Insider’s memory
- Deductive AI launched publicly with $7.5 million in seed funding to provide AI SRE agents that automatically detect faults, diagnose root causes in seconds, and guide corrective actions on complex software systems.
- Its platform connects to code, logs, metrics, traces, and events, using a continuously updated knowledge graph and agentic reasoning to replicate the way expert engineers investigate failures, reducing debugging time by up to 90%.
- Backed by CRV, Databricks Ventures, Thomvest and PrimeSet, Deductive AI is already deployed at companies like DoorDash, Foursquare and Kumo to accelerate incident resolution and improve software reliability at scale.
PRESS RELEASE — Deductive AI, a pioneer in AI-powered root cause analysis and resolution, announced $7.5 million in seed funding and public launch. The round was led by CRV, with participation from Databricks Ventures, Thomvest Ventures and PrimeSet. With its official launch, Deductive AI introduces a new generation of AI SRE agents that learn from real-world incidents to automatically detect faults, diagnose root causes, and help engineers resolve software issues in minutes instead of hours.
At its core, Deductive AI is a unified platform for deploying and training AI agents that connect to an organization’s code, logs, metrics, traces, and events. These agents can reason about a continually updated knowledge graph that maps relationships between systems, allowing them to test hypotheses, evaluate evidence, and identify precise root causes in seconds.
“The complexity and interdependencies of modern infrastructure mean that investigating the root cause of an outage or incident can be like looking for a needle in a haystack, except the haystack is the size of a football field, it’s made up of a million other needles, it’s constantly reshuffling and it’s on fire – and every second you don’t find it equals lost revenue,” said Sameer Agarwal, co-founder and CTO of Deductive AI. “Our goal is to fundamentally change the problem of how engineers troubleshoot software in production. By teaching AI agents to think like engineers, we give teams time to build and innovate rather than constantly fighting.”
Engineering teams today face the enormous challenge of diagnosing software failures across fragmented systems, sifting through petabytes of logs, millions of time series metrics, and hundreds of dashboards to find the needle in a haystack. These investigations can take hours or even days, distracting engineers from building new features and shipping products, while outages ripple through systems long after the initial outage.
“We’ve seen world-class engineers spend half their time debugging instead of building,” said Rakesh Kothari, co-founder and CEO of Deductive AI. “And as vibe coding generates new code at a rate we’ve never seen, this problem will only get worse. With new, unpredictable failures, teams that don’t evolve their approach to debugging will get stuck in maintenance instead of shipping.”
Powered by purpose-built agentic AI and reinforcement learning, Deductive mimics the workflow of experienced site reliability engineers (SREs). It generates hypotheses, tests them against actual system evidence, and provides clear, actionable insights at every stage of an investigation. While traditional observability tools stop at surfacing problems, Deductive AI goes a step further by not only identifying root causes, but also providing fixes and guiding teams through corrective actions.
“Our advertising platform operates at a pace where slow, manual investigations are no longer viable. Every minute of downtime directly affects business revenue. In these highly stressful and ambiguous early moments of an incident, AI-driven triage plays a crucial role in accelerating our path to mitigation, supporting our 2026 goal of a 10-minute resolution window. Deductive has become a critical extension of our team, quickly synthesizing signals across dozens of services and making bring out the information that matters – in minutes »
– Shahrooz Ansari, Senior Director of Engineering at DoorDash
“As the industry rushed to coding agents, our partnership with Deductive took a different path: integrating agentic AI directly into data operations. Together, we built a system that unified our telemetry feeds with Deductive’s reasoning engine, creating an AI-powered frontline agent that proactively manages our data pipelines – detecting anomalies, diagnosing root causes, and accelerating resolution.”
– Vikram Gundeti, Chief Technology Officer, Foursquare
“Deductive promises to democratize visibility into our training infrastructure which today has a steep learning curve. We are excited to partner with Deductive and believe it can automatically connect the dots between hundreds of workflow executions and tell us the root cause of any training workflow failure.”
– Virajith Jalaparti, Infrastructure Manager, Kumo
“We’ve been using Deductive for about a year now. It gives us the clarity and speed we need to turn telemetry and code data into meaningful actions. Deductive has demonstrated the ability to learn and adapt to our unique use case and deliver key insights at every stage of the investigation so our team can move forward faster, smarter and with confidence.”
– Pratik Kotkar, VP of Engineering, Apoha
“Deductive AI is a revolutionary company tackling some of the most persistent problems in software reliability. Their approach represents the next generation of intelligent reliability systems that help engineering teams move faster, innovate more, and manage complexity at scale.”
– Max Gazor, CRV Board Partner and early investor in category-defining companies such as Reflection AI, Airtable and Cribl
Deductive’s founders bring deep expertise in building and scaling some of Silicon Valley’s most successful data and analytics platforms. Sameer Agarwal received his Ph.D. at UC Berkeley, where he created BlinkDB, a cutting-edge approximate query processing framework, and was among the first engineers at Databricks to create Apache Spark. Rakesh Kothari was an early engineer at ThoughtSpot, where he led teams focused on distributed query processing and large-scale systems optimization. Together, they founded Deductive AI to put reasoning, not reaction, at the heart of software reliability.
Additional Information
For more information, visit deductive.ai where you can also book a demo.
About Deductive AI
Deductive AI creates AI SRE agents that connect codebases and telemetry data to guide incident response from detection to resolution. Founded by veterans of Databricks and ThoughtSpot, Deductive AI combines code-based reasoning, agentic AI, and reinforcement learning to accelerate root cause analysis by up to 90%. The company is based in Mountain View, California, and is backed by CRV, Databricks Ventures, Thomvest Ventures and PrimeSet. Learn more about www.deductive.ai or connect with Deductive on LinkedIn Or X (Twitter).
Media contact
Elliott Burr
[email protected]
Featured Image: Credit: Deductive AI
