After spending several days speaking at DigiMarCon, Dennis Yu and Sam McLeod noticed the same issue coming up again and again with business owners, agency leaders, and local service professionals.
Most people in the room were already using AI tools. Very few felt confident that those tools were actually helping their business.
The issue wasn’t access. It was approach.
Why Some People Get Better Results from AI
During the event, Dennis Yu shared an observation based on how he personally uses ChatGPT.
He pointed out that people who take the time to communicate clearly — even using simple phrases like “please” and “thank you” — tend to get better responses. Not because the tool needs politeness, but because those users are more likely to slow down and explain the task properly.
Dennis explained that the quality of the output often mirrors the quality of the input. When users describe what they’re trying to accomplish and react to the response the way they would with a real assistant, the results improve.
That same pattern showed up repeatedly in conversations with agency owners and local service businesses at DigiMarCon. The people getting usable output weren’t typing one-line prompts. They were giving context and treating the interaction like an ongoing task.
The Common Mistake Most Business Owners Make
Throughout their sessions, Dennis and Sam McLeod kept seeing the same habit.
Most people were using ChatGPT the way they use Google. Ask a question. Read the answer. Move on.
Dennis explained that this limits the value of the tool. He emphasized that many users rely on AI to produce information, when it becomes far more useful when asked to process work.
That idea clicked once Sam started walking through how they use Agent Mode in real situations.
Repurposing Existing Content
Sam McLeod demonstrated how a single piece of content can be reused instead of rewritten.
Using Agent Mode, they showed how an existing video or long-form asset can be reused instead of rewritten, using the same repurposing process they follow inside The Content Factory.
- Create multiple blog posts
- Adapt messaging for Facebook ads
- Rewrite versions suitable for Google ads
The point wasn’t producing more content for the sake of it. It was helping local service businesses get more value from content they already have.
Improving Paid Advertising
Dennis Yu also walked through how they review ad copy using AI.
Rather than guessing what needs to change, they ask AI to:
- Look at an ad from the perspective of someone unfamiliar with the business
- Check whether the message reflects the company’s reputation
- Suggest small adjustments instead of full rewrites
This approach helps businesses refine ads without constantly starting over.
Handling Routine Tasks
They also showed examples of AI assisting with everyday tasks that often get pushed aside, including:
- Drafting email responses
- Answering common customer questions
- Handling repetitive work that doesn’t require strategic thinking
These weren’t framed as experiments. They were shown as normal parts of the workflow.
Why the Atlas Browser Changes the Workflow
Sam also introduced ChatGPT’s Atlas browser during the session.
Instead of copying information between tools, Atlas allows AI to move through steps on its own. They demonstrated how this makes it easier to:
- Review existing content
- Navigate pages
- Complete basic tasks with fewer interruptions
For many of the local service business owners in the room, this was the first time they had seen AI used to complete actions instead of just returning answers.
Treating AI Like a Teammate
One theme that came up repeatedly in Dennis and Sam’s sessions was how people give instructions.
When users clearly explained what they wanted done, why it mattered, and what “good” looked like, the results were noticeably better. When prompts were rushed or vague, the output usually missed the mark.
Dennis made the point that this isn’t about learning prompts or tricks. It’s about explaining work the same way you would if you were delegating it to another person on your team.
Watch the repaired DigiMarCon Chicago session on Local Service Spotlight’s YouTube channel.
Add proof and guardrails before asking agents to act
The reason a clear instruction works is not that AI needs a clever prompt. It is that useful work needs a real source, a clear goal, and an accountable owner.
At DigiMarCon Chicago 2026, I opened with: “Everyone else does PowerPoint. I’m gonna show you something that’s real.” The practical point for a local business is to start with the evidence it already has: permissioned job photos and video, customer questions, reviews, team expertise, credentials, and documented work.
“We have to show evidence that we actually do the thing that we say we do.”
Before an agent drafts anything, identify what happened, what is permissioned to use, who can verify the facts, and where the finished asset should send a customer next. An agent can organize, summarize, format, and propose connections. It should not invent an experience, repeat an unverified number, or publish a sensitive claim without review.
The right brief gives the agent context: the business goal, intended customer, approved sources, constraints, destination, and what “done” means. In the session, I said, “If you’re clear about what the goal is, you don’t need to know exactly how it’s being done.” That does not remove human responsibility. It creates a boundary inside which the agent can do useful work.
A practical request might be: take this permissioned job-site interview; use only the recording and linked source files; flag anything that needs a fact check; prepare a reviewable article that embeds the approved video at the top; link it to the right service or proof page; and do not publish.
Keep a record people can inspect
AI can make mistakes, so a good workflow leaves an audit trail. Keep the original recording, transcript, notes, images, and source links together. Mark unclear names, claims, and numbers for review. Save what the agent used, what it changed, what it could not verify, and what requires a person’s decision.
During the session, I described the importance of documenting agent work so an unintended change can be inspected and corrected. The principle applies just as much to content as it does to a website: a result that cannot be checked against its source is not ready to become a public claim.
“You have to equip your agents just like regular employees. Give them names. You have to onboard them. You have to train them. You have to give them access to stuff. You have to manage them. You have to QA their stuff.”
Give each workflow one business outcome, only the access it needs, and a human gate for customer-sensitive, financial, legal, medical, reputational, or irreversible actions. The goal is not more automated output. It is a trustworthy system that turns real proof into useful assets and improves from what customers actually do next.
For the maintained Local Service Spotlight method, use Activate Your Agents: The 4-Stage Content Factory. It is the canonical hub for collecting real proof, producing and repurposing assets, distributing evidence, and measuring results. This session addition is supporting evidence for that existing system—not a second guide.
The Bigger Goal Behind the Work
Dennis emphasized that the real goal is helping local businesses control their marketing and build trust.
Most local service businesses already have a strong reputation in their community. The problem is that their marketing doesn’t always reflect that reputation, or it’s handled by someone else who doesn’t understand the business.
Dennis explained that AI makes it easier for business owners to keep marketing in-house, understand what’s being done, and tie everything back to trust and credibility instead of gimmicks.
He also mentioned training a young adult from the local community to handle marketing tasks, so business owners aren’t forced to outsource work they don’t fully see or control.
Where to Begin
For business owners unsure where to start, Dennis and Sam suggested beginning with work you already have.
Take content you’ve already created—reviews, photos, videos, customer stories, past posts—and focus on processing that material instead of trying to invent something new. That’s where AI becomes useful, because it helps organize, repurpose, and publish what already represents your reputation.
If you want help setting this up properly, this is exactly what Local Service Spotlight is built for. Through the Reputation MRI and Content Factory system, local service businesses turn their existing proof into blogs, social posts, videos, and structured content that builds trust and visibility over time.
Instead of guessing what to post or experimenting with tools on your own, you can get started with a Reputation MRI to see where your marketing stands today, then use the same repurposing workflows Dennis and Sam demonstrated to consistently publish content tied to your reputation.