San Francisco-based Autoheal has raised $7.9 million in a seed funding round to scale its self-improving software factory, designed to help enterprise platform engineering teams build, deploy, govern and continuously improve AI agents across the software development lifecycle.
Trusted by industry leaders including Nomura Bank and AvidXchange, Autoheal helps platform engineering teams build and manage AI agents for post-coding workflows such as coding cost efficiency, incident response and vulnerability remediation. These workflows are grounded in enterprises’ own systems and processes.
AI is enabling engineering teams to ship more code at a faster pace, but the acceleration is also creating a growing operational burden, including production incidents, security vulnerabilities and rising token costs.
Autoheal is designed to address these challenges and is already being used by enterprises where off-the-shelf point agents have struggled to deliver.
The $7.9 million seed round was led by Innovation Endeavors, with Harpinder Singh joining Autoheal’s board. The round also saw participation from Emergent Ventures, U&I Ventures, Darkmode Ventures, Batch Ventures and Param Hansa Values.
Why Platform Engineering Needs a New Operating Model
Repetitive software development lifecycle (SDLC) workflows, including incident response and vulnerability remediation, consume more than a third of an engineering team’s capacity.
As coding agent adoption increases, managing large language model (LLM) spending and context is becoming another challenge for engineering organizations.
To manage these demands, platform engineering teams are increasingly shifting towards a “software factory” model powered by specialized AI agents.
However, deploying these agents at scale can be challenging because of fragmented tools, a lack of shared context and strict security requirements.
A unified platform for creating, managing and continuously improving software factory agents, including existing coding agents, has therefore become a priority.
Such a platform provides agents with shared engineering context, secure production access, private evaluation infrastructure, cost controls and mechanisms to remain current as enterprise environments evolve.
“Our experience taught us that while building the first version of an AI agent is easy, scaling it consistently across the enterprise SDLC is the real challenge,” said Sid Choudhury, Co-Founder and CEO of Autoheal.
“Platform engineers need more than cloud agents that execute tasks. They need a unified platform to deploy, govern, and continuously improve those agents across complex enterprise workflows. That’s why we built Autoheal.”
Also Read: Ultraviolette Raises $85 Million Led by Yali Capital, TDK Ventures
What Autoheal Is Building
Autoheal’s software factory provides enterprises with infrastructure and tools designed to transform engineering organizations into a self-improving system.
The platform connects existing coding agents, code repositories, CI/CD systems, observability platforms, cloud runtimes and issue trackers. This gives worker agents across the software factory a shared engineering context graph.
As these worker agents execute repetitive post-coding workflows, two self-improvement agents operate in the background: the Evaluator and the Healer.
- Evaluator Agent
The Evaluator agent scores every worker agent’s run. For example, a coding agent can be evaluated based on the specifications and pull requests it generates, with downstream signals such as review comments, CI failures and incidents used as evaluation criteria.
- Healer Agent
The Healer agent addresses low-scoring worker agents by opening pull requests that improve their skills, prompts, tools or model selections.
The agent also verifies these changes against historical benchmarks to check for regressions before they are reviewed by engineers.
For platform engineering teams, this creates a continuous agent-healing loop as enterprise systems and processes change. Every behaviour change is version-controlled in Git and requires engineer approval.
Actions remain governed and audited, with visibility into access, reasoning and costs. The goal is to improve accuracy and execution speed while reducing the cost per successful task, with engineers able to expand agent autonomy as the systems demonstrate reliability.
Autoheal Traction
Autoheal is already operating inside complex regulated environments, where engineering teams are using the platform to reduce incident response times, manage customer support escalations and free up thousands of hours of engineering capacity.
“Our production operations teams spend valuable time triaging alerts and managing incidents, while also pulling engineers away from their software development activities.
Autoheal gives us a platform that takes investigation timelines down from hours to minutes. The fact that it runs entirely within our own cloud, in compliance with our controls, made it a natural fit for how we operate,” said Sameer Jain, CIO, Wholesale at Nomura Bank.
“In production incident response, Autoheal took our time to root cause to minutes, with evidence our engineers trust. That’s time our developers stay focused on feature work.
Next, we’re shifting it left into other critical parts of our SDLC, because every engineering hour we get back goes into shipping faster for our customers.” said Krish Shetty, CTO & SVP, at AvidXchange.
“Autoheal helped us tackle two major challenges at once: making our engineers faster at troubleshooting across our complex environment, and significantly optimizing our software costs across our monitoring stack.” said Vijay Pendyala, SVP Engineering & Customer Success, at Empiric Earth.
Autoheal Origin Story
Autoheal grew out of its founders’ experience building enterprise engineering and AI platforms at Harness, Microsoft Azure, ThoughtSpot and AppDynamics.
After scaling Harness to more than $200 million in annual recurring revenue (ARR), the team identified a new challenge.
While building individual AI agents had become easier, deploying those agents safely across the SDLC and engineering teams had become increasingly time- and token-intensive.
To prevent agent sprawl and enable day-two governance, the company believes a platform needs to manage agents as code while overseeing them through continuously learning meta-agents. This insight became the foundation for Autoheal’s software factory.
“Enterprises are moving quickly from experimenting with AI agents to asking how they can operate them safely and efficiently at scale across the entire software factory,” said Harpinder Singh of Innovation Endeavors.
“Autoheal is building the agent infrastructure layer that makes that possible. The opportunity is much larger than one agent or one workflow. It is giving platform teams a repeatable scalable way to deploy specialized intelligence across the engineering organization.”
What’s Next for Autoheal
Engineering processes and implicit architecture decisions are often contained within an enterprise’s boundaries or remain within engineers’ knowledge.
Frontier models have been trained on public internet data, open-source code and synthetic data, but not on enterprises’ proprietary data.
Enterprises are looking to build sovereign and cost-effective intelligence around this data, which Autoheal identifies as a competitive advantage.
The company plans to first capture this data by operating the software factory and then train small private models with different architectures for individual customers. These models are expected to power a majority of the tasks in the software factory that are not generative in nature.
Over the longer term, the same architecture could extend beyond software engineering into data and security engineering.
Autoheal is building toward a model in which large enterprises operate software factories with their own populations of specialized agents, with the company aiming to provide the platform engineering teams use to build, govern and continuously improve those agents.
