AI Agents
in Construction
Finance
Join us on Tuesday, June 23rd from 2:00–3:00 PM ET for a candid session for construction finance leaders on using AI agents in real workflows — and what the GAIA2 benchmark reveals about where they're strong vs. where they still break.
· 2:00 – 3:00 PM ET for Finance Leaders
Not a sales pitch.
A real conversation.
This is a practical discussion for construction finance and operations leaders who want to understand where AI agents can create value today — and where they still fall short in messy, real-world environments.
We'll look at how AI agents are being used in ERP, project management, AP workflows, and internal operations tools. You'll leave with a clearer picture of what the next 12–24 months looks like.
Join the Interest List
What we'll cover
A practical, hands-on walkthrough built for people who actually use — or are about to use — these tools.
Real Agent Workflows
How I'm personally using Claude Coworker, Nemoclaw, and Claude Code in day-to-day business tasks.
What Works Today
Concrete task categories where AI agents already deliver reliable value in business workflows.
Where Agents Still Break
Messy, asynchronous, real-world environments expose limits. We'll look at failure patterns honestly.
GAIA2 Benchmark
A plain-English walkthrough of GAIA2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments.
Construction Finance Lens
What this means for teams running ERP, project management, AP automation, and ops tools.
12–24 Month Outlook
Where the technology is heading and how to position your team ahead of the curve.
Built for construction finance leaders
If you're responsible for financial operations, technology decisions, or just want to understand what AI agents can realistically do for a construction business — this session is for you.
Format (roundtable, webinar, or live demo) will be shaped by interest. Sign up to weigh in.
What GAIA2 tells us about AI agents
The GAIA2 benchmark evaluates LLM agents across dynamic, asynchronous real-world tasks. The results reveal a nuanced picture — strong performance in some areas, significant gaps in others.
Structured Task Execution
Agents perform well on well-defined, sequential workflows with clear inputs and outputs.
Asynchronous Environments
Multi-step tasks with delays, handoffs, or changing context still trip up most agents.
Messy Real-World Data
Unstructured, incomplete, or contradictory inputs remain a major reliability challenge.
Join the Interest List
Leave your info below if you’d like to attend or learn more. I’ll use responses to decide whether to host this as a small roundtable, webinar, or live demo session.