Sage Pushes Finance AI Beyond Demos With 'Glass Box' Model
Sage Future 2026 opened in San Francisco with a thesis for finance leaders: high performance comes from AI embedded directly in ERP workflows, governed transparently, and paired with nonnegotiable human accountability. Across keynotes and sessions, Sage executives, AWS leadership, and the New York Yankees' CFO sketched a model where finance teams use AI agents to accelerate decisions while keeping judgment and control in human hands.
Accuracy Over Hype
Sage CEO Steve Hare compared today's AI rush to the city's gold rush era, arguing that lasting value comes from those who build trusted infrastructure rather than chase hype. In finance, "nearly right is wrong," he said, making accuracy, auditability and explainability mandatory for any AI that touches system of record data or closes the books.
Sage's response is what it calls "glass box AI"-combining a domain-specific accounting and compliance language model with proprietary data and embedding agents inside products like Sage Intacct rather than running AI as an external add-on. Every recommendation is designed to be traceable, with clear logic and adherence to existing permissions so controllers and CFOs can interrogate outputs instead of trusting opaque models.
Hare stressed that finance judgment and accountability cannot be outsourced. "There's huge potential in AI," he said, "but there's also huge accountability and responsibility. If anything, it becomes more important. They're the ones building trusted systems capable of operating in the real world."
For finance executives, that combination changes daily work patterns. Transaction-level tasks such as reconciliations, exception checks and policy enforcement are increasingly handled by agents, while leadership attention shifts to scenario evaluation, policy tuning and cross-functional alignment.
Customers Show Results
Sage highlighted customers already using embedded AI to improve performance. At Byler Holdings, finance leader Rebecca Miller reported that AI-infused workflows in Sage systems changed core processes, improved accuracy and freed more than 100 hours per month that had been spent on manual checks and adjustments. That time is now redirected toward analysis, planning and business partnering.
Hare invoked the LPGA as a metaphor for modern finance, noting golfers can analyze every swing in detail and learn more about their game than ever before. "The data is extraordinary," he said. "However, it doesn't take a shot and definitely doesn't sign the scorecard. Tools can get better, but the performance still belongs to the leader who uses them."
AWS Partnership Signals Scale
An expanded collaboration with AWS formed another pillar of the event. Sage is working with AWS AI services such as Amazon Bedrock and purpose-built chips like Trainium and Inferentia to scale agentic AI across its portfolio, including Sage Copilot and finance agents.
Julia White, CFO at AWS, described agentic AI as one of the fastest-growing categories on AWS and emphasized the goal of embedding AI seamlessly into processes rather than adding complexity for users. For CIOs and enterprise architects, that partnership signals tighter integration between Sage workloads and AWS infrastructure, including shared governance, logging and performance optimization for AI-heavy finance applications.
ERP platforms that pair domain-specific AI with hyperscale infrastructure will be better positioned to support large-scale, regulated finance operations.
Yankees Show the Operating Model
Scott Krug, senior vice president and CFO of the New York Yankees, brought a practical lens to the keynote by charting the club's finance evolution. In 2004, there was effectively no FP&A and little historical information in systems, which limited the organization's ability to evaluate major initiatives such as a new stadium.
"Most significant transactions end up on the back page of the newspaper, on ESPN and more," Krug said. "Our actions are reported to the public."
Over two decades, the Yankees shifted to a forward-looking finance model with budgets expanding from a single page to dozens of pages that now mirror the detail available to Fortune 500 executives. Adopting Sage Intacct in 2023 gave department heads and budget managers direct access to financial data, enabling continuous budget updates and real-time visibility into spending against objectives.
Krug described his role as balancing financial discipline with business ambition, which he supports by understanding each department's language and goals rather than acting as a reflexive "no." With modern tools in place, that approach translates into more frequent, data-driven conversations where leaders share accountability for trade-offs, supported by timely information rather than quarterly surprises.
Krug said the priority is to use tools that are safe and process-enhancing so staff can spend more time on performance and judgment instead of data entry. He defined true efficiency as "what we can do because we used the tool," capturing Sage Future's broader message that the value of AI and ERP modernization lies in capacity to make better decisions, not in the technology itself.
What This Means
Finance leaders will orchestrate AI agents rather than perform transaction-level work. System choices should prioritize self-service transparency for line managers, continuous planning capabilities and tight integration between operational and financial data so major decisions can be evaluated in real time with full financial context.
AI and ERP investments will be judged on hours saved, decision quality and agility, not on generic automation claims. The case studies presented-Byler Holdings and the Yankees-indicate that outcome-centric metrics, not technology features, will define next-generation ERP success.
Learn more about AI for Finance or explore the AI Learning Path for CFOs.
Your membership also unlocks:
