search icon

IDC Insights

The emerging economics of AI in finance

AI speeds up finance, but time spent verifying outputs from generic AI models limits how far autonomy can scale. Review the latest IDC research on AI in finance. Learn where agentic AI is gaining traction, why explainability is becoming essential, and what finance leaders should consider as AI adoption grows.

Join us for a webinar to explore the latest IDC research

Trust, transparency, and agentic AI
Date/time: Wednesday, Sep 30, 2026 at 2:00 p.m. ET

Key topics:

  • Reduce time on verification.
  • See where agentic AI is scaling.
  • Strengthen AI with explainability.
Register for webinarDownload report
30+ hours lost per week verifying generic AI outputs*

1 in 5 finance leaders lose 30+ hours/week verifying AI outputs

Executive summary

Everyone assumes AI economics are about intelligence and automation. The data suggests the real constraint: trust. As AI-generated work increases, finance teams spend significant time verifying outputs, explaining decisions, and ensuring results are defensible, often recreating the work AI was meant to reduce. 

Spent weekly verifying AI

26%

AI gains lost to verification

71%

Reject systems without reasoning trace

Verification is where AI productivity disappears

Insight #1

AI can generate outputs quickly, but finance teams still need to validate assumptions, review exceptions, and confirm compliance. Nearly 1 in 5 finance leaders spend 30+ hours each week validating AI-generated outputs, turning verification into a new category of finance work. As AI adoption grows, trust, explainability, and governance are becoming at least as important as speed.

Transparency can recover lost AI productivity

Insight #2

26% of AI time savings are lost to verification, explanation, and reconstruction work. When AI systems can't clearly show and explain their reasoning, teams spend hours tracing assumptions, validating logic, and recreating the analysis—wasting the time AI was supposed to save.

Lack of transparency doesn’t remove labor. It shifts it into explanation work that slows scale.

Autonomy is advancing, but only in narrow workflows

Insight #3

38% of organizations expect to deploy transactional AI agents within two years. Adoption is focused on narrow workflows where outcomes are reviewable. High-volume, rule-based processes (such as AP, reconciliations, and recurring close tasks) are where finance autonomy is advancing first.

Where autonomy is scaling first

Accounts payable (AP)

Reconciliations

Recurring close tasks

Report comparison

Accuracy is not enough to create trust

Insight #4

In finance, accuracy without explainability isn’t enough. 71% of leaders would veto a 99% accurate AI system if it lacked a reasoning trace.  

This is the divide between generic AI built for just speed and finance-grade AI that you can trust. Without visibility into how outputs are generated, even accurate and high performing AI fails to scale. Without visibility into how outputs are generated, even accurate and high performing AI fails to scale. 

Transparency accelerates adoption more than risk tolerance. Leaders are more willing to increase throughput in existing workflows than to expand AI into higher stakes decisions without explainability.

This is why finance-grade AI (AI systems designed specifically for audit trails, explainability, and compliance like Sage AI) outpaces generic solutions in adoption and trust.

Explore finance-grade AI

Sage Ai is built for accounting and finance, not from generic models. Every recommendation includes the reasoning and surfaces relevant records, so teams can act faster with confidence.

Automate repetitive, manual finance tasks like data entry and reporting

Explore Sage Ai
* 19% of surveyed finance leaders reported spending 30 or more hours per week validating AI-generated outputs. Source: Sage-sponsored IDC research, The Emerging Economics of AI in Finance, 2026.