Home/ About/ David deBoisblanc
David deBoisblanc
Co-Founder and Managing Partner, Duczer East

David deBoisblanc

Data-centric AI, AI governance and enterprise integration for regulated industries. Thirty years of leading complex technology programs, now focused on getting AI into production in a form an auditor will accept.

About

David deBoisblanc co-founded Duczer East in 2018 and leads its data strategy, AI and AI governance programs personally. He works with CIOs, CDOs, CROs and their teams in banks, insurers and healthcare organizations where data residency, model risk and audit are binding constraints.

His view, formed over three decades of enterprise programs, is that most AI initiatives stall in the same place: the meaning of the data underneath them was never settled. Without clear definitions, an ontology and a knowledge architecture, models produce answers nobody can explain or defend. Duczer East's work combines that semantic foundation with two decades of integration depth across APIs, event-driven architecture, real-time data and identity, on the platforms clients already run.

He is also an expert resource for investment fund analysts assessing enterprise technology and AI trends. He works from Atlanta and Miami, in English and Spanish.

LinkedIn profile
Areas of expertise
Data-centric AIAgentic AI systemsAgentic AI governanceEnterprise integrationSemantic layers and ontologyKnowledge graphsCloudera Data PlatformWSO2KYC/AMLAI governanceAPI managementEvent-driven architecture
Positions

What David argues

The positions behind his writing in Duczer East Insights, one sentence each, with the article that makes the case.

A system prompt is not a control.

Instructions to a model are suggestions. Enforcement lives in code, gateways and identity.

Every Prompt-Based Control Is a Future Postmortem

Governance belongs in the integration layer.

The gateway, the identity platform and the logs can enforce what policies and prompts cannot.

Governance Runs on Rails

Settle the meaning of the data before funding the agents.

Agentic projects fail on missing semantic coherence, not on compute or models.

Why Most Agentic AI Projects Fail Before They Start

Design the decision before the system.

Decide whether a workload should reach production, and on what evidence, before the architecture review.

Design the Decision Before the System

If every answer needs checking, the pilot has proven a cost, not value.

The verification tax is where most AI pilots stall.

The Verification Tax

Outside the guidance is not outside the risk perimeter.

SR 26-2 placed agentic AI outside model-risk guidance. The bank still owns the risk.

SR 26-2 Carved Out Agentic AI. The Liability Didn't Move.
Career

Three decades of enterprise programs

  1. Duczer East

    Co-Founder and Managing Partner · 2018 to present

    Leads data strategy, AI and AI governance programs for regulated enterprises. Duczer East is a Cloudera Premier Partner and a WSO2 partner.

  2. Visual Integrator Consulting

    Chief Operating Officer · 2014 to 2019

    Ran operations, client success and the P&L for an enterprise architecture and integration firm specializing in microservices and cloud.

  3. Rolta Americas

    Senior Vice President · 2009 to 2013

    Led middleware, microservices, telemetry and business intelligence solutions for clients in oil and gas, transportation, financial services and entertainment, moving the practice from SOA into early microservices work.

  4. Zanett Commercial Solutions

    Senior Vice President · 2006 to 2008

    Oracle ERP and business intelligence. Zanett was later acquired by KPMG.

  5. TracFone Wireless

    Contract CIO · 2006

    Engaged by the CEO through Titan Technology Partners to correct severe system performance problems and reorganize IT. Stabilized the systems and remediated root causes, reorganized the IT department and its governance, redesigned DevOps and deployed new Oracle ERP modules. Saved more than $6 million in IT budget while shortening time to production, and created an IT budget process that brought the supported business units into the decisions. Worked with the COO, CFO and head of audit, and helped select the permanent CIO before handing over the role.

  6. Titan Technology Partners

    Senior Vice President · 2002 to 2006

    Hired by the CEO. A key architect of Titan's move into IT outsourcing, co-developing its offerings and client strategy, and led IT strategy engagements. Titan was later acquired by Velocity.

  7. Stonebridge Consulting

    SVP Client Success and Florida Managing Partner · 1999 to 2001

    Ran client delivery and merged the sales teams after the Tactics and Stonebridge merger, leading a team of 60. During his tenure services revenue grew from $30 million to $215 million, and he helped raise $55 million in venture funding. Took over the struggling Florida region and returned it to profit.

  8. Tactics

    Vice President · 1995 to 1999

    Oracle ERP and MRP, business intelligence and analytics. Associate of the Year, 1996.

Education: Oglethorpe University; Louisiana State University. Languages: English and Spanish.

Writing

Insights by topic

David writes the practitioner posts in Duczer East Insights. Each topic collects the posts on one problem area.

Agentic AI Governance & Model Risk KYC/AML & Financial Crime Private AI & Data Architecture Semantics, RAG & Data Meaning API Management & Integration
All Insights
Get in touch

Compare notes with David

On agentic AI governance, private AI on governed data, or an integration problem that has stalled. No pitch, just a conversation.

Contact Duczer East info@duceast.com