Skip to main content
TipRanks
Go Ultimate75% OFF

MGT Scales AI-Driven Software Development With Agent-Based Workflows

MGT Scales AI-Driven Software Development With Agent-Based Workflows

According to a recent LinkedIn post from MGT, the company is developing an AI-based software “factory” built around multiple collaborating coding agents. The post portrays a shift from individual AI assistance toward automated teams that manage development tasks, reviews, testing and deployment.

The account reports that weekly output increased from about 32 points during the spring and summer to 685 points last week, representing 389 pull requests. It also describes 50 to 100 workflows running concurrently, with each workflow tracking cost, quality and speed before completing a retrospective.

The post highlights infrastructure designed to address bottlenecks in agent-based development, including on-demand worktrees, persistent memory, threaded task management and cloud-based processing. It also describes an AI engineering manager called Canary that scores pull requests for business value and helps coordinate review, correction and merging.

For investors, the reported workflow could point to higher engineering productivity and greater software output without a proportional increase in conventional headcount. However, the figures are self-reported and measure internal development activity rather than revenue, customer adoption or profitability, so the near-term financial impact remains uncertain.

The post suggests MGT is positioning its engineering model within a broader industry shift toward always-on AI agents and automated software operations. It also references recruiting engineers to build similar systems, which could support future product development while increasing competition for specialized talent and raising questions about reliability, oversight and operating costs.

Disclaimer & DisclosureReport an Issue

Related Articles