Welcome to ‘The AI Edge,’ a series exploring how professionals across industries are using AI at work.

Today, we’re talking to Jamie Juviler, a seasoned content marketer with a background in web development. He oversees the content strategy for HubSpot’s Website Blog, leads growth initiatives for the blog, and manages a team of writers.

When it comes to AI, he’s a strong proponent of it at work, using it to brainstorm website conversion strategies, write project briefs, and debug code.

He even uses ChatGPT to help him be a better people manager.

Let’s get into how he does it all.

How AI Fits Into This Content Marketer’s Day

AI has a place in almost everything Juviler does.

“Be it writing, be it content strategy, or even people management – I think AI can at least make some tasks easier,” he says. “How I use it depends on the task at hand. It definitely plays a greater role in some aspects of my job than others.”

Since you can’t talk about AI at work without productivity coming up, I wondered if Juviler could quantify AI’s impact on his work.

He estimates his efficiency has gone up at least 30%.

“It’s made me tremendously more efficient because there are projects that I probably wouldn’t be able to do by myself, if at all, if not for AI,” he says.

One of those projects is BlogBot, a tool Juviler developed (thanks to his developer background) to speed up the CMS upload process.

“I’ve learned pretty complex things in JavaScript that would have taken me probably ages to just figure out and try to implement myself,” he says. “BlogBot probably wouldn’t be in the state it’s in currently if it weren’t for AI.”

On the project management side, he uses ChatGPT to:

• Write the first draft of any project brief he’s writing.
• Analyze data and pull insights.
• Ideate and brainstorm strategies.

“I could feed it a bunch of conversion data and it would tell me, ‘Here are some specific posts to focus on,’ or ‘here’s a specific topic area to focus on,’” Juviler says.

Onto one of the most interesting use cases for AI: People management coaching.

Juviler’s been a people manager for about a year now and ChatGPT has come in handy a few times, as he refines his leadership skills.

“I’m still learning how to best coach my employees to do their best work based on how they are, what their needs are, how they receive feedback, how they prefer feedback, what their working style is – all those things,” he says. “So, I found that ChatGPT is a really helpful coach for me to, in turn, coach my reports.”

He recalled a recent conversation with a direct report about improvements they were looking to make in a particular area.

ChatGPT helped him brainstorm exercises to run through with his report.

Juviler isn’t the only one turning to AI for support at work. 

Back in February, a former Dropbox employee and manager told Axios she had used ChatGPT to write performance reviews. 

The article includes numerous accounts from employees across industries using AI for similar purposes.

Now, the question is, how do make sure you get the results you’re looking for with AI?

Can’t Do AI Without Some Prompt Engineering

Six months ago, Juviler was pretty new to conversion optimization (CRO).

He got up to speed pretty quickly with AI but it didn’t happen overnight. It took a lot of troubleshooting in the beginning.

“I found that the advice it gave, at least to me at first, was pretty basic. Like, add more CTAs,” he says. “When you get a response like that, you need to ask clarifying questions to hone its responses.”

It’s been said time and time again but Juviler confirms it: Prompt engineering is key to a high-quality output.

“I’ll start with broader questions and see if I can do less work and just get the response I want, but if I find that it’s struggling to give me precise, helpful feedback, then I definitely use some prompt engineering.”

It looks a little something like this:

• Give the AI tool a job and background. E.g.: “You’re a marketing expert with 10 years and CRO experience on blogs.”
• Give as much context as possible about the situation. E.g. “You run a blog that gets 500K average monthly views, mostly from organic traffic, and your conversion rate has been steadily decreasing in the past 3 months.”
• Provide data if relevant.
• Make a specific request. E.g. “Present 10 possible approaches to increasing our conversion rate, explaining why each strategy would work and why it might not.”

Over time, Juviler has built a reference document of free prompts and templates that he uses regularly.

Where This Content Marketer Draws the Line with AI

Juviler is clear on the value AI brings to his work, but he’s also mindful of overreliance.

“I often wonder if I’m using it too much and if it’s dulling my ability to further my own learning,” he says.

He’s not alone – whenever new tech comes around, there’s this conversation around overreliance. It happened with calculators, remember?

But there’s a line Juviler won’t cross.

“I never have it write blog posts, any reader-facing content on the blog, I only use it for research,” he says. “That’s a pretty strict line I keep for myself.”

He adds that AI’s off the table during any interpersonal communication.

“If I’m writing a Slack message or an email, I don’t ever fake authenticity with AI,” he says. “If I am coming off as a real person to somebody, I’m going to be that person – that’s just a moral thing.”

Juviler’s also keenly aware of AI’s limitations, offering a word of caution to aspiring developers.

“If you’re excited by the idea of using generative AI to help you program things, you still need to learn the fundamentals,” he says. “I wouldn’t be able to understand anything I was doing if I didn’t have a solid kind of fundamental knowledge of programming.”

Getting any mileage from AI tools like ChatGPT, he says, will require a strong foundation of knowledge.

All that said, Juviler sees AI as a value add – particularly for those in a similar career to his.

“I think it’s a big win for marketers,” he says. “If you’re a marketer and there’s an area you want to improve on, something you want to pursue that you don’t have a lot of experience in, the barrier to entry is made a lot lower with AI.”