A day or two after the 2026 Super Bowl, viewers could still describe their favorite ads joke by joke. Some of them couldn't tell us which brand paid for those ads. We wanted to know which creative choices make an ad stick to the right brand, how this year's celebrity-heavy and AI-made ads landed, and whether any of it changed what people thought or did.
So within 24 to 48 hours of the 2026 game, we ran ad testing with 56 US adults who had watched the whole broadcast and paid attention to the commercials. Each one sat down for a video interview with our AI moderator.
When we asked what made an ad stick, humor came up first. 59% named it. The brand came up far less often. Only 41% remembered an ad because the brand was part of the idea, meaning the joke or story wouldn't work without it.
TL;DR: humor and celebrities got phones down, but viewers mostly credited the right brand when the brand was the joke. Two in three did nothing afterward, and the few who acted mostly searched.
Key takeaways
Funny ads got remembered, but not always for the right brand. 59% said humor made an ad stick, while 41% remembered one because the brand was built into the idea. Some viewers could retell a joke and still not name the brand.
Celebrities helped until there were too many. 78% said the right celebrity makes an ad more memorable. But 63% felt this year's game had too many, and when we asked viewers to name the brand behind each celebrity ad, 45% said the star had overshadowed it.
AI-made ads drew strong opinions but changed little. After we told viewers one ad, from a vodka brand, was made entirely with AI, 57% called AI-made ads lazy or inauthentic. Yet 55% said knowing an ad used AI wouldn't change how they feel about it. Poor or off-looking visuals were what cost trust.
Ads moved brand opinion in both directions. 51% left the game with a better view of at least one brand. 32% left with a worse view of one, citing privacy worries, an off-putting tone or a celebrity who didn't fit.
Very few viewers acted on what they saw. 67% took no action because of an ad. The most common step was a search (23%), and only 6% bought something.
How we ran the ad test
We recruited US adults aged 18 to 70 who watched the entire game live, including halftime. We screened out anyone who spent the breaks getting food, muted the ads or didn't remember paying attention to them.
The interview moved from memory to behavior:
Recall. Which ads do you remember, without looking anything up? Was there a moment you put your phone down to watch?
Brand opinion. Did any ad change how you feel about a brand, or how much you trust it?
Action. Did you search, visit a site, sign up or buy anything because of an ad?
Celebrities. How did the celebrity presence feel? For each celebrity ad you remember, which brand were they selling?
AI. Did any ad seem AI-made? Then we told viewers that one ad, from a vodka brand, was made entirely with AI, and asked how that changed their view.
Head to head. Viewers watched four ads from the game on camera and talked through their reactions.
The AI moderator asked follow-up questions after each answer, then grouped the open-ended answers into themes. Not every viewer reached every section: 56 answered the recall questions, 53 the brand questions, 52 the action questions and 51 the celebrity and AI questions. The percentages below use those counts.
1. Humor got phones down. The brand didn't always come with it
When we asked what made an ad stick, viewers gave these reasons most:
Why viewers remembered an ad | Share of viewers |
|---|---|
It was funny | 59% |
The brand was part of the idea or setting | 41% |
It had a twist or surprise | 36% |
Viewers could name more than one reason, so the shares add up to more than 100%.
The 18-point gap between the first two rows is where ads lost their credit. Some viewers could describe a funny spot in detail and still not say who it was for. Ads they found confusing lost the brand name even when they enjoyed them.
When viewers did remember the brand, the ad usually did one of three things. It made the product part of the joke, showed what the product does, or set the story somewhere that obviously belonged to that brand.
2. Celebrities helped, and there were too many of them
Viewers liked celebrities in principle. 78% said a well-cast star makes an ad more engaging or memorable. But 63% also said this year's game had too many of them, and since those two shares add up to well over 100%, at least 21 of the 51 viewers felt both ways at once.
The real cost appeared when we asked viewers to name the brand behind each celebrity ad they remembered. 45% said the star had overshadowed the brand or its message, and some could picture the celebrity clearly without recalling the product at all.
What separated the ads that worked from the ones that didn't was fit. When a celebrity's role said something true about the product, viewers connected the two. When the celebrity appeared for a quick cameo, they took attention away from the brand.
3. An AI-generated vodka ad split viewers, but most said AI wouldn't change how they judge an ad
After asking whether any ads seemed AI-made, we told viewers that one ad, from a vodka brand, was made entirely with AI.
57% said ads made this way feel lazy or inauthentic. 49% said they feel innovative or positive. 55% said knowing an ad used AI makes no difference to how they feel about it. The three groups overlap because many viewers said more than one of these things.
Whether AI was used mattered less to viewers than how good the finished ad looked. Visuals that looked slightly off cost trust quickly. Viewers were more accepting when the work was polished, when synthetic people on screen were disclosed, and when the brand was in tech or had a playful tone.
4. Brand opinion moved both ways
51% said at least one ad gave them a more positive view of a brand. 32% said an ad left them feeling more negative about one.
The negative shifts came from three places: privacy worries about the product being advertised, a tone that put viewers off, and a celebrity who didn't fit the brand. An ad can win on recall and still cost you on opinion, and a recall-only test won't show it.
5. Two in three viewers didn't search, click or buy anything because of an ad
67% said no ad led them to do anything during or after the game. The rest took small, fast steps. 23% searched online because of an ad, 15% visited a website or scanned a QR code, and 6% bought something.
Searches came from curiosity. Viewers looked for a longer cut of an ad they liked, or checked a detail the ad raised, such as how a product handles their privacy. That moment favors one-step calls to action and landing pages that hold up under game-night traffic.
What the head-to-head ad test showed
Each viewer watched four ads from the game on camera and talked through their reactions. We're describing the brands by category.
The chip brand's ad won on humor and pop-culture pull, but some viewers couldn't say why they'd choose that chip over another. The men's grooming brand's ad was the most memorable and the most divisive, with some viewers calling it crude. The condiment brand's ad stuck through humor and catchy music, though some found the food shots unappetizing. The home appliance brand's ad drew attention with its celebrity, but some viewers couldn't connect it to what the brand sells.
All four were funny. They split on whether viewers came away knowing what the product does and why it belonged to that brand.
Where Listen fits
We ran this study on Listen Labs. Our AI moderator interviews each person on video, follows up on every answer and groups the answers into themes, so each percentage comes with the reasons behind it.
The follow-ups are where viewers who loved a funny ad admitted they couldn't name the brand.
What to do with this before your next big campaign
Once a big-moment spot airs, you can't edit it. These five checks come straight from what viewers told us, and each one catches a problem a standard recall score would miss.
Ask for unaided brand recall in every ad test. Enjoyment and brand credit came apart by 18 points (59% vs. 41%). Ask viewers who the ad was for before you show the logo again. If they can retell the joke but not the brand, the creative needs the product written into the idea itself.
Test whether the celebrity sells the brand. 78% liked a well-cast star, but 45% said stars overshadowed the brand. Ask viewers what the celebrity was selling. If they can't say, the fee is buying attention for the celebrity.
Judge AI-made creative on how the finished ad looks. 55% said AI use itself doesn't change their view, but visuals that looked slightly off cost trust fast. Put the finished cut in front of viewers and ask whether anything felt wrong, and disclose synthetic people on screen.
Score brand opinion alongside recall. 32% came away thinking worse of a brand, for reasons like privacy worries, tone or a mismatched celebrity. Ask how the ad changed their view of the brand and why, so you catch backlash while you can still edit.
Plan for what viewers do after the ad. 23% searched because of an ad, while only 6% bought. Make sure the search results and landing page answer the question the ad raised, and keep the next step to one tap, like a QR code or short URL.
Methodology
Listen Labs interviewed 56 US adults aged 18 to 70 within 24 to 48 hours of the 2026 Super Bowl. Participants came from a general population panel and were screened to include only people who watched the entire game live, including halftime, and paid at least some attention to the commercials.
Interviews were AI-moderated video conversations covering ad recall, brand opinion, actions taken, celebrities, AI-made ads, and reactions to four ads from the game. Bases vary by section: 56 for recall, 53 for brand opinion, 52 for actions and 51 for celebrity and AI questions. Open-ended answers were coded into themes, and a participant could fall into more than one theme, so shares within a question can add up to more than 100%.
No individual responses are quoted.