Why Sage Is Building Its Own AI Instead of Just Plugging In an LLM
In this episode, Doug and Matt sit down with Taylor Brown of CLA to talk about Sage Intacct’s latest release, and why Sage is building AI into the product itself instead of plugging in a generic LLM.

About this episode
Why Sage Is Building Its Own AI Instead of Just Plugging In an LLM
In this episode, Doug and Matt sit down with Taylor Brown of CLA to talk about Sage Intacct’s latest release, and why Sage is building AI into the product itself instead of plugging in a generic LLM.
Taylor is a CPA and a director in CLA’s digital practice. She spent years as a Sage solutions consultant and still teaches accounting at UT Austin. They walk cash management, three-way matching, the Financial Intelligent Agent, the MCP AI gateway, and the fixed asset and audit trail updates.
Taylor’s accuracy warning is the one to hold onto. AI gets accounting wrong 30 to 40 percent of the time, which is why a foundational understanding still matters when Copilot, MCP, and customer-built agents show up in the workflow.
They close on crawl-walk-run change, what to do after a release, and the dad joke.
CHAPTERS
Listen to learn more about
- (00:00) Meet Taylor Brown
- (01:45) Can Teachers Do
- (03:09) Accounting Paths Beyond Big Four
- (04:57) Why Release Notes Matter
- (05:32) Cash Tracking AP AR
- (07:03) Three Way Match Upgrades
- (08:50) Cash Is King Lessons
- (12:16) Finance As Profit Center
- (13:49) AI Copilot Enters Finance
- (14:26) Training AI For ROI
- (16:10) Sage Copilot Deep Dive
- (19:51) Why Build Sage AI
- (22:47) MCP And AI Gateway
- (23:13) MCP Use Cases And Data
- (25:25) MCP Access And Licensing
- (25:56) MCP Licensing Questions
- (27:17) Agent Marketplace Gatekeeping
- (28:12) How Agents Get Sold
- (29:38) Why Agentic AI Matters
- (30:11) Customer Built Agents
- (31:38) Worst AI Misuse Risks
- (33:47) AI Accuracy in Accounting
- (35:26) Fixed Assets Release Focus
- (36:04) SFAM vs SIG Evolution
- (37:52) Reversals and Audit Trail
- (40:01) Splitting Assets Efficiency
- (40:58) CIP and Migration Advice
- (43:58) International Release Differences
- (46:05) One Thing to Do Next
- (48:21) Crawl Walk Run Change
- (49:48) Dad Joke Sendoff
Related Industries
- SaaS
- Financial Services
- Technology
- Corporate Finance
Speakers
Doug Lewis
Host | Consultant
LinkedIn
Matt Lescault
Host | CEO, Lescault & Walderman
LinkedIn
Taylor Brown
Guest | CLA
LinkedIn
Related Resources
Questions from this episode
What is an LLM in this episode?
A generic large language model you'd plug in. Sage is building AI into Intacct instead of dropping one of those in.
What is MCP?
The MCP AI gateway in this release. The connector layer for Copilot and customer-built agents.
What is three-way matching?
Checking a purchase order, a receipt, and an invoice against each other. It's in the release walkthrough with cash management.
What is the Financial Intelligent Agent?
A named Sage AI agent in this release, sitting next to Copilot and MCP.
IN THIS EPISODE
Want to be on
our podcast?
If you work in the Sage Intacct ecosystem and have a story worth telling, a lesson learned the hard way or a strong opinion on where finance is heading, we’d like to hear from you.

Sage 101: Understanding the Players and Products
Doug, Emily, and Matt cover the practical differences between Sage 50, 100, Intacct, and X3, breaking down how businesses use…
How Bill is Redefining Financial Automation
Richard Corn, Director of Product Management at Bill, joins Doug and Emily to talk about getting finance teams off manual…
The Mid-Market Integration Sweet Spot
In this episode, Doug, Emily, and Matt sit down with Scott Hollrah, founder of Venn Technology, to talk about the…



