Technology & Innovation

Agentic AI explained: A smarter future for high-performing finance teams

Agentic AI is the next evolution in high-performance finance. Learn where the technology is now, according to top CFOs, technologists, and generative AI experts.

We’re entering a new world of agentic AI.

If generative AI is about producing text, images, and answers on demand, agentic AI is about acting on behalf of users autonomously.

For CFOs and finance teams staying ahead of the curve, this shift could be as significant as the move to the cloud in the early part of this century.

This article will explore what agentic AI means in practice—and how you can start preparing now. Here’s what we discuss:

What is agentic AI, and why does it matter now?

Put simply, agentic AI is like a digital analyst. It thinks for itself.

As Kamales Lardi, author of Artificial Intelligence for Business, puts it: “LLMs are the brains, agents are the worker bees.”

They’re purpose-built systems that combine language understanding with tools, data, and reasoning to complete tasks with minimal human input.

At Sage Future 2025, Sage’s Chief Technoogy Officer, Aaron Harris, stated that the next generation of AI will do more than answer questions and prompts.

Instead of waiting for dashboards to refresh or prompts to be entered, agentic AI observes what’s happening across your business, interprets the signals, and takes the initiative—often before you know something’s changed.

Agentic AI doesn’t wait for you to ask a question or click a button. It’s always running in the background, watching what’s happening in your business, spotting changes, and acting on your behalf.

Where traditional AI tools automate tasks or generate content, agentic AI can:

• Monitor data across multiple systems.
• Recognise patterns and exceptions.
• Decide what matters.
• Take action—or notify—without being prompted.
• You no longer need to look for insight—insight can find you.

As Kevin Quirk of AI Bridge Solutions explains: “AI agents are not just futuristic novelties; they’re delivering measurable value today—automating workflows, closing operational gaps, and increasing productivity without major overhead.”

It prompts a simple but important shift in thinking:

• Are you still waiting for insight—or is insight finding you?
• Do your systems offer context—or just data?
• Are you building trust—or borrowing it?

We don’t need AI that’s passively waiting for a prompt. We need AI that’s always on… Combining reasoning, memory, and autonomy.

Aaron Harris, Chief Technology Officer, Sage

How generative AI builds the agentic AI foundation

Agentic AI represents the future and Sage Copilot is already leading the way.

As Dan Miller, Sage EVP of Financials and ERP, highlighted during a keynote at Sage Future, Sage Copilot is not a prototype or concept. It’s already live, and embedded in products like Sage Intacct, Sage X3, and more.

It’s delivering on the promise of high-performance finance operations.

Sage Ai and Copilot aren’t just reacting to what’s happened. They’re helping you stay ahead of what’s coming next.

Dan Miller, Executive Vice President of Financials and ERP, Sage

Sage Copilot transforms day-to-day finance work by turning routine processes into intelligent, streamlined interactions. Examples include:

• Instantly surfacing overdue invoices by value and contact.
• Assisting with reconciliation, variance analysis, and period-close tasks in real time.
• Delivering answers sourced from Sage documentation via conversational AI.
• Providing real-time insights through 360-degree dashboards across inventory, sales, and operations.

Sage Copilot is trained on product documentation, accounting standards, and customer language—a foundation on which agentic AI operates with deep financial trust and contextual understanding.

At Sage Future, Sage announced a partnership with the AICPA to train its models on the profession’s official content.

This is a first of its kind… A signal to the industry that we’re building AI you can trust.

Aaron Harris

This move reflects a shift for a traditionally conservative profession: accounting leaders aren’t just ready to adopt AI. They want to shape it.

The building blocks are already there: reasoning, contextual awareness, and domain-specific expertise. The next phase is autonomy.

Automation to autonomy: Inside the new finance stack

Agentic AI doesn’t replace the need for clean data or strong processes. It amplifies them.

That’s why the Sage “innovation north star,” as Dan Miller explains it, is built around three connected ideas:

• Continuous accounting.
• Assurance.
• Insight.

Together, these form future architecture for finance teams, with agentic AI is poised to power each one.

1. Continuous accounting: Getting data in, friction-free

With AI-enhanced workflows like automated bill entry, PO matching, and classification, finance teams are already cutting manual work and reducing errors.

But with AI agents, it goes further: intelligent systems won’t just process your documents—they’ll monitor exceptions, detect anomalies, and correct patterns for you.

2. Continuous assurance: Real-time trust in the numbers

Traditionally, finance teams wait until month-end to reconcile data, analyse variances, and flag issues. With Sage Copilot and agentic AI, those tasks can shift upstream, creating a near real-time flow of trusted data.

“Tasks like reconciliation and variance analysis are starting earlier, flagged sooner, and completed faster,” says Dan Miller.

AI agents can proactively detect deviations in your expected cash flow, flag them before they snowball, and even suggest follow-ups, long before a report is pulled.

3. Continuous insight: From reports to recommendations

This is where the shift to agentic AI becomes clearest. Instead of running reports or refreshing dashboards, you can receive contextual, actionable updates based on business changes.

You want AI that doesn’t just pull a report. You want it to analyse sales transactions, churn rate, and tell you what matters.

Aaron Harris

This is insight that leads, not lags. It’s how agentic AI transforms finance from a reactive function to a strategic driver, and it’s something every finance leader should have in their toolkit.

How to move to fully agentic systems in finance

AI’s value to you lies in its ability to create the conditions for faster, smarter, more autonomous decision-making.

That’s exactly what Sage Copilot has begun to deliver. By embedding generative AI into core workflows, Sage Copilot offers real-time assistance, contextual insights, and time-saving automation.

But these capabilities are only the start.

In finance, expect a world of orchestrated, agentic intelligence, where autonomous systems understand priorities, coordinate complex tasks, and act on behalf of your business.

Agentic AI isn’t a sudden, unexpected leap. It’s the result of years of strategic innovation.

From task automation to orchestrated intelligence: the journey to agentic AI has gone through many iterations. For instance, for over a decade, automation, machine learning, and now generative AI has been embedded into Sage products.

Each wave of AI advancement has deepened expertise, sharpened customer focus and performance, and strengthened foundations of trust.

Here’s how that evolution unfolded:

Wave 1: Task-based AI

Sage builds machine learning into core workflows—automating repetitive tasks like invoice categorisation or anomaly detection. This delivers speed, accuracy, and efficiency for clearly defined processes.

Wave 2: Generative AI

Sage Copilot represents the generative wave. It helps finance teams:

• Draft month-end close reports.
• Answer “How do I…” questions via chat.
• Surface relevant data on request.

These tools accelerate decisions and reduce manual work—but still rely on prompts to get started.

Wave 3: Agentic AI

Today we’re entering an age of orchestrated intelligence. With agentic AI, autonomous systems:

• Operate continuously in the background.
• Understand business context.
• Connect multiple workflows.
• Take initiative across multi-step processes.

To explore the Sage AI journey and see how these technologies can power your business, visit the Sage Ai Hub.

Seeing Agentic AI in action for finance

To illustrate what agentic AI looks like in practice, during his Sage Future keynote Aaron Harris walked through a familiar CFO scenario: tracking revenue against forecast.

Traditionally, this means logging into a dashboard, comparing figures, and hoping nothing has shifted since the last update.

But that’s reactive.

It also assumes the dashboard tells the whole story. But the real signs of trouble often appear before the numbers even hit your accounting system.

The problem with this is you’ve got to remember to go to the dashboard… And it’s really possible that you’re going to miss something because what’s changed hasn’t hit your accounting yet.

Aaron Harris

For example, you might be meeting your revenue targets, but if the business is signing more short-term, lower-value contracts, that could lead to higher churn.

These early signals aren’t in your accounts and often live upstream—in sales pipelines, contract terms, or customer behaviour—long before they hit the books.

That’s where agentic AI steps in.

Instead of relying on reports, you can instruct Sage Copilot to monitor a key metric—in this case, your revenue forecast—and alert you if anything changes.

We’ve told Sage Copilot to just let us know if the forecast changes.

Aaron Harris

And here’s where the AI agent goes beyond standard automation: even though bookings and revenue are tracking to plan, Sage Copilot notices the shift in contract mix, understands the risk of churn, and proactively flags a forecast change.

Agentic AI automatically notifies the finance team, allowing them to respond quickly.

CFOs don’t just want to know the answer—they want to act on it. They want to pull others into the conversation so they can take steps to correct it.

Aaron Harris

This is the shift: from static dashboards to proactive, context-aware systems that spot changes early, explain what’s happening, and make it easier to respond.

The AI Trust label: the real differentiator

Finance teams need fast answers, but they must come from accurate, explainable, and secure systems. Any respectable approach to AI must go beyond performance to something deeper: trust.

As Aaron Harris explained, Sage deliberately decided to train its own large language models, not because off-the-shelf models like GPT-4 lacked power, but because they lack precision, consistency, and accountability in a financial context.

Generic AI is risky and unacceptable in a financial setting.

Hallucinations can’t be tolerated.

Audit trails matter.

Outcomes must be consistent.

We need these models to be experts at accounting and compliance… You don’t want AI behaviour that drifts over time.

Aaron Harris

Kamales Lardi warns: “The very autonomy that makes AI agents powerful also poses danger. Without governance, the absence of oversight raises concerns about fairness, bias, and accountability.”

Kevin Quirk adds: “Every AI agent we design operates within clear boundaries and is fully auditable. That’s how you earn trust from employees and customers alike.”

So instead of chasing scale, Sage focuses on relevance—its internal models, built on 7 billion parameters (vs GPT-4’s 2 trillion), are trained on:

• Accounting standards and exams.
• Real-world customer and compliance data.
• Industry-specific terminology.
• Sage product documentation.

This supplies predictable, audit-ready outcomes—a must for CFOs and auditors alike.

To bring that commitment full circle, Sage introduced the AI Trust Label—a built-in feature that will accompany AI-powered features within Sage software products.

This label will:

• Explain how the model was trained.
• Show how accuracy is measured.
• Outline safeguards to protect data.
• Disclose how bias is minimised.

Trust is the biggest barrier to adopting AI… We’re not just building artificial intelligence. We’re building authentic intelligence.

Aaron Harris

With trust embedded at every level, you have a software company positioned to use agentic AI and lead its responsible evolution in finance.

Final thoughts: Tips for high-performing finance teams

As Kamales Lardi warns, the autonomy of AI agents demands oversight and accountability—two areas CFOs must bake into their adoption strategies from day one.

Encourage your finance team to think about automated, proactive, intelligent workflows that scale trust, insight, and action. It’s about smarter systems that think ahead.

The narrative that AI agents replace jobs is an oversimplification. Instead, they evolve roles.

Kevin Quirk

Agentic AI represents the next big thing in finance: autonomous digital analysts that observe, reason, and act, freeing people to focus on judgment, leadership, and strategy.

As Sage Copilot continues to expand its capabilities, you have building blocks:

• Real-time access to actionable insights.
• Autonomous monitoring of key metrics.
• Seamless collaboration.
• Embedded trust through transparent, domain-specific AI.

Don’t wait for the month-end. Don’t rely on yesterday’s numbers.

The future isn’t just arriving — it’s already here. And it’s available to you right now.

Dan Miller

Build a finance stack that listens, learns, and leads.