Build Your First AI Accounting Agent

Overview

Investing in advisory is challenging because accounting professionals are under pressure to deliver clean books on time resulting in a reactive approach to client relationships. AI is creating a new opportunity for firms to transition between the manual work to more strategic advisory work that generates more per hour.

In this hands-on session you will learn how to:

• Build AI agents that transition your team from doer to advisor.
• Automate prep, free up capacity, and proactively identify advisory opportunities ahead of client conversations.
• Securely manage client data in your AI tool of choice (Claude, ChatGPT, Co-pilot, etc.).
• Transition your firm to higher revenue per hour advisory work and profit margins.

Attendees will leave with practical AI agent examples, prompts, and use cases they can bring back to their team to build a more proactive advisory practice, while keeping the accountant in control of the judgment, relationship, and advice.

Speaker Bios 

Logan Kieller

Logan Kieller is Growth Lead at Kick, where he ran the company's research bookkeeping firm through 2025 year end. 
He was responsible for the firm's process design, hiring, client offer and operational cadence, and led it through its second full busy season. The findings from that experiment were subsequently documented in a research paper Logan authored and Kick has published. Logan now works with accounting firms to build agents and modernize their practice. Prior to Kick, Logan ran his own consulting business and worked as a consultant at Boston Consulting Group.

Highlights

Participants will explore:

• A first AI agent, built live during the session on demo data, that preps a client meeting with KPIs, insights, and flagged changes before you walk in. 
• A workbook of advisory-focused use cases with copy-paste prompts, including cross-client benchmarking, cross-sell and upsell identification, and personal-to-business tax savings spotting.
• How to set up recurring routines so agents run on a schedule instead of waiting to be prompted. 
• A three-stage roadmap for moving staff from doer to reviewer to advisor, designed so attendees can take it back and teach it to their team. 
• A clear picture of where the accountant stays in control: the judgment, the relationship, and the advice.

Prerequisites

None

Designed For

• Public Practice

Objectives

Upon completion of this session, participants will be able to:

• Identify tasks within the client close and meeting-preparation workflow that can be delegated to an AI agent, and distinguish them from tasks that require professional judgment.
• Describe the steps required to configure an AI agent that compiles client financial data, key performance indicators, and period-over-period variances in advance of a client meeting.
• Recognize the data confidentiality and client-consent considerations that apply when client financial information is processed in a general-purpose AI tool.
• Apply a three-stage staff progression model (preparer, reviewer, advisor) to assign advisory delivery responsibilities across an existing team.
• Determine advisory service opportunities - including cross-client benchmarking, service expansion, and tax planning items - that can be surfaced from existing client financial data using structured prompts.

Preparation

None

Notice

FICPA Webinars are presented virtually. 
 
Our Webinars include a live presentation with speaker audio, video and PowerPoint presentation.  Attendees can ask questions. The presenter(s) will repeat questions to share with all attendees and answer questions as time allows. Individuals can also reach out to the presenter directly via email after the event.

For more information feel free to call the FICPA Member Service Center at 800-342-3197.