AI Gives Radio Time Back

How it can strengthen a programmer’s weaker skills and help talented people do better work

By Joe “Crash” Kelley | SonicAttention.com

AI can help a broadcaster become better at parts of the job that have never come naturally. That possibility deserves as much attention as the hours it saves.

A programmer with a terrific ear for music may struggle to turn a spreadsheet into a useful programming decision. A gifted air personality may have trouble translating a strong on-air idea into a visual campaign. Someone who understands an advertiser’s business may still need help developing the creative treatment that makes the message memorable.

Those are opportunities for professional growth. If AI helps people work through them, the station can gain a more capable programmer and the audience can get a better product. Whether that happens depends on what management does with the capacity those people gain.

The question Mike McVay helped bring into focus

Mike McVay’s October 20 Radio Ink column, “The Fearless Talent: Mastering AI for Stellar Performances,” makes a persuasive case for learning the tools, strengthening personal brands, reaching audiences across platforms, and protecting credibility. His closing advice is worth keeping close: “Be fearless and always learning.”

McVay has made an extraordinary contribution to radio. His decades of programming, consulting, executive leadership, and talent coaching have helped shape the industry. His commitment to compelling personalities and meaningful audience relationships has enduring value. He has earned the respect of broadcasters who care about the craft.

His column also raises a question that belongs on management’s desk: when a talented broadcaster becomes more capable, how will the company help that person turn the improvement into something listeners and advertisers can hear?

I encounter that question while handling programming, operations, and production responsibilities for five stations in Southwest Georgia. Country, classic rock, Hot AC, Urban AC, and news/talk each require a different approach. The music, personalities, promotions, and community relationships all deserve attention.

After four decades in radio, I am enthusiastic about what these tools make possible. I also know how quickly a recovered hour can disappear into the next assignment.

A strong broadcaster can still have an uneven skill set

Most of us entered radio through something we loved and learned the rest as the job expanded. We might have started with music, performance, production, or engineering. That starting point still influences what we do confidently and what we put off.

Being an excellent programmer does not mean being equally accomplished at data analysis, graphic design, video editing, sales presentations, and talent development. Admitting that is useful. It gives us somewhere to improve.

I see the opportunity in my own creative work. At SonicAttention, I write the melodies and lyrics, develop the concept, and decide what an advertiser’s message should sound like. Modern production tools help me bring those ideas to life with a range of voices, instrumentation, and arrangements that would once have required resources beyond a small production operation’s reach.

The creative decisions remain mine. I reject the beautiful vocal that sells the wrong emotion. I change the lyric that buries the client’s name. I rebuild the ending and edit the spot until it fits the available time. The technology expands what I can execute; experience tells me whether the result is worth using.

That same relationship can help a programmer approach work that once felt outside their abilities.

What the research actually suggests

There is evidence for this beyond our own enthusiasm, although the studies below examined other kinds of work. They do not establish a ratings increase, a revenue gain, or a guaranteed result for radio.

In a study published in The Quarterly Journal of Economics in 2025, Erik Brynjolfsson, Danielle Li, and Lindsey Raymond examined an AI assistant used by 5,172 customer-support agents. Productivity, measured as issues resolved per hour, increased by about 15 percent on average. Less experienced and lower-skilled workers gained more, improving both speed and quality. The most experienced, highest-skilled workers saw small speed gains and small declines in quality. The benefit depended on the person and the task. Read the study.

A 2025 working paper, The Cybernetic Teammate, examined product-development work at Procter & Gamble. Without AI, commercial specialists tended toward commercial proposals and technical specialists toward technical proposals. With AI, their ideas incorporated a more balanced mix of both perspectives. That is especially relevant to a programmer who understands the audience but needs help thinking through an advertiser’s business objective, or a sales-focused colleague who needs a stronger creative starting point. Read Harvard’s account of the research.

There is also a useful creative caution. A 2024 Science Advances experiment found that AI-supplied ideas improved evaluations of short stories, particularly for writers who initially scored lower on a creativity measure. But the resulting stories became more similar to one another. For radio, I take that as a reason to use assistance thoughtfully and insist on local detail, distinctive voices, and an original point of view. Read the paper and abstract.

My takeaway is that AI may help us work more effectively beyond our strongest specialty. We still need to recognize where it is helping, where it is flattening the work, and where another person’s expertise is needed.

Five places a programmer can put that idea to work

These are practical applications I would test, rather than outcomes established by those studies.

1. Making better use of music and audience data. A programmer with strong musical instincts may be less comfortable building spreadsheet comparisons. With data the station is permitted to use in an approved tool, AI can help organize spin histories, compare category distribution, flag apparent repetition, and explain the calculations. Ask it to distinguish an observation from a possible explanation. Verify the numbers against the original export before changing a clock or rotation. Use the analysis to ask sharper questions, especially when the data is too limited to establish a trend.

2. Turning an idea into a clearer creative brief. A programmer may know a promotion needs more energy without knowing how to describe the desired music, pacing, or visual treatment. AI can help translate that instinct into a brief a producer, designer, or salesperson can use. Start with the audience, the local reason to care, the advertiser’s goal, and the station’s personality. Explore several approaches, then supply the judgment and specificity that make one worth developing.

3. Preparing more useful talent coaching. Some programmers hear a problem immediately but struggle to explain it constructively. Using an authorized aircheck and a checked transcript, AI can help organize examples, locate a buried premise, or suggest a focused practice exercise. Listen to the actual audio before judging timing, warmth, humor, or chemistry; a transcript leaves much of a performance out. The programmer should deliver the feedback and understand the person receiving it.

4. Carrying the station’s identity into digital work. An experienced broadcaster may have little formal training in visual design or web writing. AI can help develop a storyboard, prepare a first layout, adapt a promotion for a short video, or identify where a web page is hard to follow. Keep the station’s real logo, verified event details, and brand voice. Check the result on a phone and test the links. A polished-looking mockup still needs to work when a listener uses it.

5. Connecting creative work to a business result. A programmer who speaks comfortably about imaging may be less confident explaining a campaign to an advertiser. AI can help organize a proposal around the customer’s problem, the offer, the creative idea, and a sensible way to measure response. Rehearse the questions a skeptical owner might ask. Then work with sales to check the assumptions, pricing, and promises. The goal is a better conversation about the client’s business.

In each case, the broadcaster brings something essential: knowledge of the audience, the market, the station, and the people involved. Assistance is most useful when that knowledge directs the work.

Ask for an explanation you can use again

Finishing a task and developing a skill are related, but they are not identical. A programmer can receive a useful answer without understanding how it was produced.

I would make explanation part of the assignment. Ask why a proposed approach fits the objective, what assumptions it depends on, and how to check the result. Request an alternative and compare the tradeoffs. When the task involves numbers, ask for a calculation you can reproduce. When it involves creative work, ask what makes the idea specific to this station or advertiser.

That gives you something to carry into the next project. You become better equipped to brief the work, recognize a weak answer, and decide when to call a specialist.

It also keeps a useful boundary in place. Agreement among several AI systems is not independent verification. Local facts need reliable sources. Replicated voices need the performer’s permission and clear terms. The person approving the finished work remains accountable for it.

Give the improvement a real assignment

Here is a practical exercise for a programmer and manager to try together: choose one recurring task the programmer tends to avoid because it is outside their strongest skill set.

Record how long the task normally takes, where the frustration occurs, and what an acceptable finished result looks like. Complete a comparable assignment with AI assistance and keep the time spent checking and correcting its output in the total. Have a qualified colleague review the result against the same standard. Repeat on another comparable task before drawing a conclusion.

Then decide what the improvement will make possible. If a verified workflow saves an hour, protect that hour for an aircheck session, a stronger client idea, or a local relationship that has been waiting for attention. If it improves quality without saving time, that may still be a worthwhile investment.

The experiment should tell us whether the programmer is producing better work, learning something useful, and gaining capacity that reaches the station’s product.

Management decides where the time goes

This is the responsibility I keep returning to. AI can help a person become more capable. There are still limits to the number of brands, people, and decisions that person can give thoughtful attention to in a day.

The assignments that tend to slip are often the ones with no immediate deadline: developing talent, listening critically, planning an original promotion, and spending time with a business owner before the next contract comes up. These are exactly the activities that recovered time could support.

I want radio to earn a creative dividend from its technology investment. Measure the better programming, advertiser ideas, and local relationships that the investment makes possible. Hours saved matter because of what people can do with them.

The same principle applies when personalities develop new revenue. A host’s local credibility may support a campaign that combines an honest endorsement, short videos, and an appearance. That work needs planning, support, and a fair compensation arrangement. Put the person whose relationship helps create the value into the revenue model.

What I want radio to gain

McVay’s encouragement to learn and grow is sound advice. A programmer who uses AI thoughtfully may discover that a former weak spot becomes a useful part of their contribution to the station.

I want ownership to make room for that development. Give people training, clear priorities, and time to apply what they learn. Let some of the gain reach the listener and the advertiser. Make the work more ambitious, more local, and more worth paying attention to.

After four decades in this business, I am still excited by what I can create. I want more broadcasters to experience that feeling, including in parts of the job they once found intimidating.

AI gives us a chance to become better broadcasters. Radio management has to give that improvement somewhere to go.

Joe “Crash” Kelley is a four-decade broadcaster, operations and production leader, 2026 GABBY Award winner, and creator of SonicAttention.com.

Expanded from an essay first published in The Signal, with additional research and practical applications for radio programmers.

Hear the work. At SonicAttention.com, original human ideas become custom jingles, commercials, and radio imaging. Bring me a business you want to win or renew. I’ll create a free custom spec. You pay nothing unless your client buys it. Tell me about the project.

Related reading:AI Should Make Radio Better—NOT SMALLER and Your Voice. Their AI. Who Gets Paid?.