ATL246: Difference between revisions
No edit summary |
No edit summary |
||
| Line 36: | Line 36: | ||
“Show me all the meals in the past year where I’ve spent more than $100—boom—and it just goes and gets that data.” | “Show me all the meals in the past year where I’ve spent more than $100—boom—and it just goes and gets that data.” | ||
== | == Companies Mentioned == | ||
{| class="wikitable" | {| class="wikitable" | ||
Revision as of 01:46, 14 February 2026
ATL246 – Model Context Protocol
Release Date: February 13, 2026 Duration: 17:21 Hosts: Randy Johnston, Brian F. Tankersley Source File: 20260220 Model Context Protocol ATL246 (Audio)
Episode Summary (200 Words)
In Episode ATL246 of the Accounting Technology Lab, Randy Johnston and Brian Tankersley explore the rapidly emerging Model Context Protocol (MCP) and why it may become one of the most important developments in Artificial Intelligence for accounting firms in 2026.
MCP acts as a standardized connector that allows AI tools to securely communicate with accounting and business systems without building custom integrations for every application. Think of it as a universal adapter—or perhaps more accurately, a slightly “janky” Bluetooth layer—that enables AI to query multiple systems simultaneously.
The hosts explain how MCP improves consistency, enhances security through permission-based access, and gives firms greater control over what AI can see and do. Unlike traditional APIs, MCP standardizes machine-to-machine conversations, allowing tasks like drafting reconciliations, summarizing transactions, explaining variances, and detecting anomalies without exporting data to spreadsheets such as Microsoft Excel.
However, governance is critical. Firms must define permissions (preferably read-only to start), authenticate securely using OAuth or token-based access, test with non-production data, update Written Information Security Plans (WISP), and revise engagement letters to address AI usage and client consent.
MCP is not a replacement for accounting systems or autonomous AI. Instead, it is the practical integration layer that makes AI usable today—transforming how accountants access, analyze, and act on financial data within their systems of record.
Memorable Quotes (with Timestamps)
- 00:11 – Randy Johnston
“An MCP is a standardized connector that lets AI tools securely talk to business systems… You can think of an MCP like a universal adapter.”
- 03:06 – Brian Tankersley
“It’s very much like this machine-to-machine conversation that you used to have in person to get the information you needed.”
- 09:54 – Brian Tankersley
“Writing back to data—you do it at your peril.”
- 15:40 – Randy Johnston
“The real work is going to be in your learning and in the governance, not so much the technology.”
- 16:15 – Randy Johnston
“Show me all the meals in the past year where I’ve spent more than $100—boom—and it just goes and gets that data.”
Companies Mentioned
| Product / Company | X (Twitter) | |||
|---|---|---|---|---|
| CPA Practice Advisor | @CPAPracAdvisor | CPA Practice Advisor | CPA Practice Advisor | @cpapracticeadvisor |
| Dext | @dextofficial | Dext | Dext | @dextofficial |
| Xero | @Xero | Xero | Xero | @xero |
| Zoho Books | @ZohoBooks | Zoho | Zoho | @zoho |
| QuickBooks Online (Intuit) | @QuickBooks | QuickBooks | Intuit QuickBooks | @quickbooks |
| Microsoft Outlook | @Outlook | Microsoft | Microsoft | @microsoft |
| Expensify | @Expensify | Expensify | Expensify | @expensify |
| Zoho Expense | @ZohoExpense | Zoho | Zoho | @zoho |
Strategic Promotion Angles
- MCP standardizes AI access to accounting systems.
- Reduces risk while improving consistency.
- Governance, permissions, and policy updates are critical.
- Makes AI usable today without exporting data.
- Enables fraud detection, advisory insights, and anomaly identification.
- Firms should begin experimenting now to build competency.
