Upfront Highlights: Canela’s AI Microdramas, Amazon’s Shopping Play

1. Canela’s AI Microdramas

If I had to pick two stories from this week’s Upfronts that are emblematic of major changes in the industry, they would be Canela’s new microdrama app, Zully, and Amazon’s focus on commerce.

Zully is in many ways the more important of the two stories as it is the forerunner of a whole new era of content creation.

It was something I had predicted last August, though honestly it was not all that prescient—my premise was that AI actors could not be much worse than the real ones and that, more than that, the audiences would not really care if the actors looked slightly unreal. That it was all part of the vibe.

So kudos to Canela for realizing that the tech was ready and rolling out an app that uses AI to produce a series of original microdramas, many of them Latino-themed adaptations of classics like Pride and Prejudice.

Why It Matters

AI video is having a moment in the political arena which meant that it was only a matter of time before it migrated elsewhere. 

Cases in point: the Iranian Lego videos and Spencer Pratt’s LA mayoral campaign.

Microdramas are well-suited to AI because they rely on close-ups and exaggerated facial expressions rather than beautiful scenery and artfully lit backdrops.

Studios, like the one Zully is using in Mexico, can crank them out at warp speed and even make adjustments on the fly. Meaning you can A/B test different storylines in real time to see which one is doing better. Make different versions for different regions by tweaking the way the actors look or even their accents.

It’s not art, but it’s not supposed to be.

Zully was a particularly smart move for Canela Studios, which produces streaming content aimed at Latino audiences in the US. 

That’s because Zully, with its Spanish language microdramas, can be deployed all across Latin America, which is a prime market for this sort of content.

What’s more, Zully can be promoted all across Canela’s content line-up, thus solving one of the biggest problems for all of the emerging microdrama apps: how do you promote it?

So that is all that is good about Zully and it’s easy to see how this sort of production capability then migrates to other categories too: soap operas/ telenovelas/ K-dramas/ dizis are easy, then non-fiction TV shows, then, eventually comedies and dramas.

Will AI write the scripts too?

People may not want to hear this, but for many of the categories it eventually will. Not all of them—high production value dramas and comedies will still be written by humans.

And if you are about to send off an angry comment, ask yourself if House Hunters really needs to be written by an actual human.

What You Need To Do About It

If you are Canela, take a bow. Getting out there first with something that is likely to become table stakes is a big win, provided the shows are what the audience actually wants, which is still a big risk.

But the rest of the industry is not far behind you: Janko Roettgers is reporting today that Netflix is busy making AI-created animation. 

And if Netflix is doing it, everyone else can’t be too far behind.

If you are the rest of the industry, I suspect that I am correct, that none of you are all that far behind Canela. But if you are not, then WTF?

If you are considering an acting or writing career, realize that this is both good and bad news. On the one hand, you don't really want to be the one writing or acting in these sorts of shows. On the other, they do serve as an easy source of income in hard times. 

Everything is a trade-off.

2. Amazon’s Shopping Play

The other harbinger of the Upfronts was Amazon’s emphasis on how it could use all of its data to enable performance marketing, e.g. selling stuff.

And I use the words “all of its data” in a somewhat Talmudic sense, to mean not just the data it has around its TV series or their audiences, but rather all of the data it has collected since 1994 around what we buy and what people like us buy.

Which is a lot.

Meaning Amazon knows that no matter how much you talk about Starbucks and pay attention to their TV commercials and Instagram Stories, you have a recurring subscription for Gevalia and are unlikely to actually buy anything from Starbucks.

Making their targeting all the more valuable.

Why It Matters

Amazon is doing what we always suspected they’d start doing, which is using AI to make ads that are hypercustomized to the consumer in question.

It’s sort of what you can do with microdramas—change the accents from Minnesota to Mississippi, but only in the service of sales, not engagement. 

This has been the promise of streaming for the past decade or so, the sort of personalization you can get via online, only not quite as creepy, and Amazon is finally making it happen.

And it’s not as if Amazon is the only one who has all sorts of data. Walmart may have a couple of loyal shoppers (or so I hear) and they also own VIZIO, giving them a place to make use of all of that data.

Ditto retailers like Target, credit card companies and the like.

The key of course will be to make the ads look useful and relevant and not creepy and stalkery, something the industry does not have an ideal track record of doing, so there’s that and it’s no small thing.

But should they come to their senses and keep things on the up-and-up, then consumers will start to see more ads for things they want and less ads for things they don’t want—minute-long pharma ads in particular.

The key will be to keep those “buy now” ads largely relegated to the non-premium inventory and keep the branding and image ads in with the premium inventory.

Branding, of course, being a premium endeavor. 

What You Need To Do About It

If you are Amazon, well done. If your plan works as promised, you will have made TV advertising more relevant and less annoying. Kudos for that one.

If you are VIZIO/Walmart, you’re on deck. (Not that you didn’t already know that, but…)

If you are the advertising industry, remember that performance on non-premium, branding on premium is a guideline, not a rule. And rules were meant to be broken, depending on the product and the content.

That said, relying on very granular data for very granular targeting is the future, especially as the type of goods that are sold online expands and opportunity expands along with it.

If you are a consumer, rejoice—your ads should be getting a lot more relevant.

Should being the operative word here.