Sage has announced a joint initiative with PwC aimed at making AI in finance transparent and auditable, backed by IDC research showing that more than 70% of finance leaders would reject AI outputs they could not explain — even when those outputs are accurate.
The announcement came at Sage Future in San Francisco on 30 April. The partnership frames its approach around what Sage calls "glass box" AI — systems designed to show their reasoning rather than deliver opaque recommendations — combined with PwC's experience deploying AI at scale internally.
The trust gap the initiative targets is significant. IDC's data, commissioned by Sage, shows that a majority of finance leaders place explainability above accuracy as a precondition for adoption. In regulated industries, where decisions need to be defensible to auditors and regulators, that preference is commercially rational rather than simply cautious.
Alongside the PwC collaboration, Sage also announced a broader expansion of AI agents across its core product suite. Sage Intacct, its cloud accounting platform for mid-market businesses, gains agents for finance workflow automation. HCM products gain workforce management agents, and Sage X3 gets operational intelligence capabilities. The company characterises the expansion as directly embedding AI into existing workflows rather than offering it as a separate layer.
The pairing of explainability-first design with agent automation addresses a common tension in enterprise AI deployments: automation that users cannot understand tends to generate compliance risk and end-user resistance. Whether the glass-box framing holds up to scrutiny — and how Sage and PwC define "explainable" in practice for complex multi-step agent decisions — will determine whether the initiative advances beyond positioning.
