At the Sage Future Conference in San Francisco, Manav Thiara, SVP and Technical Fellow, AI and Data Systems at Sage, described how agentic AI is transforming financial operations from basic workflow automation into systems capable of autonomous support.
Thiara explained that CFOs are increasingly looking beyond simple efficiency gains toward AI agents that can proactively identify anomalies, forecast financial outcomes, and coordinate across finance, payroll, and HR systems in real time. Throughout the conversation, he emphasized that Sage’s core differentiator is trust: building “glass box” AI systems that are explainable, auditable, and governed through multiple layers of validation, including AI models checking other AI models.
Thiara also discussed Sage’s partnership with PwC, the company’s “AI Factory” framework, and the broader societal shift toward AI-augmented work, arguing that curiosity and experimentation will be essential as enterprises adapt to rapid technological disruption.
Core Takeaways
From Automation to Autonomous Finance: Finance teams are moving beyond simple workflow automation toward autonomous AI agents that proactively identify anomalies, generate forecasts, automate reporting, and continuously monitor business operations across systems.
“Glass Box” AI Transparency: Sage’s AI trust strategy focuses on replacing opaque “black box” systems with explainable AI frameworks that provide visibility into data lineage, calculations, auditability, and override controls for enterprise customers.
AI Checking AI: Sage uses layered validation systems where LLMs and AI agents validate the outputs of other AI systems, creating scalable governance frameworks capable of monitoring millions of transactions without requiring human review of every action.
AI Expands Human Capability: Thiara argues that AI will not eliminate human responsibility, but instead free workers from repetitive tasks while enabling broader, more generalized skill sets across engineering, finance, and business operations.
Key Quotes
The Shift Toward Fully Autonomous Financial Systems
“I think even in 2024, we were just kind of starting to talk about: how do we automate the workflows and how do we reduce the time? And fast forward to now, just two years later, we are now talking about fully autonomous agents.
The CFO doesn’t just want to know the opportunities where there could be efficiencies. Rather, how can we have the agents running end-to-end that can truly tell us proactively that this is where the gap is? What does my future forecasting look like? How do we flag any anomalies?”
Turning the AI “Black Box” Into a “Glass Box”
“What we are really saying with PwC is: let’s take that black box and take that work that we have done within AI Trust Label and expand it to create a blueprint and a framework where organizations and our customers can really understand not only where their data’s coming from, but what the tools and resources are that can help to accelerate their businesses using these AI technologies.
Instead of the black box, we are having a glass box, which is transparent. So you have full visibility into what’s happening, you have the confidence and you have the trust, which is at the center of everything Sage is doing, so you know not only where your data is being used, but how is it used?”
Why Sage Uses AI Systems to Validate Other AI Systems
“If we can have a large language model or frankly a small language model spit out an output, we will send it to another agent that will validate that output. So now we have that validation layer. In fact, we have sometimes double or triple validations on agents on top of agents.
If you are now running millions of transactions every day, every month, it’s obviously not scalable to have human beings check every single thing, but what is scalable is that we can build the framework and the blueprint of our AI development in a way that it is always going through those controls.”
AI as the Next Great Technological Disruption
“So you’re right, [our current time period] is quite a disruption. It is happening at an unprecedented scale and speed that has never happened before.
But if you are curious and if you’re open-minded, there is no better time, frankly. Roles are changing. I tell my engineering teams that now you can be product managers. To my product manager friends, I tell them, you can be engineers now. Everybody, the way AI has democratized these tools, it’s like everybody can be there.”
