AI is everywhere. Results are not.
AI is everywhere in finance right now: There are new tools, copilots, and promises about productivity and automation seemingly every day.
In a recent piece, I argued that automation alone isn’t enough and that real progress comes from embedding intelligence directly into financial workflows, where decisions actually happen.
That idea is gaining traction. But it’s also exposing a gap. Because for most organizations, the outcomes still haven’t caught up to the hype.
According to recent Forrester research, 45% of organizations are already using AI agents and another 25% are piloting them. But only 17% have scaled those efforts across the enterprise.
That gap matters. Because it tells us something important: AI adoption is accelerating, but real impact is not.
The issue isn’t access to AI. It’s where and how it’s applied.
Learn how Emburse Expense Intelligence embeds AI as financial infrastructure.
AI without context is just noise
One of the most telling findings in the Forrester study is this: 76% of leaders say context is what unlocks the true power of AI. But only 17% say their systems are actually context-aware.
That disconnect is at the heart of why so many AI initiatives stall or don’t deliver the intended value.
Emburse research has demonstrated how most AI today operates outside the systems where decisions are made. It analyzes data after the fact. It suggests actions without full visibility into policy, workflows, or financial impact. It’s powerful in theory, but limited in practice.
Finance doesn’t run on isolated insights. It runs on connected decisions:
- What’s being spent
- Why it’s being spent
- Whether it complies with policy
- How it impacts budgets, forecasts, and risk
Without that context, AI can generate output. But it can’t generate trust. And without trust, it won’t scale.
Fragmentation is the real barrier—not intelligence
Another challenge is structural.
Enterprise data isn’t centralized. It’s fragmented—spread across systems, teams, and workflows. On average, organizations store content across 24 different systems, and 73% of enterprise data is unstructured.
That fragmentation creates friction for AI. It limits access. It slows down decisions. And it forces organizations to layer intelligence on top of systems that were never designed to support it.
The result is a familiar pattern: more tools, more complexity, and more manual effort to bridge the gaps.
But the problem isn’t that AI lacks capability. It’s that it lacks connection and therefore contextual awareness.
Governance, compliance, and trust can’t be afterthoughts
There’s also a growing realization that AI in finance can’t just be powerful. It has to be accountable.
Only 18% of organizations report having advanced governance capabilities in place for AI.
At the same time, regulatory pressure is increasing. Compliance expectations aren’t loosening. If anything, they’re becoming more complex and more immediate.
That creates a clear requirement: AI must operate within the same controls that govern financial systems today.
It needs to be auditable, policy-aware and secure by design, as well as aligned to regulatory and fiscal requirements. Not bolted on after decisions are made, but embedded directly into how decisions happen.
The shift from AI tools to financial infrastructure
For years, innovation in finance has been framed as adding capabilities: more automation and analytics, or more tools layered onto existing workflows.
AI has followed the same path. But that model is breaking because intelligence doesn’t belong on top of financial systems, but within them.
What finance teams actually need is not another AI tool. It’s infrastructure that brings intelligence into the system of record—where policy, approvals, auditability, and spend decisions already live.
That’s the shift from AI as a feature to intelligence as a foundation.
Innovation only matters when it drives outcomes
At Emburse, we’ve been deliberate about not positioning AI as a standalone productivity feature or simply building prompt-based assistants or copilots that sit outside workflows.
Instead, we’re embedding intelligence directly into the flow of financial operations—across travel, expense, AP and payments. And when it’s embedded in the right place, the outcomes become clear:
- Higher productivity: Administrative work disappears as routine tasks are automated at the source. At the University of New Mexico, that shift reduced reimbursement timelines by 75% and eliminated thousands of hours of manual processing each year.
- Stronger control: Risk is identified before it becomes a problem, not after the fact. Sasser Family Companies regained control over their spend management with real-time insights and customized reporting, achieving 75% time savings and a 60% reduction of expense systems after switching to Emburse.
- Better visibility: Finance leaders gain a continuous, real-time view of spend across systems. At BASF Catalysts, that visibility translated into more than 350 hours per month saved on analytics preparation alone, replacing fragmented reporting with a unified, decision-ready view of financial data.
- Improved compliance: Policy becomes part of the workflow itself. Bosch used automated flagging to streamline out-of-compliance transaction approvals, scaling the process to 100+ entities across 50 countries.
- Faster decision-making: Spend signals turn into action in the moment, not at month-end. At General Motors, embedding intelligence into travel workflows enabled continuous rate optimization and automated reshopping—capturing savings and improving cost control without disrupting the employee experience.
This is what intelligence looks like when it’s operational, not theoretical.
A different model for AI in finance
Many solutions in the market still treat AI as an add-on:
- A separate tool to adopt
- Another layer of cost
- Another workflow to manage
It’s an approach that creates more work, not less.
We’ve taken a different path. Emburse Expense Intelligence is built to be embedded directly into the workflows finance and travel teams already use. It brings intelligence into the system of record—so decisions happen with full context, full visibility, and full control.
That includes a new generation of AI capabilities designed for real outcomes, not just incremental improvements.
For example, instead of asking users to generate or complete expense reports manually, or through prompts, AI can now complete them automatically. Receipts are captured, categorized, validated, and routed without intervention.
It’s a touchless experience. Not a faster version of the same process, but a fundamentally different one.
The future of AI in finance is embedded—and invisible
We’re already seeing signs of where the market is heading. Standalone AI tools are becoming harder to justify. They require additional investment, introduce new risks, and often struggle to deliver sustained value over time.
At the same time, organizations are looking for solutions that are:
- Integrated into their existing systems
- Secure by design
- Scalable across the enterprise
- Focused on measurable outcomes
In other words, they’re looking for intelligence that doesn’t sit beside the system, but operates within it.
Let’s face it: AI will continue to evolve, and models will improve as capabilities expand. But the real breakthrough in finance won’t come from better algorithms alone. It will come from where intelligence lives.
Not in dashboards or layers added after the fact, but inside the infrastructure that governs how money moves.
That’s where decisions are made and where control is enforced. Because in the end, AI isn’t the goal. Better outcomes are.
Want to see Emburse Expense Intelligence in action? Get a demo.