Smartly launches AI to predict and explain creative performance across Social, Google, Programmatic, and CTV

Smartly adds AI creative predictive tools across advertising channels

Smartly introduced two AI creative intelligence products that connect creative decisions with performance across social, Google, programmatic, and Connected TV. The pitch is simple: predict how your ads will perform before launch, then learn exactly why they worked (or didn't) after.

For creatives, this means fewer guesses, fewer "run it and see" tests, and more clarity on which visuals, tones, and edits move the needle.

What launched

Creative Predictive Potential runs pre-flight. It uses computer vision and eye-tracking models to estimate attention, sentiment, and engagement-before you spend a dollar. You get recommendations to tighten messaging, pacing, and visual emphasis.

Creative Insights runs post-flight. It analyzes themes, tones, and visual patterns across your ads and turns performance data into explanations-with AI summaries and theme analysis-so you know what to scale and what to cut across every channel.

Why this matters for creatives

Smartly manages over $6B in ad spend for 800+ brands and sits across major inventory partners (Amazon, Google, Meta, Pinterest, Reddit, Snap, TikTok). That reach lets creative intelligence feed directly into media activation-one workflow, less friction.

Most teams still reuse assets: 72% repurpose creative across platforms, and only 25% tailor for both social and CTV. Meanwhile, CTV's budget share is set to double from 14% (2023) to 28% (2025). Quality matters: enhanced CTV formats have shown 78% higher attention rates than standard units according to Kargo.

How Creative Predictive Potential works (pre-flight)

The system estimates viewer attention and engagement via computer vision and eye-tracking models. It flags the moments where attention drops, where pacing lags, or where the message confuses-and suggests edits before launch.

• Storyboard with "attention anchors": opening visual hook in the first 1-2 seconds, logo or value cue by 3-5 seconds, clear CTA by the end.
• Test multiple intros and thumb-stopping frames; keep the first scene change under 1.5 seconds for social and 2-3 seconds for CTV.
• Use contrasting color blocks and large typography for mobile. Avoid overcrowded frames that dilute focal priority.
• Map variants to placements (vertical, square, horizontal) so the tool can score attention by format, not just a single master.

How Creative Insights works (post-flight)

After the campaign, AI Theme Analysis clusters high-performing creative patterns across visuals, tone, and messaging. AI Summaries explain what drove outcomes across channels-useful when you're managing thousands of variants.

• Identify high-ROI motifs: product close-ups vs. lifestyle, voiceover style, tempo, color palettes, on-screen text density.
• Spot cross-channel winners and "channel-only" winners (e.g., social cuts that outperform but don't translate to CTV).
• Build a living playbook: validated hooks, pacing rules, CTA styles, and brand codes that consistently hold attention.
• Feed the next brief: roll winning components into templates; retire patterns that correlate with attention drops.

CTV is the big unlock

Smartly is bringing its social-first intelligence to Connected TV. That matters because CTV is sensitive to creative quality and context. Repurposing online video often underperforms in living-room environments.

• Design for distance: bigger text, fewer words, clear focal points, and slower cut cadence than mobile.
• Front-load story clarity: viewers decide quickly whether to keep watching-even on the couch.
• Use episodic sequencing: three 15s variations that build on each other can outperform one 30s catch-all.

One workflow from idea to outcome

Smartly's platform ties creative production, media activation, and optimization together. Automated templates and modular editing enable 150M+ variants monthly, with cross-platform reporting in one place.

They cite PwC-validated outcomes: 5.5x return on ad spend and saving 42 minutes every hour through operational efficiency. If you're shipping high volumes of creative, that time delta compounds fast.

Industry moves you should know

The timing fits a wider push toward AI in ads: Amazon launched AI agents for campaign management on November 11, and Google made Ads Advisor widely available on November 12 details here. Marketers plan to increase AI use (72%), but only 45% feel confident applying it. Tools that turn creative data into clear actions help close that gap.

What to do next (practical checklist)

• Pre-brief: define the attention goal by placement (e.g., 3-second hold on Reels, message clarity by 5 seconds on YouTube).
• Variant plan: 3-5 intros, 2 pacing options, 2 CTA formats per concept. Keep swaps modular.
• Pre-flight scoring: run Predictive Potential; implement the top three fixes (opening frame, text density, pacing).
• Channel fit: produce platform-native cuts (vertical for social, cinematic for CTV). Don't force-fit.
• Post-flight: use Creative Insights to tag winning themes. Update your creative playbook the same week.
• Scale with guardrails: lock brand codes (color, logo, VO tone), then let templates iterate scenes and hooks.

Why this is good for creative teams

• Fewer blind tests. You enter launch week with data-backed creative choices.
• Cleaner feedback loops. Insights explain why an ad worked, not just that it did.
• Faster iteration. Templates plus theme analysis turn learnings into new variations quickly.
• Better cross-channel fit. Social learnings inform CTV cuts without copy-paste mistakes.

Resources

• AI courses by job: level up creative and marketing workflows
• AI Certification for Marketing Specialists

Bottom line: creatives who treat attention as a design system-not an afterthought-will win across social, Google, programmatic, and CTV. Predict pre-launch, explain post-launch, then ship more of what works.