Neon&SlateBook a call

Case study · Agentic AI cohorts for finance professionals

Nine weeks. Non-technical finance professionals became builders.

Investment professionals with no coding background built their own custom AI apps, on their own workflows, and demoed them to a jury from the association and its sponsor firms.

Request a cohort brief All results
9 weeks2 of foundations, 7 of live building
50finance professionals per cohort
0 codecustom apps built in plain English, one per participant

Objective

From AI user to AI builder.

Nobody in the room writes code. Participants direct the build in plain English and ship a working app on their own workflow: a research agent, a deal-flow parser, a client-reporting copilot. Tool-agnostic method, Claude as the build environment. The goal is daily adoption, not experimentation.

Two layers

Foundations first. Then seven weeks of building.

Module 01

LLMs and key vocabulary

  • Types of agents
  • How LLMs actually work
  • Core terminology for investment professionals
Module 02

Three steps to build an agent

  • Workflow deconstruction: sourcing, diligence
  • Process mapping
  • Connectivity and integration
Module 03

Data confidentiality and governance

  • Data classification: deal and LP data
  • What to share, what to protect
  • Firm-level AI governance and model risk

Week by week

Every agent in this program is a real engineering artefact.

  1. Weeks 1 to 2Foundations. AI key concepts, agent thinking and context engineering, data confidentiality and governance. A task-delegation model: what to build, when to use AI, when to co-produce. Two 75-minute sessions with Q&A. Pods formed.
  2. Weeks 3 to 4AI for daily productivity. The universal starter. Assistant agents installed, first demos, how to build your AI clone. Then prospect research, daily brief, meeting notes and to-do centralization. Every participant ends week 3 with a first working master agent.
  3. Week 5Research, signals and document intelligence. The major LLMs for fact checking. Industry research, competitive analysis. Research and news consolidation agents. IC memo agent.
  4. Week 6Reporting and dashboards. LP and client reporting, CIM parsing, IC memo first draft.
  5. Weeks 7 to 8Personalization sprint. No new content. Each participant extends one agent into their own workflow, in pods of three to four, with office hours. Each pod sends an async status before Demo Day.
  6. Week 9Demo Day. A formal three-hour session. Each pod presents one working agent, before/after metrics, lessons learned and an enhancement plan, five minutes plus five of Q&A, to a panel of judges from the association and sponsor firms.
Pre-recorded tutorials before classOne 60-minute live class a weekOne hour of optional office hours a weekAssistant agents, dummy data sets and connections provided by Neon & Slate

Agents built

One shared starter, then a track per role.

All participants

Agentic chief of staff

  • Tone of voice configuration
  • Daily morning briefings
  • Prospect research before a meeting
  • Meeting notes consolidation
  • CRM updates
Track: private equity

PE agents

  • IC memo draft
  • Deal flow ingestion with CIM parser and scorecard
  • Dashboard to track diligence progress
  • AI-assisted search
Track: asset management

AM agents

  • Client reporting and commentary
  • Research copilot
  • Pitch deck drafting
  • Compliance Q&A

Outcome

Every participant leaves a builder: a custom app that runs, shown to a jury, with before and after metrics.

Keeps alumni competitive in an AI-reshaped finance industry. Strengthens the association's differentiation in preparing its members for investment leadership. A proven pilot for the next cohorts.

Bring a cohort. Leave with running agents.

Request a cohort brief charlotte@neon-slate.com