CDOIQ 2026 · Data Governance · Agentic AI · Semantics

CDOIQ 2026: My Honest, No-BS Takeaways

Juan Sequeda's grounded recap of the 20th Annual CDOIQ Symposium — foundational governance work, the tech-company divide, the changing CDO role, agentic AI in production, and semantics as first-class infrastructure

KG curated by kg-generator, rdf-infographic-skill, and Claude Sonnet 5 on behalf of Kingsley Idehen

"These are not prerequisites you check off before the real work begins. They are the real work."

Part One

🏛 Most Organizations Are Still in the Middle of Fundamentals

You cannot shortcut the foundation. Governance, data quality, ownership models, and semantics are not prerequisites to the real work — they are the real work.

Making Trust Measurable

Bill Snider presented the Trusted Data Score, a five-category rating (ownership, curation, quality, protection, observability) applied to ~100 priority lake houses out of 12,500 schemas.

Governance focused on ~100 of 12,500 schemas

Sustained and Disciplined Execution

Caroline Serio and team rebuilt catalog adoption over four years through gamification, community building, and a shift to data product ownership.

Data Summit grew to 400+ attendees, 20 booths

The Diamond Model

Kris Mork, CDO of Leidos, presented a diamond model (Attention, Direction, Energy, Congruence) via a horse-herd leadership metaphor.

Don't Boil the Ocean

Narayanan Nair, CDO, identified 15-20 critical data elements across 15-20 systems most likely to feed AI, and consolidated unstructured documents before applying AI.

Part Two

The Tech Company Divide

ADP and Capital One built genuinely impressive data platforms over a decade because technology is native to their organizational DNA. The honest no-bs question: if they can do this, can everyone else?

The Muruntau Gold Mine Analogy

Amin Venjara, CDO of ADP, compared enterprise data to the Muruntau gold mine: gold discovered in 1958 wasn't realized until the mining infrastructure was built.

Federated Hub-and-Spoke Data Products

Amy Lenander (CDO) and Christina Egea (SVP Product Management) presented Capital One's model. Capital One formalized the ontologist job title in 2022.

Related: Bethany Sehon's 2019 knowledge graph pilot

"Institutional delusion, thinking you are what you're not."

— discussed at the #HonestNoBS Dinner
Part Three

🧭 The CDO Role, What's Actually Changing

The role has moved from defensive, compliance-oriented work to offensive value creation and business transformation.

The CDO Panel

Bhagyesh Phanse (CDAO, Humana), Caroline Buckley (CDAIO, Ford), and Chandhu Nair (SVP Stores, Data, AI & Innovation, Lowe's) discussed the shifting CDO role, moderated by Randy Bean.

The CDO Turnover Question

"Expectations are up, patience is dramatically down, and the gap between what leaders are asked to deliver and the timeline it actually takes is the real driver."

Bhagyesh Phanse, Humana

"Data storytelling has evolved to value realization storytelling — telling the story of what changed in the business because of the data."

Chandhu Nair, Lowe's

Caroline Buckley framed high CDO turnover as a positive signal of demand; Chandhu Nair saw it as a natural technology-adoption cycle of centralization, democratization, and recentralization.

Part Four

🤖 Agents, AI, and the Reality Check

Agentic AI is non-deterministic: the next action depends on prior output and can always change. Practitioners argue for keeping the agentic surface area as small as possible.

AI Agents Panel

Radha Kuchibhotla (Lead Director, AI Solutions Design, CVS Health) and Art Morales (VP, Technology Enabled Science, CSL Limited) discussed what counts as an "agent" and the production risk of non-determinism.

"The goal in production should be to minimize the non-determinism of agentic AI and convert as many steps as possible to deterministic processes."

— Art Morales, at CDOIQ 2026

Financial Services Panel

Nachiket Mehta (VP of AI and Data Engineering, The Hartford), Andy Ghosal (VP Product Management, JPMorgan Chase), and Narayanan Nair (CDO, Farm Credit Services of America) discussed AI and data engineering in financial services.

Farm Credit Services of America

Agentic Credit Narratives

~12,000 credit narratives annually, automated with five to six specialized agents pulling from a consolidated enterprise data warehouse.

Writing time cut ~50%; ~20,000-24,000 hours saved/year

The shadow agent problem is real: employees build their own agents, then leave, with no documentation of what those agents do.

Part Five

🕸 Semantics, The Thread Running Through Everything

Semantics and ontologies came up in more sessions, in more ways, than anything else at CDOIQ 2026. Chandhu Nair noted that ontology PhDs are now in high demand after years of being overlooked — a moment Sequeda credits Kimberley Herrington with capturing on camera.

Scar Tissue and 20 Lessons from 20 Years of Building the Foundation AI Actually Needs

Sequeda's own talk, delivered in one-minute lessons: ontologies and knowledge graphs deliver a 3x improvement in LLM query accuracy; governance is what moves AI POCs into production; knowledge is a first-class citizen, not a byproduct of data.

Three Questions Every Data Executive Should Be Asking

🎯 Closing Thoughts

"The leap from solid foundations to actual business transformation hasn't happened yet for most of these organizations, and nobody quite knows how to close it."

Juan Sequeda

Semantics is the differentiator, and the organizational challenge is harder than the technical one. The CDOs who understand this haven't been chasing hype — they've been busy building real foundations.

Additional Takeaways Posts Referenced by Sequeda

Frequently Asked Questions

Fourteen questions about the CDOIQ 2026 conference takeaways

Sequeda characterizes the conference as grounded rather than hype-driven, with data executives emphasizing foundational work over transformation theater.
A five-category star-rating framework (ownership, curation, quality, protection, observability) that Bill Snider's team applies to roughly 100 priority lake houses out of 12,500 schemas.
Caroline Serio's team drove adoption over four years through gamification, community building, an internal Data Summit, and a shift to a data product ownership model.
Leidos CDO Kris Mork's data leadership model: Attention first, Direction second, Energy third, Congruence throughout.
An agentic solution automating credit narrative writing cut analyst time roughly in half, saving an estimated 20,000-24,000 hours per year.
ADP and Capital One built mature data platforms over a decade because technology is native to their organizational DNA, raising the question of whether non-tech enterprises can replicate that path.
That the goal in production should be to minimize the non-determinism of agentic AI and convert as many steps as possible to deterministic processes.
As the thread running through everything, citing Capital One's ontologist titles since 2022, ADP's enterprise knowledge graph, and Lowe's use of AI to parse unstructured documents into taxonomies.
Whether vendors support true data and semantics optionality or create lock-in, whether governance infrastructure supports taking AI POCs to production, and whether incentives reward durable knowledge infrastructure over fast wins.
Foundational work is proceeding steadily, but the leap from solid foundations to actual business transformation hasn't happened yet for most of these organizations.
Bhagyesh Phanse (Humana), Caroline Buckley (Ford), and Chandhu Nair (Lowe's), moderated by Randy Bean.
Caroline Buckley sees it as demand for great talent; Bhagyesh Phanse sees rising expectations outpacing patience; Chandhu Nair sees a natural technology-adoption cycle.
Employees building their own AI agents and then leaving, with no documentation of what those agents do.
"Scar Tissue and 20 Lessons from 20 Years of Building the Foundation AI Actually Needs," covering topics from knowledge graphs improving LLM accuracy 3x to governance being what moves AI POCs into production.

📚 Glossary

Thirteen key terms from the CDOIQ 2026 takeaways

The Chief Data Officer and Information Quality Symposium, now in its 20th year.
Nationwide's five-category rating framework applied to its priority lake houses.
A data architecture combining data lake and warehouse traits; Nationwide governs ~100 as its priority set.
An ungoverned AI agent left behind by a departed employee, with no documentation of its behavior.
Variability in agentic AI outputs across runs; Art Morales argues production should minimize it.
A formal job title, adopted by Capital One in 2022, for ontology practitioners.
A semantic entity-relationship structure; Sequeda cites a 3x LLM query-accuracy improvement from grounding in one.
Accuracy, completeness, and reliability of enterprise data — one of Nationwide's five Trusted Data Score categories.
Organizational policy for managing data as an asset; framed as the real work, not a preliminary step.
High Chief Data Officer departure rates; attributed to a mismatch between rising expectations and shrinking patience.
Kris Mork's data leadership model: Attention, Direction, Energy, Congruence.
A model where business units own their data as a product with real accountability — KeyBank's turning point.
Thinking your organization is what it's not — why non-tech enterprises may struggle to transplant tech-company data DNA.

🕸 Knowledge Graph

Explore the entity relationships in the CDOIQ 2026 takeaways. Click nodes to open entity descriptions via URIBurner.

Mode
Density

SPARQL Workbench

Execute live queries against the companion Turtle knowledge graph hosted on URIBurner (Virtuoso-backed).

Q1 List all named people with job titles and organizations
PREFIX schema: <http://schema.org/>

SELECT ?person ?name ?jobTitle ?org
FROM <https://linkeddata.uriburner.com/DAV/demos/daas/cdoiq-2026-honest-takeaways-claude_code-1.ttl>
WHERE {
  ?person a schema:Person ;
          schema:name ?name .
  OPTIONAL { ?person schema:jobTitle ?jobTitle }
  OPTIONAL { ?person schema:worksFor ?org }
}
ORDER BY ?name
Q2 Retrieve all FAQ question-answer pairs
PREFIX schema: <http://schema.org/>

SELECT ?question ?answer
FROM <https://linkeddata.uriburner.com/DAV/demos/daas/cdoiq-2026-honest-takeaways-claude_code-1.ttl>
WHERE {
  ?q a schema:Question ;
     schema:name ?question ;
     schema:acceptedAnswer/schema:text ?answer .
}
ORDER BY ?question
Q3 Retrieve organizations with descriptions
PREFIX schema: <http://schema.org/>

SELECT ?org ?name ?description
FROM <https://linkeddata.uriburner.com/DAV/demos/daas/cdoiq-2026-honest-takeaways-claude_code-1.ttl>
WHERE {
  ?org a schema:Organization ;
       schema:name ?name ;
       schema:description ?description .
}
ORDER BY ?name
Q4 Retrieve case studies with organizations and outcomes
PREFIX schema: <http://schema.org/>
PREFIX : <https://juansequeda.substack.com/p/cdoiq-2026-my-honest-no-bs-takeaways#>

SELECT ?case ?name ?org ?outcome
FROM <https://linkeddata.uriburner.com/DAV/demos/daas/cdoiq-2026-honest-takeaways-claude_code-1.ttl>
WHERE {
  ?case a :CaseStudy ;
        schema:name ?name ;
        :featuresOrganization ?org .
  OPTIONAL { ?case :hasOutcome ?outcome }
}
ORDER BY ?name
Q5 Entity-type summary — SAMPLE-based canonical recipe
PREFIX rdf:  <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>

SELECT
    ?type
    (SAMPLE(?s)     AS ?sampleEntity)
    (SAMPLE(?label) AS ?sampleLabel)
    (COUNT(?s)      AS ?entityCount)
WHERE {
    GRAPH <https://linkeddata.uriburner.com/DAV/demos/daas/cdoiq-2026-honest-takeaways-claude_code-1.ttl> {
        ?s rdf:type ?type .
        OPTIONAL { ?s rdfs:label ?label }
    }
}
GROUP BY ?type
ORDER BY DESC(?entityCount)

About This Page

This knowledge graph infographic was generated from Juan Sequeda's Substack post "CDOIQ 2026: My Honest, No-BS Takeaways" — the source content was transformed into RDF-Turtle and rendered as this HTML infographic. Full generation provenance (skills, model, server platform) is listed once, in the footer below.