UGC-style fitness ads with AI: why imperfect sells more

A video shot on a member's phone at your gym, mediocre lighting, background noise, typically gets double or triple the CTR on cold traffic compared to the €800 spot you commissioned. That's not an anomaly: it's been the dominant pattern on Meta for years, and in local fitness it plays out with almost boring regularity. The reason is simple. Instagram's feed is full of content from regular people, and an ad that looks like content from a regular person doesn't trigger the "this is advertising, keep scrolling" reflex.

This article explains what UGC-style actually means (it's not just "shoot it badly"), the three ways to get it with real prices, what happens when you bring AI into the equation, and three concrete scripts you can shoot this week.

What UGC-style really means

UGC stands for user generated content: content made by users, not the brand. But when advertisers talk about "UGC-style ads" they mean something more specific: ads designed to look like a regular person made them on their phone, even if the brand commissioned and paid for them.

The visual codes matter more than the content. Here's what makes a video "read" as UGC:

The counterintuitive part: the script is thought through and the structure is deliberate. UGC-style isn't amateur content, it's professional content dressed up as amateur. A good UGC video has a hook in the first 2 seconds, a central argument and a CTA, just like any ad. What changes is the skin.

Why does it work so well for gyms specifically? Because joining a gym is a trust decision, not a product decision. Nobody doubts you have dumbbells. What they doubt is whether they'll fit in, whether people will stare, whether they'll quit after three weeks. And that doubt is better resolved by someone who looks like them saying "I thought the same thing, and look" than by any aerial drone shot of your gym floor. It's the same mechanism behind why member testimonials are the most underused advertising asset in the industry.

The honest caveat: UGC-style wins on cold traffic, where the barrier is distrust. On retargeting, where the person already knows you, a more produced ad with a clear offer can perform just as well or better. Don't convert your entire creative mix to UGC; use it where it hits hardest, which is at the top of the funnel.

Route 1: real members (the best option and it's free)

Your best UGC creator trains at your gym three times a week and doesn't know it. Videos from real members have something no paid creator can replicate: when someone in the neighbourhood sees the ad, they might run into that person at the supermarket. The social proof is literal.

The challenge isn't getting the content, it's asking for it without making things awkward. What works:

Ask the right person at the right moment. Don't send a mass message to all your members. Identify 3–5 people who already speak well of you, have been coming for over six months, and have a story (lost 15 kg, came back from an injury, started at 55). Ask them right after a good moment: they just did their first pull-up, they just renewed for the year. The "yes" comes easily.

Make it easy. Nobody wants to "record a testimonial." Propose something specific: "Can I film you for 60 seconds on your phone telling me what scared you before you started and what happened after? I'll send you the three questions beforehand." Three questions, their phone, two minutes. If you ask them to prepare something, it'll never happen.

Offer something in return without it feeling like a payment. A month's membership, two PT sessions, their favourite smoothie for a month. It's not a commercial transaction, it's a thank-you, and that framing completely changes their willingness.

And the paperwork, always. Before using a member's face in an ad you need a signed image rights release that's specific to advertising (the generic consent in their membership contract usually only covers internal communications or organic social). One page, plain language, stating it will be used in paid Meta ads, with a duration and a right to revoke. Your accountant or lawyer can have it ready in a day. Skipping this gets very expensive the day a member leaves unhappy and asks you to pull your best-performing campaign.

The full guide on how to build a UGC content system at your gym covers the whole process, from internal casting to the content calendar.

Route 2: paid UGC creators

If you don't have willing members or need volume, there's a fairly mature UGC creator market in Spain. Current real prices: between €50 and €150 per video for creators without a large audience (which is what you want, you're paying for the video, not reach). A pack of 3–4 videos with hook variations is usually negotiated at €200–400. Find them on marketplace platforms or directly on Instagram by searching "UGC creator" + your city.

Advantages: speed, consistent execution quality, zero social friction. A skilled creator delivers a video with hook, subtitles and three variants within a week.

Downsides nobody mentions: the creator doesn't know your gym. If you send a two-line brief, they'll deliver the same generic video they gave a gym in another city last week, interchangeable phrases, catalogue-level enthusiasm. The brief is 70% of the result: give them your exact offer, the main objection your leads raise, two or three phrases your members actually say, and ask them to record it as if they were telling a friend. And if they can come and work out at your centre once before filming, the video improves by a whole category.

Second downside: the face in the ad will never be inside your gym. For cold traffic from a national brand, that doesn't matter. For a neighbourhood gym, it loses part of its power. It works, just not quite as well as a real member.

Route 3: synthetic UGC with AI

It exists, it's cheap, and we need to talk about it honestly. Today you can generate a video of a person who doesn't exist speaking to camera about your gym for under €10, using avatar and generative video tools. The technology has improved so much that many of them pass at scroll speed in a feed.

My position: for local fitness, it almost never makes sense. Three reasons.

First, the uncanny valley is still there. Current avatars fail at exactly what makes UGC good: the human micro-imperfection. They gesture too uniformly, they breathe oddly, the skin lighting doesn't quite land. Part of your audience won't be able to say what's wrong, but they'll feel something is off, and that feeling is the exact opposite of what you're trying to create with UGC.

Second, the trust risk is asymmetric. If nobody notices, you get a cheap video. If someone notices and comments ("this gym uses fake AI testimonials"), the damage in a neighbourhood business, where everyone knows everyone, far outweighs what you saved. And every month people are better at spotting synthetic content.

Third, there's a clear red line: a synthetic testimonial presented as real is deception, full stop. An AI-generated person saying "I lost 12 kg at this gym" describes a result that never happened. Aside from the ethical problem, Meta's policies on misleading content expose you to ad rejections and degraded account health.

Where does AI make sense in UGC? In everything around the human video: generating the scripts and hook variations, subtitles, cutting 15/30/60-second versions, translation if your neighbourhood is multilingual. The human part of the video, keep it human.

On disclosure: if you use synthetic content, say so. Meta already labels AI-generated content in several contexts and the regulatory direction in the EU is unambiguous. Getting ahead of it costs little; getting caught costs a lot.

Three UGC scripts that work for gyms

The structures below are tested in local fitness. Shoot them on front camera, vertical, in a single take if you can. Target duration: 30–45 seconds.

1. The converted sceptic

Hook: "I was the typical person who signs up in January and quits in February."

Structure: first-person objection (2–4 seconds) → what was different this time, with one specific detail about the gym (15–20 seconds) → current state with a data point ("I've been going 11 months, I can't believe it myself") → soft CTA ("if you're in the same boat I was, check this out").

It works because it disarms the number one objection (dropping out) in the first second, said by someone who lived it.

2. The normal day

Hook: "Let me show you what coming here looks like on a random Tuesday at 7pm."

Structure: handheld walk-through of the entrance, the floor, a class starting (15 seconds) → a spontaneous comment about the atmosphere ("nobody watches you here, everyone's focused but they'll say hi") → CTA with the offer.

It works because it answers the question no polished ad answers: "what am I actually going to walk into?" It's the easiest video to shoot and the one that generates the most booked visits on cold traffic.

3. The uncomfortable question

Hook: "How much does this gym actually cost? I'll tell you straight."

Structure: exact price and what it includes, no hedging (10 seconds) → honest comparison ("there are cheaper places, here's what you get here that you don't get there") → who it's not for ("if you just want a treadmill, don't come, a budget gym is fine for you") → CTA.

It works because aggressive price transparency generates disproportionate trust, and the "who it's not for" part qualifies the lead before they arrive. You get fewer clicks and better ones.

Shoot all three structures with two different people and you have six creatives for 6–8 weeks of campaign. When one starts to fatigue, change the hook before changing the whole video, creative fatigue signals show up early if you're watching for them.

Where to start this week

If you have a member with a good story and the right attitude, start there: it's free, it converts best, and your accountant sorts out the release form in a day. If not, a paid creator with a detailed brief is a solid shortcut for €100. AI, for now, belongs in the scriptwriting room rather than in front of the camera: at Pilotium we use it daily to generate and vary scripts and ad copy for gyms, which is precisely how we know where it stops being a good idea.

The definitive test is uncomfortable to admit: show your ad to someone without telling them it's an ad. If they ask "who is this person?", you've got UGC. If they say "oh, a gym ad," start over.