--- schema: "swft.publication/v1" id: "software-factory-definition" title: "What is a software factory?" description: "An AI software factory turns clear product requests into checked software through agents, human decisions, guardrails, and learning." summary: "An AI software factory is a repeatable way to turn a clear product request into working software. People choose the goal and quality bar. Coding agents do bounded work inside a prepared system, and automated checks plus human judgment decide what can ship." canonical: "https://swft.io/what-is-a-software-factory" author: "SWFT Editorial" author_type: "Organization" published: "2026-08-31" modified: "2026-09-02" kind: "explainer" section: "Guides" tags: ["what is a software factory", "AI software factory", "agentic software development", "coding agents"] evidence_labels: ["INFERENCE", "OBS", "SELF-REPORT"] source_ids: ["ai-engineer-worlds-fair-2026", "bemer-software-factory-history", "hraness-building-software-factory", "latent-space-factory-loops", "ona-software-factory", "strongdm-software-factory", "uber-efficient-software-factory", "usaf-acquisition-report"] authorship_disclosure: "AI-drafted from the cited public sources and independently checked by a second AI editorial-review agent (Codex) for source fit, claim boundaries, overlap, and reader utility. SWFT Editorial is responsible for corrections." --- # What is a software factory? A plain-English guide to the people, agents, checks, and learning loop behind an AI software factory. > **Authorship:** AI-drafted from the cited public sources and independently checked by a second AI editorial-review agent (Codex) for source fit, claim boundaries, overlap, and reader utility. SWFT Editorial is responsible for corrections. ## Quick answer An AI software factory is a repeatable way to turn a clear product request into working software. People choose the goal and quality bar. Coding agents do bounded work inside a prepared system, and automated checks plus human judgment decide what can ship. Imagine a customer reports that checkout fails when a discount code expires. In an ordinary AI-assisted workflow, someone asks a coding tool for a fix and then shepherds the result through the rest of the process. In a software factory, that whole path is designed: the request becomes clear work, an agent gets the right context and tools, the change is checked, a person handles the decisions that need judgment, and the released result feeds the next improvement. That complete loop is the factory. The coding agent is one worker inside it. ## What does a working AI software factory need? ### 1. A task brief A **task brief** is a specification in product language: the goal, relevant limits, owner, risk, and definition of done. “Fix checkout” is too vague. “Expired discount codes should show an explanation, preserve the cart, and never charge the customer” gives the team and the agent something testable. ### 2. Supplied context **Context engineering** decides which code, product decisions, customer evidence, and rules the agent sees for this job, what it can fetch later, and what would only distract it. The supplied context should be small enough to use and complete enough to prevent avoidable mistakes. ### 3. An agent harness The **agent harness** surrounds the model with instructions, tools, context, memory, permissions, and a safe workspace. A better model can help, but the harness determines whether it can find the right file, run the right check, recover from a failure, and stay inside its authority. ### 4. A proof loop Before asking for approval, the agent should check its own work. We call this the **proof loop**. In AI engineering, the repeatable scenarios and scoring rules are **evals**. Ordinary software tests, security checks, screenshots, and human review can all supply proof. The version that passed must be the version that ships. ### 5. Guardrails and named decisions **Guardrails** answer practical questions: What may the agent change? How much may it spend? Which actions need approval? When should it stop and ask for help? “Human in the loop” is too vague unless the human decision is named. ### 6. A learning loop The **learning loop** turns repeated failures and production feedback into better specifications, supplied context, tools, evals, and rules. It improves the way work gets done, not just the one result that failed. [Latent Space's 2026 field dispatch](https://www.latent.space/p/aiewf-daily-dispatch-loops) describes this wider loop across signals, prioritization, implementation, review, deployment, and learning. The [AI Engineer World's Fair 2026 program](https://ai.engineer/worldsfair/2026) reflects the same shift in the field: software factories, harness engineering, context engineering, evals, sandboxes, memory, and agentic engineering appeared as distinct areas of practice. ## Why call it a factory? The word emphasizes repeatability. One clever agent run is a demo. A factory can handle another valid request through a known path, show what happened, and get better when that path fails. The idea predates generative AI. An [IEEE Computer Society profile](https://history.computer.org/pioneers/pdfs/B/Bemer.pdf) records Robert W. Bemer's early software-factory work, and the [United States Air Force](https://www.af.mil/Portals/1/documents/2021SAF/04_Apr/FY19_FY20_Dept_of_the_Air_Force_Acquisition_Biennial_Report_final.pdf) later used the term for organizations and platforms built around continuous software delivery. Today's AI software factory keeps the focus on the complete production system while giving agents more of the implementation work. For a personal account from someone building one, read Ben Guo's [Building a software factory](https://hraness.com/writing/building-a-software-factory). ## The simplest test A system is becoming a software factory when it can take another clear task brief, produce a checked result through a repeatable path, show the proof, and learn from failure. If success still depends on one person's hidden memory or heroic cleanup, the factory is not ready to scale. ## How we know - **Observed artifact (OBS)** The phrase software factory has documented pre-generative-AI lineage. Sources: [Robert W. Bemer](https://history.computer.org/pioneers/pdfs/B/Bemer.pdf); [Department of the Air Force Acquisition Biennial Report FY19–FY20](https://www.af.mil/Portals/1/documents/2021SAF/04_Apr/FY19_FY20_Dept_of_the_Air_Force_Acquisition_Biennial_Report_final.pdf). - **First-party report (SELF-REPORT)** Current operators use the term for systems spanning agents, context, validation, delivery, and operations. Sources: [Running a Software Factory Efficiently at Uber Scale](https://www.uber.com/us/en/blog/efficient-software-factory/); [The StrongDM Software Factory: Building Software with AI](https://www.strongdm.com/blog/the-strongdm-software-factory-building-software-with-ai); [Software Factory](https://www.software-factory.dev/). - **Analysis (INFERENCE)** SWFT's definition joins the recurring ideas in historical and current usage: repeatable work, clear boundaries, proof, human control, and feedback that improves the system. Sources: [Robert W. Bemer](https://history.computer.org/pioneers/pdfs/B/Bemer.pdf); [Running a Software Factory Efficiently at Uber Scale](https://www.uber.com/us/en/blog/efficient-software-factory/); [The StrongDM Software Factory: Building Software with AI](https://www.strongdm.com/blog/the-strongdm-software-factory-building-software-with-ai); [AIEWF daily dispatch: loops and software factories](https://www.latent.space/p/aiewf-daily-dispatch-loops); [AI Engineer World's Fair 2026](https://ai.engineer/worldsfair/2026). ## Sources - **Observed artifact (OBS)** [AI Engineer World's Fair 2026](https://ai.engineer/worldsfair/2026) — AI Engineer; accessed 2026-09-01. The official 2026 program groups software factories, harness engineering, context engineering, evals, sandboxes, memory, and agentic engineering into dedicated tracks. - **Observed artifact (OBS)** [Robert W. Bemer](https://history.computer.org/pioneers/pdfs/B/Bemer.pdf) — IEEE Computer Society; accessed 2026-08-31. Secondary historical profile recording Bemer's work defining the nature of the software factory. - **Observed artifact (OBS)** [Building a software factory](https://hraness.com/writing/building-a-software-factory) — Ben Guo; accessed 2026-08-31. Ben Guo's earlier field note about the concrete factory he operates. - **Observed artifact (OBS)** [AIEWF daily dispatch: loops and software factories](https://www.latent.space/p/aiewf-daily-dispatch-loops) — Latent Space; published 2026-07-01; accessed 2026-09-01. A dated field dispatch connecting software-factory practice to the complete loop from signals and prioritization through execution, review, deployment, and learning. - **First-party report (SELF-REPORT)** [Software Factory](https://www.software-factory.dev/) — Ona; accessed 2026-08-31. Ona's public experiment and livestream series about the inputs, steering, and operation of software factories. - **First-party report (SELF-REPORT)** [The StrongDM Software Factory: Building Software with AI](https://www.strongdm.com/blog/the-strongdm-software-factory-building-software-with-ai) — StrongDM; accessed 2026-08-31. StrongDM's public account of its software-factory operating model. - **First-party report (SELF-REPORT)** [Running a Software Factory Efficiently at Uber Scale](https://www.uber.com/us/en/blog/efficient-software-factory/) — Uber Engineering; published 2026-08-27; accessed 2026-08-31. Uber's account of factory layers, adoption, token economics, managed agents, quality signals, and context infrastructure. - **First-party report (SELF-REPORT)** [Department of the Air Force Acquisition Biennial Report FY19–FY20](https://www.af.mil/Portals/1/documents/2021SAF/04_Apr/FY19_FY20_Dept_of_the_Air_Force_Acquisition_Biennial_Report_final.pdf) — United States Air Force; accessed 2026-08-31. Official account of Cloud One, Platform One, Kessel Run, and software-delivery modernization. ## Read next - [The rise of software factories](/the-rise-of-software-factories) - [Software factory vs coding agent, copilot, harness, and CI/CD](/software-factory-vs-coding-agents-ci-cd-devops-platform-engineering) - [How an AI software factory works](/software-factory-architecture)