CAST launches conversational AI tool for software portfolio analysis
CAST has launched 'Ask CAST', a conversational AI capability for its Highlight platform that allows IT leaders to query their software portfolios in natural language and receive what the company describes as deterministic, analysis-backed responses. The tool is available immediately to existing CAST Highlight clients.
The feature is accessible through Microsoft Teams and via third-party AI assistants including Claude, ChatGPT and Google Gemini, as well as any tool supporting the Model Context Protocol (MCP). CAST says no additional training is required; users interrogate their portfolio data within the tools they already use day-to-day.
What it does
CAST positions 'Ask CAST' as a means of collapsing the lag between a business question and an actionable answer across four enterprise IT priorities: technical debt reduction, application modernisation, cloud migration readiness, and AI-deployment planning. The company says the assistant can identify which applications are candidates for consolidation or retirement, sequence remediation work by business risk, and flag which applications are best suited to agentic augmentation.
Greg Rivera, VP of Product at CAST, said the capability was designed to remove friction from decision-making. "Whether it's a strategic priority or an urgent threat, an executive decision or a complex project, 'Ask CAST' removes the friction that historically slows everyone down," he said.
CAST did not disclose the number of Highlight clients eligible for the feature, benchmark response-time figures, or specific customer case studies in the release.
Market context
The software intelligence and portfolio management market has grown sharply as enterprise codebases age and AI-transformation programmes raise the stakes for getting modernisation sequencing wrong. CAST competes in a space that also includes platforms from ServiceNow (IT Business Management), LeanIX (acquired by SAP), and a range of application portfolio management tools offered by Broadcom and IBM. The addition of a conversational interface follows a pattern visible across enterprise software: vendors are wrapping existing analytical engines in natural-language layers to reduce the specialist skills required to extract value.
The MCP integration is notable. Anthropic's Model Context Protocol has attracted rapid adoption as a standard for connecting AI assistants to enterprise data sources, and CAST's decision to support it alongside its own Microsoft Teams integration suggests the company is betting on protocol-level interoperability rather than a single-vendor AI stack. For enterprise architects evaluating the tool, the security model governing which portfolio data the AI assistant can access, and under what conditions, will be a key due-diligence question that the release does not address.
Regulatory read-across
As organisations accelerate AI deployment on top of existing software estates, regulators are paying closer attention to the provenance and auditability of the data informing AI-driven decisions. The EU AI Act's transparency obligations for high-risk AI systems, and broader expectations under NIS2 around software supply-chain risk management, create commercial incentives for vendors that can demonstrate their intelligence layer is grounded in auditable, code-level analysis rather than probabilistic inference. CAST's emphasis on "100% deterministic data" appears calibrated to address exactly this concern, though the company has not published an independent audit of that claim.
CAST was founded in 1990 and describes itself as the pioneer of the software intelligence field. The company serves enterprise clients, consultancies and cloud providers across multiple continents. The next milestone to watch is whether the company publishes usage metrics or named customer outcomes for 'Ask CAST' as the capability scales beyond its Highlight base.