, detect probable causes, rank hypotheses, and return evidence to developers in real time — running fully autonomously or in a human-in-the-loop mode. The platform carries the state of each investigation durably, so a long-running analysis survives failure and resumes where it left off.
Agents connect to GitHub, Prometheus, Datadog, and New Relic, ingesting telemetry at minute-level granularity tuned to each source's rate limits. Akka coordinates event-based agent workflows, with built-in guardrails and checkpoints that prevent runaway token usage or infinite analysis loops. The result is elastic data ingestion feeding a live entity-relationship graph, resilient concurrency that keeps telemetry pipelines stable under pressure, and clean, maintainable ingest code that a 10-person team can operate at enterprise scale.
This is the Never Fail guarantee: Akka handles clustering, resilience, durable state, and traffic steering, so the agents investigating an incident do not lose their work and stay responsive as event volume spikes.
Scale. The system ingests millions of events in real time and builds knowledge graphs spanning millions of nodes and edges from billions of time series, petabytes of logs, and hundreds of millions of lines of code changes. Investigations run across that graph without losing state.
Real-time healthcare. Health professionals can instantly and securely discuss patients' cases and protocols with colleagues and peers through rich media messaging or voice and video calling without interrupting time-sensitive workflows.
Security and compliance. This solution maintains the highest of global standards and complies with the GDPR, E-Privacy, the NHS information governance, DCB 0129, ICO, ISO 27001, and NEN 7510, 7512, and 7513.
Cost to operate. A 10-person team delivers and operates enterprise-scale RCA. Durable state and automatic recovery remove the operational overhead of keeping long-running, concurrent investigations alive under load.
Speed to production. Root cause analysis is accelerated up to 90%. Agents that once required hours of manual correlation now return ranked probable causes in minutes.
Business impact. Time-to-mitigation fell 70% on average. As DeductiveAI reports, even when an investigation does not surface the exact root cause, the evidence the agents gather is a substantial time-saver for engineers.
DeductiveAI is an AI-native company: autonomous agents that investigate incidents across telemetry are the product, not a roadmap item. That system reached production through Akka Specify, the delivered-outcome path — specifications go in, and a governed production system comes out in weeks, fully operated. Instead of assembling stateful orchestration, durable investigation state, and recovery on their own, the team defined the behavior they needed and received a running system built to carry millions of events and long-lived, concurrent agent workflows.
The same platform serves both paths. Enterprises build agentic AI directly on the Akka Agentic AI Platform, or have a governed system delivered and operated through Akka Specify. Both provide the production guarantees DeductiveAI runs its autonomous investigations on today: durable state, resilience under peak load, and recovery that survives failure.
