From the levee by my home, I can see global containerships move down the Mississippi River. These ships can carry 10 stories of cargo. They are long-lasting, older than I am, and they can’t turn easily. By their side, a small tugboat nudges them around bends too sharp for them to maneuver alone. Tugboats have the power of 40 racecars. Their teams know their waterways and landscapes intimately. They are first responders and the everyday engine of their ecosystem.
As large public systems, workforce regions, and institutions endeavor to navigate labor gaps and a future shaped by AI, we can take lessons from the tanker and the tugboat.
The rotational inertia of our large systems, that resistance to changing direction, makes realigning the education-to-career pipeline exceptionally difficult, especially when the pipelines are shifting at unprecedented speeds. Meanwhile, a sense of urgency dominates the atmosphere: the belief that you won’t last as an organization without being “AI-native.” Expectations are accelerating as Workforce Pell data hurdles emerge, talent marketplaces become top priorities, and the National Science Foundation announces a $234M grant for AI Readiness.
Many of these transitions require connecting people to opportunities at scale, which often means new technology and digital products. So it is reasonable to worry that surviving the shifts ahead will require ripping and replacing legacy enterprise software and processes. And if organizations are paying enterprise prices nowadays, the software should fit like a wedding dress. The need for expensive tools that force the work to contort around them is gone.
The message to the worker is: adapt or lose. But, for public systems and legacy institutions, standing up entire development teams or replacing the hull is expensive, risky, and often too slow for the moment we are in. If our tankers can’t turn amidst massive labor shifts, the gap between people and good jobs will stubbornly widen.
So, what’s the tugboat?
Subject Matter Experts Enter the Codebase
AI has reached a new level of coding quality. That means that subject matter experts, with their “local waterway” knowledge of spreadsheets, relationships, and institutional memory, can surprise us all with the precision of what they build and solve. Any tool built to make the institution more nimble applies. Perhaps it’s triaging connections between worker and local opportunity, or an unsexy, automated crosswalk between old and new taxonomies. It might be an AI front end like the growing fleet of career chatbots or the back end like a cheaper jobs ingestion pipeline.
At WhereWeGo, a workforce technology company I co-founded, the timelines for our lab builds are in days, not months. Yet the market is still oriented towards the complex, million-dollar software transformation. RFPs come with exhaustive requirements, timelines stretch over years, prototypes are circulated for weeks, and funding is focused on the big launch and demo days.
But what the workforce system most desperately needs isn’t more slow builds. We need the core capacity to continuously repair, upgrade, and integrate light ones. Fortunately, we can now do that by leveraging AI-assisted development and training the people closest to the problem to define solutions that will have the most immediate impact.
Choose the First Useful Turn
Where to start these lightweight repairs? Starting gets easier when we look at the repeatable builds workforce systems need again and again. It is likely one of these five:
- Helping people understand what is available
- Helping people decide whether an opportunity fits
- Helping people take a next step
- Helping staff and partners keep information accurate
- Helping the system learn from real behavior
And then, locate the subject matter expert closest to the problem. Examples throughout the career navigation industry abound:
- A state’s education department knows a common application for dual credit courses will increase enrollment, but their systems are incompatible. Who knows best just how inflexible they are?
- A sector partnership needs scalable alignment between curriculum and job postings. Who is the convener that knows both the curriculum designers and the hiring managers?
- A governor’s office is stitching together incompatible datasets. Who is the champion that has worked with them all?
Across each of these, top-down mandates, lengthy RFP processes, or lengthy builds are likely not the best approach. The best is combining a subject matter expert, AI-assisted coding, and simple boundaries that keep the solution secure and useful. That is how teams can deploy something workable in weeks. Workforce development can start treating this as a core capacity.
A Process for Doing No Harm
AI has made building faster, but stewardship is no less critical. Otherwise, workforce tools depreciate quickly, with an average life expectancy of under three years. When we gave ourselves a charge of one free and public tool every 30 days (as long as someone’s asked for it), we knew that we’d continue to tend some and let others die. That’s okay.
In the career navigation industry, a “viable” product or MVP might just mean it’s sellable to a workforce board or is called “useful.” But, a workforce tool has a different mission and risk profile. It has to avoid a wasteful cost-per-outcome and it relies on the time, money, and hope of a worker or learner, which are increasingly one and the same.
When a job board has 40% fake job listings, dead links, or biased chat interactions, that’s manufactured harm. A worker is discouraged and the labor gap widens. We must commit to creating tools that don’t create additional harm through dead ends, confusion, inaccuracy, bias, or neglect. We propose that those of us in workforce development designing technical solutions adopt a standard of what we call the “minimum ethical product.” The MEP asks: “what valuable tool can we responsibly ask those we serve to rely on?”
Subject matter experts—especially those who interact with workers and learners—managing the day-to-day work are best positioned to answer that question. Our workforce system has exactly who it needs. We are ready to move beyond extra-large projects to see extra-large impact.
If we want practical solutions that can help large systems make nimble turns, we need to give those experts the authority and opportunity to act.
We need tugboats to turn this tanker.