How Plai Doubled Annual Revenue and Ships 30,000+ Ad Creatives a Month
Plai is an AI-powered advertising platform founded by growth marketer Logan Welbaum, a former Meta and Google operator who once grew a TikTok account to 450,000 followers before turning his attention to ad automation. Plai’s promise is simple: type a prompt, and the platform generates brand-aligned creative and launches campaigns across 11 publishing channels, for clients ranging from global acts like Green Day and Red Hot Chili Peppers to local service businesses.
The platform runs its in-product creative editor on IMG.LY’s CE.SDK. Here’s how it came together, and what changed inside Plai as a result.
The Challenge: AI Got to 80%, Then Users Had to Leave
Plai was already good at the hard part. Its models produced highly on-brand images and videos from a prompt, the kind of output that would normally take a designer or a Canva subscription to assemble. But generating an asset and finishing an asset are two different problems, and the second one kept tripping users up.
AI could draft a creative, but it couldn’t reliably control the details that decide whether an ad actually ships: the exact crop, the headline copy, a logo nudged a few pixels, a color matched to a brand guide. Those are the last 20% of the work, and they were exactly the part the model couldn’t be trusted with.
The most important part of an ad is what people see. Before IMG.LY, we couldn’t give customers the flexibility to edit and optimize AI-generated assets quickly.
Without an editing layer inside Plai, that last 20% became a detour. Users exported the AI output, opened a separate design tool, made their tweaks, and re-uploaded the result, breaking the one-prompt promise that made Plai compelling in the first place. Every round trip added friction and pushed campaign readiness further out.
Why They Chose CE.SDK
Logan’s team looked at the obvious alternatives, including pointing users at external editors like Canva. None of them fit. Sending customers to a separate product with its own account and its own login defeated the purpose: the whole value of Plai was that you never had to leave it.
Building the editor in-house was the other option, and Logan was clear-eyed about what that would cost a team whose core competency was ad automation, not rendering engines.
If you’re not a design-first company, you’ll pull your hair out trying to build an editor yourself.
CE.SDK won on the axes that mattered to a product company shipping fast. It was white-label, so it could be styled to feel like a native part of Plai rather than a bolted-on iframe. Its configurable, form-based editing UX suited Plai’s users, who wanted to adjust an ad, not learn a design tool. And it could be integrated quickly enough to matter to a roadmap measured in weeks, not quarters.
The decisive frame was the gap CE.SDK closed: AI takes the creative to 80%, and the editor lets the user finish the last 20% without ever leaving the flow. That’s the refinement layer Plai had been missing.
How They Implemented CE.SDK
Plai embedded CE.SDK directly into its creative flow, so the editor opens on the AI-generated asset the user just produced. From there they can adjust layout, rewrite text, fix proportions, refine colors, and apply brand-specific edits, then publish to any of the 11 connected channels, all inside Plai.
The integration was fast. What would have been a multi-year platform commitment if built from scratch went live in about a month.
Fast to launch, that’s the biggest thing. Building this ourselves would’ve taken months. With IMG.LY, it was live in a month and now drives thousands of ad campaigns.
Because the editor was white-labeled, users didn’t experience it as a third-party tool. It read as part of Plai, which is exactly what kept them in-product instead of tabbing out to finish their work elsewhere.
The Results
With the editing layer in place, the AI-to-publish workflow finally closed end to end, and the output numbers reflect it. Plai now exports between 30,000 and 37,000 ad creatives per month, with users finishing and shipping assets in the same session they generated them.
The time savings were just as stark. Campaign prep that used to stretch across weeks, with external editing and re-uploads in the middle, collapsed to minutes once the refinement step lived inside Plai. For Plai’s clients, that translated directly into budget: businesses that would otherwise hire a designer or pay for separate design software saved between $1,000 and $10,000 per month.
Since integrating IMG.LY’s editor, we now deliver 30,000+ ad creatives per month, dramatically shorten time-to-launch, and help our clients cut design spend by up to $10,000 per month.
The editor also became a moment in the sales conversation, not just a feature in the backlog. The visible polish of the in-product editing experience changed how prospects reacted to a demo.
Customers’ eyes light up when they see it. They know they’d otherwise hire a designer or buy Canva.
Overall Impact
The headline result is the one that shows up in the board deck: over the period Plai integrated and scaled creative automation, annual revenue doubled, with the editing layer as a core driver. But the structural win sits underneath that number.
By treating CE.SDK as the refinement layer on top of its AI, Plai kept its engineering focus where its advantage actually lived, on ad generation, channel integrations, and campaign automation, rather than sinking a team into building and maintaining a creative engine. The result was a product that finishes what its AI starts, keeps users inside a single flow, and turns thousands of prompts a month into ready-to-ship campaigns.
That’s the shape of the result for an AI-first product: the model gets you to 80%, and an embedded editor turns the last 20% from a reason to leave into a reason to stay. With video editing and AI audio next on the roadmap, Plai is extending the same pattern to its next set of formats.
