America wants more ships, more chips, more advanced manufacturing, and a workforce ready for AI. Who is responsible for preparing the people who will do that work?

We call it the workforce system. But employers, colleges, training providers, workforce boards, and government agencies each own a piece. Who connects them? Who makes sure a training program leads to a job? And who helps workers navigate the distance between the two?

Paul Fain, editor at Workshift and writer of The Job newsletter, joins Ian Davidson to examine what the workforce system actually is, how it is funded, and why so much responsibility falls to states and regions. Paul reports across the boundaries between education, workforce policy, and employer demand. That perspective helps reveal what happens when institutions pursue related goals without the coordination to make them work together.

The conversation starts with the Workforce Innovation and Opportunity Act, or WIOA, the main federal workforce framework. Paul describes a system smaller than many people realize, with funding divided among programs and administered through states and local workforce boards. The question quickly becomes larger: if employers need qualified workers, who should pay to prepare them, and what should employers commit to in return?

AI makes the coordination problem more urgent. Near the end, Paul describes the possibility of labor-market disruption on the scale of industrialization. Ian's conclusion: no one is coming to assemble this for us. States, employers, educators, and workforce organizations will have to build the capacity to respond together.

This is the second episode in Act 3, our continuing investigation into why the workforce system does not function like a system. Next, Human Cloud co-founders Matthew Mottola and Tony Buffum help us examine work spreading across projects, clients, and platforms.

EPISODE CHAPTERS

00:00 - A workforce system built around figuring it out alone

03:00 - What is the workforce system?

06:00 - How workforce funding reaches people

09:00 - Who should pay to prepare workers?

13:00 - Connecting education to jobs that pay off 15:00 - Oklahoma and employer-funded training

]18:00 - What coordinated infrastructure requires

20:00 - Industrial policy and local responsibility

21:00 - Why states are being asked to do more

23:00 - State talent marketplaces and connected data

28:00 - National ambitions meet regional economies

34:00 - AI disruption and support beyond tuition

38:00 - No one is coming to assemble this for us

Read Paul's reporting: https://workshift.org/ Connect with Paul: https://www.linkedin.com/in/paulfain-thejob/ Explore Signal Mapping: https://signollabs.com/blog-signal-mapping/ Learn about Signol Labs: https://signollabs.com/ Working to connect training and employer demand? Contact Ian at ian@signollabs.com or https://signollabs.com/contact-us/. Subscribe to The Signal Lab and share this conversation with someone building your region's workforce.

How do we prepare workers for disruption if we cannot see how work is changing?

Artificial intelligence is already changing how employers operate, organize work, and value different capabilities. Yet much of the workforce ecosystem still relies on job postings to understand employer demand. Those postings can be useful, but they are advertisements. They do not reliably tell us which opportunities are real, who was hired, what work that person performs, what the job pays, or whether it lasts.

Economist Julia Lane has spent decades working on this problem. She is a professor at New York University, and author of Democratizing Our Data.

Julia explains why technological change reaches workers through employer decisions. Companies choose whether to adopt AI, how to redesign their operations, and which workers and capabilities they need. Understanding the future of work therefore requires better information about what employers actually do - not only what they advertise.

This is a conversation about data, but its consequences are deeply human. Bad information wastes employer resources, directs training dollars toward uncertain outcomes, and leaves workers making consequential choices without a clear view of opportunity. Better information could change the narrative from inevitable disruption to informed adaptation.

EPISODE CHAPTERS

00:00 - The Planes That Never Returned: How selection bias distorts what we think we know.

01:10 - Why We Cannot Prepare for Change We Cannot See: The limits of job-posting data.

05:00 - Learning at Work: Why on-the-job training matters across a working life.

06:40 - Labor Supply, Employer Demand, and Technology: Where economic disruption enters the workforce.

08:20 - The Missing Longitudinal View: Why surveys struggle to capture labor-market churn.

10:35 - The Employment Data We Already Collect: What unemployment-insurance wage records reveal.

13:10 - Data Built for Yesterday's Questions: Moving from surveys to administrative records.

14:15 - Industries of Ideas: Tracking how technology shocks move through local economies.

16:10 - What Early AI Adoption May Be Showing: Initial findings from Ohio employers.

17:10 - What JEDx Could Add: Better work locations and faster reporting.

18:25 - What Better Data Means for Learners: Local jobs, earnings growth, and durable pathways.

20:00 - Why Job Postings Are Not Enough: Representation, documentation, and validation.

21:35 - Building the Missing Feedback Loop: Connecting training to actual worker outcomes.

23:05 - A More Optimistic Future of Work: How better signals could help employers and workers adapt.

Learn more about Julia Lane and Industries of Ideas: https://industriesofideas.ai/ Connect with Julia: https://www.linkedin.com/in/julia-lane/ Explore Signal Mapping: https://signollabs.com/blog-signal-mapping/ Learn more about Signol Labs: https://signollabs.com/ Connect with Ian Davidson: https://www.linkedin.com/in/ianwdavidson/ Recommend a future guest: connect@signollabs.com

What do employers actually want?

What skills matter? What credentials will they trust? Can better data improve hiring? After three months investigating how hiring decisions actually get made, Ian Davidson thinks we're starting with the wrong question.

The better question is: Which decision are you trying to change?

Hiring isn't one decision. Job boards decide which opportunities people see. Applicant tracking systems decide who gets surfaced. Recruiters decide who gets a conversation. Hiring managers decide who gets the job.

Together, those decisions allocate one of hiring's scarcest resources: human attention.

And the decisions affecting the most candidates often happen when the system knows the least about them.

Episode 10 brings together insights from Allison Danielsen, Matt Plummer, Andrew Cunsolo, Jordan Franklin, Matt Charney, Brad Robinson, and James Harris.

The investigation began as a search for better signals. It became the Hiring Decision Architecture: a framework for identifying which hiring decision we're trying to improve, what information could make it better, and where to intervene.

EPISODE CHAPTERS

00:00 — What Do Employers Actually Want? Why skills and credentials may be the wrong starting point.

05:00 — What Hiring Managers Really Evaluate Motivation, potential, work ethic, learning, and what resumes miss.

12:00 — Skills-Based Hiring and the Replacement Problem What replaces degree and experience requirements?

14:00 — Application Overload, ATSs and AI Hiring What happens when employers receive 1,000 applications?

16:00 — Hiring Is an Allocation of Human Attention Why consequential decisions happen with limited candidate information.

18:00 — The “Cake Lady” and the Candidates Systems Miss A great hire who never would have survived the conventional process.

21:00 — Better Candidate and Opportunity Profiles Why resumes and job descriptions both leave out critical context.

26:00 — The Hiring Decision Architecture Start with the decision, then identify the information needed to improve it.

28:00 — Talent Marketplaces and Workforce Intelligence What if matching systems understood people's needs and constraints?

33:00 — Fix the System Before the Hiring Manager Improve the decisions determining who reaches a human.

37:00 — AI, Machine-Readable Data and Verifiable Skills Technology can use richer information without overwhelming recruiters.

40:00 — Better Talent Marketplace Matching Matching capability, potential, motivation, needs, and opportunity.

42:00 — Learning and Employment Records + Trusted Career Profile How portable, verifiable data could improve hiring decisions.

45:00 — The Demand-Side Case for Better Workforce Data Employers aren't asking for LERs. They're asking for help with 1,200 applicants.

48:00 — Which Hiring Decisions Should We Fix? Stop starting with data. Start with the decision.

Research supported by the U.S. Chamber of Commerce Foundation and T3 Innovation Network.

Explore the Hiring Decision Architecture: https://signollabs.com/hiring-decision-architecture/

Signol Labs: https://signollabs.com/

Subscribe to The Signal Lab wherever you listen to podcasts.