Live Replay WebinarBasic level

Your Firm's Personal AI: Custom Copilots and GPTs That Know Your Business

Most practitioners have experimented with AI, but few have moved beyond using it for simple emails or basic summaries. The real breakthrough happens when practitioners move past generic interactions and learn to steer the AI toward hyper-focused outcomes. By utilizing detailed, precise prompts and providing the model with specific context, you can transform a general-purpose tool into a specialized assistant. This course explores how solo practitioners and small firms are moving their actual data into the AI workflow to get detailed answers that are grounded in their own firm’s knowledge and professional standards. We will navigate the landscape of “Custom AI,” from the simplicity of the GPT Store to the privacy-first world of self-hosted models like Ollama and OpenWebUI. We delve into how emerging technologies like Retrieval-Augmented Generation (RAG) and Model Context Protocol (MCP) are “opening doors” for firms to connect their files directly to the AI without traditional coding. This session is designed to point you in the right direction, helping you compare various tools and API options like OpenRouter so you can determine which technical path best fits your firm’s comfort level and privacy requirements.

CPE Credits2

Field of Study Information Technology

Instructor [email protected]

PublishedDec 3, 2025

NASBA sponsor ID167619

Ways to save

Save $10.56 with a credit pass

At the 40-Credit Pass rate ($22.22 a credit with the bonus), this 2-credit course comes to about $44.44: $10.56 less than buying it on its own.

Or start smaller: the 10-Credit Pass ($275.00) pays for it and leaves you 8 credits for more.

Year-end bonus: up to 5 free credits through Dec 31.

Certain conditions apply: not valid for the Federal Tax Update. Pass terms

Add the 10-Credit Pass Compare passes →

Watch

4:02 Create QuickBooks Records by Just Asking Your AISee how a custom AI copilot can execute real tasks in your business software without manual entry. In this clip, Stephen Yoss demonstrates accessing QuickBooks data, creating new customers, and generating invoices—all by simply describing what you need in plain English. Watch how automation eliminates repetitive data entry and frees your team for higher-value work.

3:09 MCP: The Universal Plug for AI and Your DataStephen M. Yoss explains how Model Context Protocol (MCP) simplifies connecting your business data to any AI tool—no complex database work required. Just like USB-C works universally across devices, MCP acts as a standard plug for your accounting, customer, and course systems to talk to ChatGPT, Gemini, or anything else. Learn how to stop building custom integrations and start asking natural questions that pull the right data instantly.

3:09 Don't Ask AI for Finished Output on First TryStephen Yoss breaks down the critical principles for effective AI prompts: give it the right context (jurisdiction, dates, specifics), never paste entire files when snippets work, and expect to build dozens of specialized prompts—not just one. Learn why treating AI responses as drafts and iterating until perfect is essential risk management for firms. Watch to master the fundamentals of custom AI tools that deliver exactly what you need.

2:14 Why Your AI Output Is Bad (Hint: It's Not the AI)Stephen M. Yoss breaks down the core principle behind getting useful results from custom AI: input quality directly shapes output quality. Learn why vague prompts deliver generic answers, and how providing specific context, entity details, and up-to-date information transforms AI from boilerplate generator into a powerful business tool. Watch to understand what your AI actually knows—and what you need to tell it.

What's included

Every registration comes with the course materials — yours to keep.

Course handout PDF
Additional course files PDF, ZIP

Course details

Recommended CPE credit: 2
Field of study: Information Technology
Program level: Basic
Delivery method: Group Internet Based
Prerequisites: None
Advance preparation: None
Course number: 2149

CPE Today (Devmatics, LLC) is registered with the National Association of State Boards of Accountancy (NASBA) as a sponsor of continuing professional education on the National Registry of CPE Sponsors (Sponsor ID 167619). State boards of accountancy have final authority on the acceptance of individual courses for CPE credit. Complaints regarding registered sponsors may be submitted to the National Registry of CPE Sponsors through its website, nasbaregistry.org. For information about our refund, complaint, and program cancellation policies, see Company Policies or contact [email protected].

💬 Ask this course

Not sure it covers what you need? Ask a real question — answered from the actual course content, so you can decide before you enroll.

Answers come from this course's own material and can be imperfect — the full details are in the course itself.

Common questions

Will this course help me if I'm new to AI tools and have no prior experience?

Yes—this course requires no prerequisites and is designed for solo practitioners and small firm owners at any experience level. The instructor starts with the fundamentals of writing better prompts and moves through progressively more advanced techniques like custom GPTs, projects, and Model Context Protocol (MCP), so you can learn at your own pace and apply what works for your firm immediately.

Can I actually use these tools with my firm's real data, or is it risky to feed client information to ChatGPT and Claude?

The course covers a complete privacy spectrum—from consumer tools where your data may be used to train models, to business-tier subscriptions where it's not, to self-hosted options where you keep complete control. The instructor shows how to redact sensitive details (like Social Security numbers and dates of birth) while keeping enough context for AI to work effectively, and emphasizes using business accounts with proper internal controls rather than free personal accounts.

Will learning this actually change how I work day-to-day, or is it just theory?

This is entirely hands-on and practical. The instructor demonstrates building engagement letters, running analyses, creating invoices directly in QuickBooks through Claude, and generating presentations from live data—all shown step-by-step in the classroom. You'll leave with working templates, a prompt library approach, and real examples you can adapt immediately for your own engagements.

I use QuickBooks and Office 365—can these AI tools actually connect to my existing systems without me rebuilding everything?

Yes, through Model Context Protocol (MCP), which the course covers in depth. QuickBooks, Outlook, Google Drive, and hundreds of other business systems now have official connectors that plug into ChatGPT, Claude, or Gemini—no rebuilding required. The instructor demonstrates creating a QuickBooks customer and invoice directly from Claude using only natural language, showing how one connector works across any AI tool you choose.

How much hands-on building versus lecture is in this course?

The course is heavily weighted toward live demonstration and practical examples, especially in the middle and later sections. You'll watch the instructor build custom GPTs in ChatGPT and Claude, upload firm templates, create projects that remember context, and connect real business systems like QuickBooks—all with real file uploads and step-by-step walkthroughs that you can replicate in your own practice.