At TPMA, we recently surveyed economic and workforce development organizations on a question that is increasingly difficult to ignore: How prepared are these organizations for an economy increasingly shaped by artificial intelligence? The answers revealed an interesting tension.
AI is no longer a distant consideration for economic and workforce development organizations. It is becoming part of their day-to-day operations and strategic priorities, and, perhaps most importantly, the questions they are receiving from the businesses they serve.
But the survey taught us something else, too. It offered a firsthand reminder that AI isn’t just changing the subjects we study. It is also changing how we conduct research and how we engage with the public. That may be one of the most important lessons organizations should take away from this work.
AI Has Arrived in Economic and Workforce Development
The survey results make one thing clear: AI has rapidly entered the mainstream of economic and workforce development. Eighty-six percent of respondents said AI will be a strategic priority for their organization over the next 12–24 months. Ninety percent reported already using AI, either regularly or occasionally, and 91% reported at least moderate confidence in its use.
Organizations are putting those tools to work in practical ways. Research and data analysis was the most frequently cited use or planned use of AI, cited by 82% of respondents, followed by marketing and content creation at 74%. Administrative tasks, grant writing and proposals, and CRM or business outreach were also common applications.
But perhaps the more consequential finding is what organizations are hearing from the businesses they serve. Nearly four in five respondents—79%—said that businesses are frequently or occasionally reaching out to their organizations for AI-related guidance.
Those questions aren’t limited to “What is AI?” Businesses are seeking help with data management and analysis, workforce impacts and upskilling, tool selection, practical use cases, marketing, ROI, implementation, and other issues. That represents an important shift for economic and workforce development organizations.
AI enablement is becoming part of business assistance.
There is a problem, however. While 79% of respondents reported receiving AI-related inquiries frequently or occasionally, only 45% described their organization as prepared to support businesses in AI adoption. Another 39% said they were only somewhat prepared. In other words, demand for assistance is growing faster than the support infrastructure available to meet it.
The barriers help explain why. Data privacy and compliance were the most commonly cited obstacles, followed by a lack of internal expertise, the rapidly changing AI landscape, insufficient funding and resources, and limited staff capacity. Notably, the problem does not appear to be a lack of interest.
Organizations are seeking practical ways to build capacity. Respondents identified partnership opportunities, staff training, tools and frameworks to support businesses, and funding and resources as their most important needs. For economic and workforce development leaders, this suggests that the next phase of AI adoption should focus less on experimentation and more on institutional readiness.
Organizations need to answer questions such as:
- Where can AI meaningfully improve our own operations?
- What policies and governance structures do we need?
- What should our staff understand about AI?
- What role should our organization play in helping businesses adopt it?
- Which businesses and industries in our region are most vulnerable to disruption—or positioned to benefit?
- What partnerships are necessary to provide expertise we do not have internally?
The organizations that answer those questions early will be better positioned to help their regions navigate the economic transition ahead.
Then AI Showed Up in the Survey Itself
Another lesson emerged from the research process. During TPMA’s data-quality review, we removed nearly 43% of responses based on multiple criteria, including unusually fast completion times, suspected spam emails or responses, and respondents identifying organizations unrelated to economic development, workforce, or training. Just over 57% of total responses are included in the final analytical dataset.
Those challenges are familiar to anyone who conducts online research. Bots, duplicate responses, professional survey takers, and low-quality submissions have been around for years. But generative AI raises the stakes considerably.
A fraudulent respondent no longer needs to submit obvious nonsense. AI can generate convincing organization names, plausible job titles, professional-sounding open-ended responses, and seemingly thoughtful answers almost instantly. That means one of the assumptions underlying online public engagement is becoming increasingly fragile:
A completed survey response is not necessarily evidence that a person meaningfully participated.
Public Surveys Have a New Vulnerability
This matters far beyond a single AI survey. Communities increasingly rely on online surveys to develop comprehensive plans, CEDS documents, workforce strategies, housing studies, transportation plans, economic development strategies, and other public initiatives. Those surveys can influence priorities involving millions of dollars in public investment. If responses are manipulated or simply overwhelmed by automated or low-quality submissions, the resulting findings can create a false picture of community sentiment.
Imagine a community survey receiving 2,000 responses. That sounds like extraordinary engagement. But what if 400 responses are automated? What if another 300 originate outside the community? What if coordinated respondents complete the survey repeatedly?
The question isn’t simply: How many people responded?
Increasingly, organizations must ask:
How confident are we that the responses represent genuine human participation?
That is a fundamentally different standard.
The Answer Isn’t Abandoning Surveys
Surveys remain an extremely useful engagement tool. But they should rarely stand alone. Organizations need to become more sophisticated in how they design, distribute, monitor, validate, and interpret them.
At TPMA, our experience reinforces several principles.
Design engagement around people, not platforms. A survey is a tool, not an engagement strategy. Strong public engagement combines surveys with interviews, focus groups, workshops, stakeholder conversations, community meetings, and other opportunities to hear directly from people.
Build data-quality safeguards into the process from the start. Organizations should consider response timing, duplicate patterns, geography, email validity, organizational affiliation, unusual response clusters, open-ended response patterns, and other indicators when assessing survey integrity.
Monitor responses while the survey is live. Waiting until the survey closes to examine data quality can allow problems to compound.
Triangulate findings. If survey respondents identify housing as the region’s greatest economic challenge, do employer interviews, demographic data, housing market analysis, and stakeholder conversations point in the same direction? When multiple sources converge, confidence increases. When they don’t, that discrepancy is itself a finding worth investigating.
Be transparent about methodology. Removing questionable responses isn’t hiding data. Done carefully, consistently, and transparently, it is protecting the integrity of the research.
Human-Centered Engagement Matters More in an AI-Enabled World
Ironically, the growth of artificial intelligence may make human-centered engagement more valuable, not less.
Technology can dramatically improve research. AI can help analyze large datasets, identify themes, accelerate literature reviews, summarize qualitative information, and uncover patterns that would previously have required enormous amounts of staff time.
But technology does not eliminate the need to understand people.
- A business owner explaining why she cannot find workers.
- A manufacturer describing why an expansion project stalled.
- A young professional explaining why he plans to leave the region.
- A workforce provider describing the barriers preventing participants from completing training.
Those conversations provide context that a survey response alone often cannot.
TPMA’s approach to engagement increasingly reflects that reality: use technology to expand our analytical capacity while designing research processes that keep people at the center. That means combining quantitative and qualitative research, validating findings across multiple sources, building safeguards against unreliable data, and recognizing when the methodology itself may be shaping the conclusions.
Two Lessons From One Survey
Our AI Enablement Survey ultimately reinforced two important lessons.
The first is about economic development.
AI adoption is moving faster than the capacity of many economic and workforce development organizations to support it. Businesses are already asking questions, and organizations have an opportunity to become trusted guides—but doing so will require new expertise, partnerships, tools, training, and governance.
The second lesson is about engagement.
In an era when AI can imitate human participation, collecting responses is no longer enough. We have to design engagement systems that can establish confidence in who—and what—we are hearing.
For organizations responsible for making decisions based on public input, that distinction will only become more important. The future of engagement isn’t choosing between technology and people. It’s using technology intelligently while becoming even more intentional about keeping the work human-centered, evidence-based, and trustworthy.