Thoughtworks unveils AI/works for legacy modernisation
Thoughtworks has launched AI/works, an agentic development platform aimed at legacy system modernisation and new software development.
The consultancy said AI/works combines legacy system analysis, requirements work, automated specification generation, and agent-led code generation and testing in a single platform. Thoughtworks positioned the product for complex enterprise environments where organisations run a mix of older and newer systems.
Legacy focus
Thoughtworks said many organisations have increased their interest in AI while continuing to rely on long-established applications. It said this gap has shaped the design of AI/works. The company also framed the market as one where many tools focus on new code creation rather than the renewal of existing platforms.
"Every CEO and CIO I meet is trying to unlock AI value inside the reality of their existing systems, not in idealised greenfield environments," said Mike Sutcliff, CEO, Thoughtworks. "AI/works is built for those conditions. It understands the systems organisations have, accelerates the systems they need next and keeps everything current as the landscape shifts. The magic comes from the combination of the platform and our deeply talented technologists. Together they deliver results with speed and confidence."
The company said the platform uses AI-enabled reverse engineering to interpret legacy applications. Thoughtworks said the process converts existing systems into structured specifications. It said it enriches those specifications with regulatory, security and industry context.
Thoughtworks said the specifications then guide agentic workflows across engineering tasks. It said the platform generates code, automated tests, and deployment pipelines.
Continuous updates
Thoughtworks said AI/works supports ongoing change after deployment. It said the platform regenerates affected components when requirements change. Thoughtworks said this approach reduces reliance on manual patching. It also said it avoids large rebuild projects.
In its announcement, Thoughtworks linked AI/works to its 3-3-3 delivery model. The company described that model as a route from idea to production in 90 days.
Thoughtworks also said AI/works integrates with cloud and data platforms. It named AWS, Google Cloud, Microsoft Azure, Databricks and Snowflake as ecosystems it works with.
The company said it has a collaboration with Mechanical Orchard. Thoughtworks said the work supports mainframe renewal.
Early deployments
Thoughtworks said early clients using AI/works have shortened modernisation cycles. It said work that previously took years has moved to a timeframe of months. The company said those projects included cost reductions and time-to-market improvements. It also said the work generated higher quality code.
Agentic engineering has become a fast-moving area of software development. Thoughtworks said several major firms have entered the category. It said many offerings focus on improving productivity for new code writing.
"AI/works stands out because it addresses the entire lifecycle, from understanding and renewing legacy systems to building what comes next," said Wang. "This sets a new bar for the category."
Thoughtworks said AI/works is available through a co-innovation programme. It said broader availability will expand through its AI that works initiative and Q1 launch activities.