Factory and Carahsoft Technology Corp. have formed a distribution partnership that will make Factory’s artificial intelligence software-development platform available to federal, state and local agencies through established public-sector contract vehicles.
Carahsoft will serve as Factory’s master government aggregator and distribute the platform through its reseller network. The product is now available through NASA’s Solutions for Enterprise-Wide Procurement V contract, The Interlocal Purchasing System, E&I Cooperative Services and The Quilt.
The agreement is a procurement and distribution development, not the announcement of a completed government deployment. No agency customer, contract value or production outcome was identified.
Factory describes its platform as “agent-native,” meaning AI agents can perform connected software-development tasks rather than simply suggest individual lines of code. The system is designed to turn requirements and operational signals into software by assisting with planning, coding, testing, review, security checks and release activities.
From coding assistant to development workflow
The broader goal is to automate more of the software lifecycle while keeping engineers in control of policies, approval gates and technical standards. Agencies could use that model to modernize legacy applications, accelerate routine maintenance or support DevSecOps teams facing staffing constraints.
Factory says its platform is model-independent and can work with open or proprietary AI models through customer-controlled keys and endpoints. A routing function selects a model based on the task, performance, cost and reliability. That flexibility may appeal to agencies that want to avoid tying a development workflow to a single model provider.
The company offers cloud, on-premises and fully air-gapped deployment options. Security features listed in the announcement include single-tenant environments, role-based access controls, single sign-on, automated user provisioning, customer-managed encryption keys, audit logs and zero data retention. Factory also reports that it has completed a SOC 2 examination.
Those capabilities address some public-sector requirements, but agencies will need to evaluate them against the sensitivity and authorization level of each intended environment. The partnership announcement does not state that the offering has a FedRAMP authorization, nor does it establish that every supported AI model or deployment configuration meets a particular government security baseline.
Governance becomes part of software assurance
Autonomous development tools introduce risks beyond conventional code generation. Agencies need to know what source material a model used, whether generated dependencies are approved, how secrets and sensitive code are protected and whether automated changes can bypass established reviews. They also need evidence that generated software is tested for vulnerabilities, licensing conflicts and operational reliability.
Human approval gates are therefore more than a product feature. They are a core control for preserving accountability when an AI agent proposes or executes changes. Agencies considering the platform should establish which tasks an agent may perform independently, which require review and how every change is logged and reversible.
Procurement access through SEWP V and cooperative contracts could shorten the administrative path for eligible buyers. It does not replace technical evaluation, acquisition planning or authorization to operate. Buyers will still need to define measurable objectives, such as reduced remediation time, faster release cycles or fewer defects, and compare those results with the cost of licenses, integration and oversight.
Factory and Carahsoft have lowered a purchasing barrier for public-sector organizations interested in AI-assisted development. The more important test will be whether agencies can use autonomous workflows to deliver software faster without weakening security, code quality or responsibility for the final product.