← Multimodal Sources 🕐 4 min read
Multimodal Sources

What 30K JPMorgan AI Agents Taught Me

> Over the course of those first few months there, we went from zero users to 250,000 users.

Show: Beyond the Pilot · Publisher: VentureBeat · Host: VentureBeat editorial

Episode URL: https://www.youtube.com/watch?v=rLxJzeRGzV8

Publish date: 2026-04-13
Duration: 2790.0s
Default source credibility: HIGH — Named F500 practitioners on-record with production metrics. VentureBeat editorial vetting. Treat vendor-sponsored segments as MEDIUM.

  • JPMorgan’s AI strategy focuses on a central platform for connectivity, not on building models, which they view as a commodity. The real value is integrating AI into the firm’s data and processes.
  • They use a dual strategy: a bottom-up approach empowering employees with reusable AI tools, and a top-down approach to re-engineer entire cross-functional business processes for AI.
  • An “innovation flywheel” drives adoption: the central AI team identifies common user-found gaps, builds solutions centrally, and expands platform capabilities for all, sparking new use cases.
  • The platform scaled from 0 to 250,000 users in its first few months. Today, 1 in 2 JPMorgan employees worldwide use the internal AI tools daily, a testament to their viral adoption strategy.
  • Data privacy was paramount from day one. Building their own platform was a strategic choice to ensure full control and understanding of data lineage, never sending proprietary data to external LLMs.

Extracted quotes

# Credibility Speaker Org Timestamp Topic Quote
1 HIGH Derek Waldron (Chief Analytics Officer) JPMorgan Chase 00:02 02-corporate-tools Over the course of those first few months there, we went from zero users to 250,000 users.
2 HIGH Derek Waldron (Chief Analytics Officer) JPMorgan Chase 00:14 02-corporate-tools Today, one in two JP Morgan employees around the world use it nearly every day.
3 HIGH Derek Waldron (Chief Analytics Officer) JPMorgan Chase 03:00 01-ai-native-landscape The actual long-term bottleneck for driving maximum value from this technology was not going to be about the model. It was going to be about how the technology connects into the technology estate and data and process estate of an enterprise.
4 HIGH Derek Waldron (Chief Analytics Officer) JPMorgan Chase 05:07 07-adoption-challenges We didn’t give our data to large language models. Data privacy was the first and foremost consideration. Part of the reason that we wanted to build a platform ourselves was that we understood the lineage of all of the data, because security was the primary concern.
5 HIGH Derek Waldron (Chief Analytics Officer) JPMorgan Chase 27:40 02-corporate-tools We have a team that surveils all of this and surfaces what those gaps are. Then we have a process to triage those so that we can solve them centrally. When we solve them centrally, we not only solve those individuals’ problems, but we also expand the capabilities, which then creates more ideas, and we’re in this flywheel effect.
6 HIGH Derek Waldron (Chief Analytics Officer) JPMorgan Chase 43:36 12-agent-workers Businesses run on long processes that cross multiple different types of teams. If we want to be able to really move the needle on those processes, there has to be a strategic element to actually rethink what the process itself will need to look like in a world of AI and AI agents.

Per-quote detail

1. Derek Waldron — JPMorgan Chase (00:02)

Over the course of those first few months there, we went from zero users to 250,000 users.

  • Stat: 250,000 users in the ‘first few months’ of launch.
  • Credibility: HIGH — Named executive at a Fortune 500 company providing a specific user adoption metric and timeline for a production system.
  • Topic tag: 02-corporate-tools

2. Derek Waldron — JPMorgan Chase (00:14)

Today, one in two JP Morgan employees around the world use it nearly every day.

  • Stat: 1 in 2 employees use the platform daily as of mid-2024.
  • Credibility: HIGH — Named executive at a Fortune 500 company providing a specific, current, firm-wide adoption metric with a clear denominator.
  • Topic tag: 02-corporate-tools

3. Derek Waldron — JPMorgan Chase (03:00)

The actual long-term bottleneck for driving maximum value from this technology was not going to be about the model. It was going to be about how the technology connects into the technology estate and data and process estate of an enterprise.

  • Credibility: HIGH — Named executive at a Fortune 500 company sharing a core strategic insight that drove their successful AI platform development.
  • Topic tag: 01-ai-native-landscape

4. Derek Waldron — JPMorgan Chase (05:07)

We didn’t give our data to large language models. Data privacy was the first and foremost consideration. Part of the reason that we wanted to build a platform ourselves was that we understood the lineage of all of the data, because security was the primary concern.

  • Credibility: HIGH — Named executive at a Fortune 500 company explaining the security rationale behind their decision to build an internal AI platform.
  • Topic tag: 07-adoption-challenges

5. Derek Waldron — JPMorgan Chase (27:40)

We have a team that surveils all of this and surfaces what those gaps are. Then we have a process to triage those so that we can solve them centrally. When we solve them centrally, we not only solve those individuals’ problems, but we also expand the capabilities, which then creates more ideas, and we’re in this flywheel effect.

  • Credibility: HIGH — Named executive at a Fortune 500 company describing a specific internal process (‘innovation flywheel’) for scaling AI capabilities based on user feedback.
  • Topic tag: 02-corporate-tools

6. Derek Waldron — JPMorgan Chase (43:36)

Businesses run on long processes that cross multiple different types of teams. If we want to be able to really move the needle on those processes, there has to be a strategic element to actually rethink what the process itself will need to look like in a world of AI and AI agents.

  • Credibility: HIGH — Named executive at a Fortune 500 company outlining their top-down strategy, which focuses on process re-engineering rather than just task automation.
  • Topic tag: 12-agent-workers

Extracted 2026-04-13T23:34:59 via scripts/podcast_mine.py (Gemini gemini-2.5-pro).