Incident management · Production

Factory.ai: Enterprise A-SWE Droid Platform for the Full Software Development Lifecycle

The problem

Enterprises running large legacy codebases are underserved by AI coding tools built for solo developers or greenfield projects, and existing IDE-based tools impose latency and cost constraints that prevent true delegation of software tasks to AI agents.

First attempt

Early agentic approaches like baby AGI and auto-GPT were seen as unreliable endless-loop systems that took many actions without guidance, making the 'agent' label inadequate for describing a more structured, goal-oriented system.

Workflow diagram · grounded in source
1
User submits ticket
trigger
“I'm going to paste a ticket into our platform”
2
Droid semantic search
ai_action
“the droid starts working. It's doing a semantic search on part of my query in that code base”
3
Initial plan presentation
ai_action
“It gives an initial crack at a plan, right? Presents that really clearly to you”
4
Clarification and user input
human_review
“The system knows when you give a very detailed answer or request to follow your instructions, when you give more ambiguous requests to ask for clarification”
5
Agentic task execution
ai_action
“the system has access to a bunch of different tools. Here, memory, project management tools, GitHub, web search”
Reported outcome

Factory.ai reached general availability with an enterprise code agent platform backed by Sequoia and Lux Capital, with enterprise customers including MongoDB's CEO publicly endorsing the platform.

Reported metrics
Reliability droid leverage for incident responsesuper high leverage
Reported stack
Factory BridgeLinearJIRASlackGitHubPagerDuty
Source
https://www.latent.space/p/factory
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Factory.ai reached general availability with an enterprise code agent platform backed by Sequoia and Lux Capital, with enterprise customers including MongoDB's CEO publicly endorsing the platform.

What tools did this team use?

Factory Bridge, Linear, JIRA, Slack, GitHub, PagerDuty.

What results were reported?

Reliability droid leverage for incident response: super high leverage (source-reported, not independently verified).

What failed first in this deployment?

Early agentic approaches like baby AGI and auto-GPT were seen as unreliable endless-loop systems that took many actions without guidance, making the 'agent' label inadequate for describing a more structured, goal-orie…

How is this incident management AI workflow structured?

User submits ticket → Droid semantic search → Initial plan presentation → Clarification and user input → Agentic task execution.

Related incident management cases