Stockland builds Sage assistant to surface SAP data

Stockland has built an AI assistant called Sage that connects about 700 property developers and project managers to procurement and purchasing data held in SAP, ITnews reports. Sebastian Gray, Stockland's head of technology, said at AWS Summit Sydney that Sage lets users "log in, start talking in natural language, and Sage then speaks 'SAP' [to SAP] and gets the actual data back". Gray told the Summit that typical SAP lookups take about five to six minutes, and that the assistant was developed and launched in production in 18 weeks after the business case was approved, ITnews reports.

Per comments reported by ITnews, Sage runs on Bedrock and is surfaced inside Microsoft Teams, so employees with existing SAP and Teams credentials can access it without additional sign-on. Gray described common user flows-purchase order and project budget queries-as examples where the assistant translates natural-language requests into the SAP transactions needed to retrieve results.

Embedding conversational agents in collaboration platforms is a common enterprise pattern to lower friction for infrequent users. Companies adopting this pattern typically aim to reduce time-to-data for nontechnical staff and to avoid retraining people on complex ERP navigation, which can unlock measurable productivity gains across distributed teams.

This implementation is a practical example of combining a managed foundation-model service with an established ERP and a collaboration layer to shorten operational workflows. For practitioners, it demonstrates an integration stack that relies on Bedrock for model hosting, SAP for authoritative data, and Teams for user access.

Observe whether Stockland or similarly structured organisations publish metrics on accuracy, audit trails, or access controls for SAP queries. Industry observers will also look for how organisations balance conversational convenience with governance, role-based access, and auditability when surfacing ERP data via AI assistants.