Data Faces Podcast

Leigh Felton, President of the AI for Job Security Foundation, joins Data Faces on location at the 20th annual CDOIQ Symposium in Cambridge, Massachusetts. Had AI been screening job candidates when she started out, she says it would have rejected her. She was a pregnant, homeless teenager and a college dropout who went on to spend 20 years at Microsoft, where she headed the company's first ethical AI enablement program.Leigh shows her two-volume book series, I Could Have Been CEO Yesterday, and explains why she believes you cannot de-bias a system that is built to recognize historical patterns. She makes the case for governing AI as an environment that shapes how people think rather than as a tool you pick up and put down. She also challenges the assumption that faster AI output equals productivity, pointing to research showing AI has reduced productivity for some engineers by about 19 percent.What you will learn:1. Why de-biasing an AI system trained on historical patterns may be impossible2. How environmental-scale governance, in the spirit of OSHA, differs from governing AI as a tool3. Why collaborating with AI and pushing back on it beats publishing its first draft, even when the task takes longer4. Why junior employees and Gen Zers will question the things 30-year experts stopped questioningChapters:0:00 Welcome and the presidential icebreaker1:06 Inside the AI for Job Security Foundation1:36 Two books on AI bias and who gets hired4:48 AI as an environment6:55 Speaking the CDO's language on productivity10:09 Don't let AI genericize you13:30 Will AI ever be smarter than people?15:45 Where to find Leigh's booksWatch more Data Faces on location: https://tinytechguides.com/data-faces-podcast/?utm_source=youtube&utm_medium=video&utm_campaign=cdoiq2026-leigh-felton&utm_content=descriptionConnect with Leigh Felton on LinkedIn: https://www.linkedin.com/in/leighfelton/#DataFacesPodcast #CDOIQ #AIEthics #FutureOfWork

Stacie Christensen, Senior Data Leader at H-E-B, joins Data Faces on location at the 20th annual CDOIQ Symposium in Cambridge, Massachusetts. After 22 years in the data industry, she grades our collective data quality somewhere between a D plus and a C minus, and she argues that data initiatives fail long before technology is ever the problem.

Stacie explains why AI is a multiplier and a scaler that has nothing positive to scale when the foundation underneath it is missing. She makes the case that data governance belongs at data creation, since bad data becomes exponentially more expensive the further down the pipeline it travels. She also introduces her book series, including Built on Data and the quietly released Built on Simplicity, and shares why circular data flows are her favorite red flag for hidden complexity.

What you will learn:

1. Why the foundation under the technology, including accountability, ownership, and governance, decides whether data initiatives succeed

2. Why hoarding data creates noise, backlog, and risk without helping critical business decisions

3. Why governing data at creation beats governing it downstream in analytics and reporting

4. How to spot circular data flows and use them as a starting point for simplification

Chapters:

0:00 Welcome from day three at CDOIQ

0:50 The hardest class on campus

1:58 Why data initiatives fail before technology

3:35 AI is a multiplier and a scaler

5:16 Grading industry data quality

8:49 Governance starts at data creation

11:07 Built on Simplicity and circular data flows

14:26 Where to find Stacie's books

Watch more Data Faces on location: https://tinytechast/?utm_source=youtube&utm_medium=video&utm_campaign=cdoiq2026-stacie-christensen&utm_content=description

Connect with Stacie Christensen on LinkedIn: https:iechristensen/

Stacie's website: https://dataisnotpersonal.com

#DataFacesPodcast #CDOIQ #DataGovernance #DataQuality

Most of what makes an expert good at the job was never written down. Ken Sanford, who leads go-to-market and commercial strategy at Clarifeye, argues that this tribal knowledge is the part of the AI context layer nobody talks about, because document retrieval is close to solved and the judgment to interpret those documents still lives in people's heads.

Ken and David Sweenor go back to their SAS days. They get into how Clarifeye's AI interviewer, Clara, pulls expertise out of people by hunting for edge cases and exceptions, and why the most tech-forward people in a company are the wrong ones to write its AI skills. Ken also explains why the memories your AI harness builds amount to your judgment, why your company should own them, and where he thinks the next generation of experts comes from.

What you will learn:
1. Why near-perfect document retrieval still leaves AI without the judgment to interpret what it finds.

2. How interviewing experts about exceptions, five or ten minutes a day, captures knowledge a wiki never will.

3. Why the people with 25 or 30 years of intuition should shape your AI skills, and how to make that capture a pull.

4. Why owning the artifacts your AI harness builds lets a company move between models and tools.

Chapters:

0:00 Intro

1:17 Ken Sanford and Clarifeye

2:29 A surfboard at the trade show booth

3:42 Iron Man 3 extras and a first job in Naples

5:58 Tribal knowledge and the unstructured data problem

8:20 How Clara runs an expert interview

10:51 Kids, chatbots, and sycophantic AI

12:02 Do experts resist? Harness memories as your judgment

13:29 Where customers start: onboarding and software migrations

15:20 Data, context, and the reasoning layer

16:17 Keeping knowledge current, and banning the word "ship"

18:58 Would leaders use AI to replace themselves?

19:50 The wrong people are writing your AI skills

22:16 Where the next generation of experts comes from

24:41 Hand integration and the skills we lose26:16 Hiring for agency27:32 "Just quit your company" and a team of 12

29:22 Why process blueprints beat a vector store

30:46 Build your own career knowledge base

31:40 End of days? A techno-optimist view

34:58 West of Jesus and the science of flow

37:05 Where to find Ken

Read the full article: https://tinytechguides.com/blog/data-faces-ken-sanford-ep53-tribal-knowledge-ai/?utm_source=youtube&utm_medium=video&utm_campaign=ep53-ken-sanford&utm_content=description

Connect with Ken Sanford: https://www.linkedin.com/in/ekenomics/Clarifeye: https://www.clarifeye.ai/Connect with David Sweenor: https://www.linkedin.com/in/davidsweenor/

#DataFacesPodcast #TribalKnowledge #AISkills

Randy Bean, founder and CEO of the Data & AI Leadership Exchange, joins Data Faces on location at the 20th annual CDOIQ Symposium in Cambridge, Massachusetts. His test for data leaders is blunt. If you're not getting measurable business value from your data and AI investments, or a clear path to it, go back to the office this afternoon and shut them down.Randy has organized the CDOIQ Chief Data Officer panel for 12 years, and this year's edition featured data leaders from Lowe's, Ford, and Humana. He traces how the role has moved from defensive risk and compliance work to an offensive, business-focused mandate, shares fresh survey numbers on where CDOs report, and explains why every chief data officer from last year's panel will be out of that role by the end of this year. He also gives his mid-year read on the AI bubble.What you will learn:1. How the CDO role moved from risk and compliance to a business-value mandate2. Where CDOs report today, with 42% under technology and 33% under business leadership3. Why churn at the top of the panel signals a fight to control AI's future in the enterprise4. Randy's bubble call, AI is overestimated in the short term and underestimated in the long termChapters:0:00 Welcome back to Data Faces0:38 Poet or rock star1:55 Twelve years of CDO panels3:10 Shut down what doesn't deliver value7:26 Where should the CDO report10:44 A role changing in real time11:55 The AI bubble, short term and long term14:24 Sign-off from CambridgeWatch more Data Faces on location: https://tinytechguides.com/data-faces-podcast/?utm_source=youtube&utm_medium=video&utm_campaign=cdoiq2026-randy-bean&utm_content=descriptionConnect with Randy Bean on LinkedIn: https://www.linkedin.com/in/randybeannvp/#DataFacesPodcast #CDOIQ #ChiefDataOfficer #DataLeadership

Everyone blames AI for the tech job market. An executive recruiter who talks to the hiring decision-makers every day says the story started two years before the AI headlines did.Jim Jinright, Managing Partner at JRJ Search Group in Dallas Fort Worth, joins David Sweenor to explain the perfect storm behind the layoffs: pandemic overhiring, the 2022 interest rate increases, and labor arbitrage, with AI arriving last and taking the blame. He describes the move from a no-hire, no-fire economy to a low-hire, no-fire one, and why his phone is ringing more than it was a year ago.We also get into where the junior work went, what a recruiter does when 500 applications arrive in an hour, ghost jobs, age screens, why Jim calls the AI interview a slap in the face, and his advice for early-career engineers who are watching coding become a commodity.Key takeaways:1. The tech job market slowdown traces to overhiring, interest rates, and cost cutting. AI added the "10 engineers or none" idea, and that idea is starting to correct.2. AI is powerful and has no context. Senior people with judgment are more in demand, and junior people lost the small problems they used to learn on.3. Every resume is perfect now, so the conversation decides. Bring the receipts: problems you owned and outcomes you produced.4. Ghost jobs are real. If you are not getting responses, it is not always you. Sometimes it is the moment.5. Move up the stack. Stop competing on a language and become the person who builds and runs the software factory.Chapters:0:00 Intro1:11 JRJ Search Group and twelve years in executive search2:16 A would-be pilot and a demo tape3:33 The perfect storm behind the layoffs5:45 From no-hire, no-fire to low-hire, no-fire6:16 Where the junior work went8:23 Every resume is perfect now9:04 500 applicants in an hour11:19 Ghost jobs13:35 Surviving a layoff and staying sane16:03 Age discrimination and showing the last ten years18:24 What "AI literate" really means to clients21:31 Demonstrating emotional intelligence24:43 Fractional work26:31 The AI interview28:44 Job hopping and the resume that explains itself32:57 Advice for early-career engineers36:57 What Jim is watching38:11 Where to find JimRead the full article: https://tinytechguides.com/blog/data-faces-jim-jinright-ep51-perfect-storm-tech-job-market/?utm_source=youtube&utm_medium=video&utm_campaign=ep51-jim-jinright&utm_content=descriptionConnect with Jim Jinright: https://www.linkedin.com/in/jimjinright/JRJ Search Group: https://jrjtx.comConnect with David Sweenor: https://www.linkedin.com/in/davidsweenor/#DataFacesPodcast #AIJobs #TechCareers

Veenit Shah, Senior Manager of Data Engineering, and Puneet Singh, Senior Data Engineer, both at Intuit Credit Karma, join Data Faces on location at the 20th annual CDOIQ Symposium in Cambridge, Massachusetts. Their team keeps credit scores current for about 140 million members, and an AI remediation agent now compresses root cause investigations that once took 30 to 40 minutes into just a few.Veenit and Puneet walk through the data quality program they presented at the conference, from startup-era firefighting with nightly calls and manual checks to five pillars monitored across more than 100 tables and 40,000 columns. They explain why data quality is everybody's job when a single alert can touch millions of members, how agents now generate production-ready pipeline code in minutes, and why the next levels of their agentic maturity model remain under security review.What you will learn:1. How Credit Karma moved from reactive firefighting to five data quality pillars measured across 40,000 columns2. Why every alert gets logged, measured, and analyzed weekly, with brutal prioritization of the noisiest issues3. How an AI remediation agent runs and refines queries until it lands on root cause and blast radius4. Where guardrails, anonymized data, and selective agent access fit as the team climbs the maturity curveChapters:0:00 Welcome and introductions2:21 The firefighting years4:58 Five pillars across 40,000 columns6:48 Whose job is data quality?8:14 The AI remediation agent10:10 Level two on the agentic maturity curve12:22 Guardrails and security for agents14:11 Wrap-up with Veenit and PuneetWatch more Data Faces on location: https://tinytechguides.com/data-faces-podcast/?utm_source=youtube&utm_medium=video&utm_campaign=cdoiq2026-intuit-credit-karma&utm_content=descriptionConnect with Veenit Shah on LinkedIn: https://www.linkedin.com/in/veenitshah/Connect with Puneet Singh on LinkedIn: https://www.linkedin.com/in/puneet-singh-5614b065/Intuit Credit Karma: https://www.creditkarma.com#DataFacesPodcast #CDOIQ #DataQuality #AIAgents

Every vendor says your data needs to be "AI-ready." Almost none of them define it the same way.Malcolm Hawker, Chief Data Officer at Profisee, former Gartner analyst, and author of The Data Hero Playbook, joins David Sweenor to take the term apart. His definition is short: data is AI-ready when it supports the use case in front of it. That makes readiness a spectrum, not a state, and it means the same customer data can clear the bar for a marketing campaign and fall far short for a regulated decision.We also get into why a semantic layer cannot solve identity, why "garbage in, garbage out" is a career-limiting thing to say to your CEO, and how the industry manufactures new names for problems that already had them.Key takeaways:1. Data is AI-ready when it fits the purpose. There is no universal threshold, and treating readiness as binary is why the term is ambiguous and nearly meaningless.2. The cost of being wrong, not a quality score, decides whether a use case can go into production.3. Semantic layers define what a customer means. They cannot tell you which of fifteen customer records is the real person.4. Most failed proofs of concept put a probabilistic system into a process that always ran on deterministic rules.5. New vocabulary in data management is often an old idea with a new label, and the loop that produces it is predictable.Chapters:0:00 Intro1:18 From Dun & Bradstreet to Gartner to Profisee4:22 The record label Malcolm never started7:02 What does AI-ready data even mean?7:46 Fitness for purpose, and why ChatGPT works anyway11:56 Context, semantic layers, and the fifteen David Sweenors14:51 Why "garbage in, garbage out" gives Malcolm hives15:44 Data quality for unstructured data19:26 The semantic pedantic feedback loop23:30 How data literacy became a top-three problem overnight25:17 Does MDM apply to unstructured data?29:54 The Data Hero Playbook and the growth mindset34:43 What Malcolm is reading37:13 Where to find MalcolmRead the full article: https://tinytechguides.com/blog/data-faces-malcolm-hawker-ep49-ai-ready-data/?utm_source=youtube&utm_medium=video&utm_campaign=ep49-malcolm-hawker&utm_content=descriptionMentioned in this episode:Bridging Knowledge, Data, and AI by Joseph Hilger, Lulit Tesfaye, and Zachary WahlSoftware Wasteland by Dave McCombConnect with Malcolm Hawker: https://www.linkedin.com/in/malhawker/Connect with David Sweenor: https://www.linkedin.com/in/davidsweenor/#DataFacesPodcast #MasterDataManagement #AIReadyData