T3's AI University: Agents, Prospecting And Transformation
Published Mar. 10, 2026 - New Orleans
T3 kicked off its conference with a full day on AI. The focus: building practical agents, systematizing digital prospecting, and putting guardrails in place so teams can move faster without creating new risks.
Why it matters
AI is shifting from experiments to execution. For management, the mandate is clear: translate models and demos into measurable workflow wins, while keeping controls tight.
Sessions made the case for agent-driven automation, signal-based prospecting, and human-in-the-loop review so errors don't reach clients.
The details
Craig Iskowitz, CEO of Ezra Group, showed how advisors can build custom AI agents with standard LLM interfaces and connect them to CRM and planning tools. His point for managers: treat agents like software products-expect ongoing debugging, versioning, and support.
Victoria Toli, President of FINNY AI, outlined a practical prospecting pipeline: detect signals, identify prospects, score them, and automate outreach. She flagged limits around deep personalization and the risk of hallucinations without tight prompts and data checks.
John O'Connell, Founder and CEO of The Oasis Group, walked through use cases from text-to-speech to virtual advisory boards. He stressed three levers leaders control: prompt quality, data quality, and clear checkpoints where a human reviews the output before it ships.
The AI University event ran March 9, 2026, the first day of the T3 Technology Conference (March 9-12, 2026) in New Orleans.
What they're saying
"Whatever it can do, we have to do more." - Craig Iskowitz (Wealth Solutions Report)
"Prompting matters significantly." - John O'Connell (Wealth Solutions Report)
Manager playbook: From pilot to production
- Start with one workflow: Pick a repetitive, rules-heavy task (meeting prep, note summarization, data cleanup). Define success metrics upfront: time saved, error rate, and rework percentage.
- Build minimal agents: Use LLMs to read context, call internal tools, and return structured outputs. Add logging from day one so you can trace failures and fix them quickly. For deeper guidance, see AI Agents & Automation.
- Prospecting pipeline: Ingest signals (news, filings, website changes), score prospects, and queue compliant outreach via CRM. Keep humans approving the final message. If you oversee growth teams, this primer helps: AI for Sales.
- Data hygiene first: Centralize sources, label sensitive fields, and strip PII from prompts when possible. Better data beats bigger models.
- Prompt governance: Standardize prompts as assets with owners and versions. Test them like code.
- Human in the loop: Require review on client-facing work, model updates, and any action that moves money or changes records.
- Security and compliance: Document data flows, vendors, and model behavior. Align with a risk framework such as the NIST AI RMF.
- Measure what matters: Track cycle time, cost per task, error categories, prospect-to-meeting rate, and advisor satisfaction. Publish a simple monthly scorecard.
Practical use cases discussed
- Agentic meeting prep: Pull CRM notes, holdings, and life events into a one-page brief with talking points and compliance flags.
- Signal-based outreach: Trigger emails when a prospect changes jobs, files a Form 4, or sells a company-reviewed by a human before sending.
- Client education: Text-to-speech summaries of quarterly letters and FAQs for accessibility and speed.
- Virtual advisory boards: Aggregate client feedback into themes, priorities, and a backlog that leadership can act on.
Risks and how to handle them
- Hallucinations: Force citations, reference only approved data, and reject outputs without evidence.
- Bias: Use diverse training examples, run fairness checks, and document known limitations.
- Over-automation: Keep a manual escape hatch. If confidence is low, route to a human by default.
- Integration drift: Monitor API changes from CRM/planning tools and pin versions where possible.
The players
- Craig Iskowitz - CEO and Founder, Ezra Group
- Victoria Toli - President, FINNY AI
- John O'Connell - Founder and CEO, The Oasis Group
What's next
The T3 Technology Conference continues through March 12 with more sessions on trends and tools shaping wealth management. Expect deeper dives into integrations, data strategy, and hands-on build sessions.
The takeaway
AI is moving from novelty to operations. Leaders who pick the right workflows, invest in data and prompts, and keep humans in the loop will see faster cycles, sharper personalization, and fewer client-facing issues.
Start small. Measure relentlessly. Ship improvements every month.
Your membership also unlocks:
