Bannerflow today put a name and a structure on the AI tools it has been building into its Stockholm-developed creative platform. The suite, called Creative Intelligence, groups image generation, predictive attention scoring, a performance analysis layer and feed-driven dynamic ads into four stages the company labels Create, Evaluate, Understand and Act and scale. The release, dated October 12, 2026, gives no availability dates, no pricing and no validation data for its predictive models.

In Short

Bannerflow, a company that sells software for making online ads, has bundled several AI tools so teams can make ads, predict whether people will notice them, see which design choices worked and turn product lists into many ad versions. This matters to brands that produce large numbers of banner, social and video ads, because it tries to pull the guessing out of creative decisions before money is spent. What changes is that more of this happens in one place, though Bannerflow has not yet published evidence that its attention predictions match real results.

What Bannerflow actually put into the suite

According to Bannerflow, Creative Intelligence is "a connected suite of AI capabilities built into the advertising workflow." The company frames it against a pattern it sees elsewhere in the market. "While many advertising technologies have added standalone generative AI features," the release states, Creative Intelligence "applies AI throughout the campaign lifecycle - from producing assets and configuring product feeds to evaluating likely attention, understanding performance patterns and building on those insights at scale."

The four stages break down as follows.

Create

Teams can generate, edit, expand and upscale images inside Bannerflow's Creative Studio, according to the company. A second element, AI-assisted feed setup, lets users describe the product feed they need in plain language rather than configuring it field by field. The release does not say which image model powers generation, whether it is Bannerflow's own or licensed from a third party, or what rights apply to generated assets.

Evaluate

This is the stage with the most direct bearing on media budgets. Predictive scoring, according to Bannerflow, uses "attention heatmaps, focus analysis and clarity scoring" to indicate how a video or static creative "is likely to capture attention before campaign budget is committed." The tool also produces recommendations meant to flag distracting elements so teams can revise a creative before launch.

Bannerflow does not explain how the attention model was trained, what eye-tracking or panel data underpins the heatmaps, or how closely its predictions have tracked observed outcomes. Those omissions matter. Attention metrics have become a currency in their own right in programmatic buying, and the vendors that sell them typically publish at least some validation against in-market results.

Understand

A component called Creative Insights is designed to show how individual creative elements, such as calls to action, colour choices and animation timing, relate to campaign performance. According to the company, those insights stay inside the platform, "creating an evidence base that can inform future briefs and creative decisions." The release does not state which performance signals feed the analysis, whether data from external ad platforms is ingested, or how attribution between an element and an outcome is handled when many variables change at once.

Act and scale

The final stage is the one Bannerflow has sold for longest. Feed-driven dynamic advertising converts live product information into on-brand variations across products, formats and markets. Pricing, stock and availability remain synchronised with the source feed, according to Bannerflow, which the company says reduces manual work and helps keep campaigns accurate. The plain-language feed configuration from the Create stage reappears here; the release describes it in both sections.

Underneath all four sits what the company calls its data platform, described as the foundation for "richer analytics, creative-performance modelling, predictive scoring and more reliable testing and optimisation." No technical detail on that platform is given.

The research behind the pitch is not new

The release justifies Creative Intelligence with survey data. According to Bannerflow, its study of 300 marketers across Europe, the US and Canada found that nearly 80% regularly review performance metrics when shaping creative strategy, and 93% track multiple creative performance measures. Respondents named data analysis, performance reporting, optimisation and iteration among the most disconnected parts of their workflows.

Those figures come from The Bannerflow Study on Creativity and Data, which the company published on its own blog on March 26, 2026, and updated on August 17, 2026. The blog labels the study "2025-2026" and gives no fieldwork dates. Neither the blog nor the release names a research firm, describes the sampling method or gives a margin of error. The survey was conducted by a vendor whose product is designed to solve the problem the survey identifies.

The release also cites the IAB, stating that 56% of US ad buyers expected to devote more time or resources to marketing mix modelling. The release does not say which IAB report this comes from or when it was published, and the figure relates to budget-level measurement rather than to creative decisions. Marketing mix modelling estimates the contribution of channels and spend to sales, usually with weekly or monthly aggregates; it is not typically built to say whether an orange call-to-action button beat a blue one. Its inclusion signals a general appetite for measurement, not demand for the specific tools in this launch.

The underlying argument is better supported elsewhere. An EMARKETER report produced with Perion, covered by PPC Land in April 2026, found that 89.2% of 111 US agency and marketing professionals called creative important for performance, but only 3.6% said it was well understood and actively optimised. In that study, 72.1% waited for performance to decline before changing creative, and 41.4% received creative insights two to four weeks after launch. That research was also commissioned by a vendor with a product to sell, but it at least disclosed sample size, fieldwork month and spend threshold.

What the company said

Moses Velasco, Chief Product Officer at Bannerflow, set out the rationale in the release. "Marketing teams have more performance data than ever, but too little of it inspires the next piece of creative. Insights sit in dashboards, production happens elsewhere and, by the time teams connect the two, the opportunity has often passed," Velasco said.

"Creative Intelligence closes that gap by embedding AI throughout the workflow - removing repetitive work, turning campaign signals into practical creative decisions, and carrying those learnings directly into what teams create next," he continued.

Velasco also addressed control. "The result is advertising that is not only more efficient to produce and optimise, but more relevant and valuable to the people it reaches, helping brands create stronger connections and better experiences with their audiences, while keeping creative judgement and brand control firmly with people."

That last clause is worth noting. Bannerflow positions the suite as advisory: scores and recommendations inform a human decision rather than triggering automatic changes to live creative. The release does not describe any automated optimisation that acts without approval.

What is new and what is renamed

Separating fresh capability from repackaging is difficult with this release. Creative Studio and feed-driven dynamic ads were part of Bannerflow's platform well before today. When SEGA reported in August 2026 that it had cut production time by about two-thirds across seven markets, and attributed a 44% lower CPM and 14% lower CPC to Bannerflow's Social Dynamic Ads, the feed-driven tooling was already in use.

The Creative Intelligence label is also not new. Bannerflow's own notes to editors described the company as an "AI-native Creative Intelligence platform" in the release accompanying its video data in September. That analysis, published September 14, 2026, found that 78% of more than 300 brands on the platform produced video in the first eight months of 2026, but only 46% served any video impressions. The same boilerplate appears in today's release.

The elements that read as genuinely new are predictive attention scoring with heatmaps and clarity scores, the Creative Insights analysis of element-level performance, plain-language feed configuration, and image expansion and upscaling. The release does not say whether these are generally available today, in beta, or limited to particular customer tiers. Bannerflow did not publish a rollout schedule.

Where this fits in the market

Creative tooling has moved in the same direction across several vendors over the past year, which puts Bannerflow's announcement in a crowded field rather than ahead of it.

Innovid, now owned by Mediaocean, released four AI features in August 2025, including Auto Classification, which uses image recognition to label creative assets and feed an insights dashboard - functionally close to what Bannerflow calls Creative Insights. Mediaocean later said its NIVO AI agents cut campaign setup by up to 90% in pilots with five brands and agencies.

Platforms have built similar analysis directly into their own interfaces. Meta added a Creative breakdown in Ads Manager in July 2025 that reports performance by individual creative element and compares AI-generated image variations against originals. Google's Asset Studio gained Gemini Omni video generation and one-click creative testing at Google Marketing Live in May 2026. For advertisers who run most of their spend through Meta and Google, the question is what an independent layer adds when the platforms already score and test creative natively.

Predictive scoring has also been tested publicly. A whitepaper from Kroger, Vidmob and MMA Global, covered by PPC Land on October 4, 2026, examined 1,934 video and image ads across Meta and DV360 and reported that its model predicted high and low performers on unseen data with 81% accuracy. Of 231 attribute levels tested, 105 passed statistical testing and 22 became guidelines. That study found roughly half of Kroger's historical media investment had gone to the weaker-scoring creative tier. Whatever its limits, it published a dataset size, a period and an accuracy figure. Bannerflow's announcement offers none of these for its own scoring.

Attention data is also moving into bidding. Google said in July 2026 that it would put Adelaide's Attention Unit scores into DV360 custom bidding. That is measurement of where ads run and how they are seen in market. Bannerflow's heatmaps operate on the creative file itself before it runs. The two are complementary in principle, but a pre-launch attention prediction and an in-flight attention measurement are not the same thing, and the release does not suggest any link between Bannerflow scores and third-party attention vendors.

Why it matters for advertisers

For brands that produce creative at volume across display, social, video, CTV and DOOH - the channels Bannerflow lists in its boilerplate - the practical gap Velasco describes is real. Performance reports and design tools tend to live in separate systems, and creative teams often receive feedback too late to change a running campaign. A tool that brings element-level analysis back into the place where the next ad is built addresses a workflow problem that independent research has documented.

The weak point is evidence. Predictive attention scoring is only as useful as its correlation with outcomes the advertiser cares about. A heatmap that predicts where eyes will land is not the same as a forecast of clicks, conversions or sales, and a model trained on generic eye-tracking data may behave differently for an iGaming banner than for a hotel video. That vertical concern is not hypothetical for this company: Bannerflow's own September data showed iGaming accounted for 76.1% of impressions served through its platform in 2026, a concentration that could shape any model trained on its customers' performance history.

There is also the question of where the insights end up. According to Bannerflow, Creative Insights "remain within the platform." For a customer, that creates a useful archive. It also deepens dependence on one vendor, since the evidence base built from a brand's own campaigns sits in a system the brand does not control. The release does not say whether insights can be exported.

Company background

Bannerflow was founded in Stockholm in 2010, according to earlier company materials. Today's release states that it has passed EUR 25 million in annual recurring revenue, employs more than 135 people across five international offices and counts more than 2,000 brands as customers, including Boozt, SAS, Ooni, SEGA and Meliá Hotels International. Those figures are unchanged from the company's September release.

The company also describes itself as "recognised as a category leader by G2 and Forrester." The release does not identify which Forrester evaluation it refers to. In the Forrester Wave for Creative Advertising Technologies in Q4 2020, Bannerflow was named a strong performer rather than a leader, with Innovid, Jivox, Clinch and Flashtalking in the leader category. Whether a later Forrester report placed Bannerflow differently is not established by the release.

Open questions

Several basic details are absent from the announcement:

  • When each capability becomes available, and whether any is in beta.
  • Whether Creative Intelligence carries additional cost or is included in existing contracts.
  • What data trained the attention model and how its predictions were validated.
  • Which generative model powers image creation, editing, expansion and upscaling.
  • Whether Creative Insights ingests performance data from Meta, Google, TikTok and other platforms directly, or relies on Bannerflow-served impressions only.
  • Whether insights can be exported outside the platform.

Until Bannerflow publishes validation data for its scoring, the Evaluate stage remains a claim about prediction rather than a demonstrated result.

Timeline

Summary

Who: Bannerflow, a Stockholm-founded creative automation company with more than 2,000 brand customers, more than 135 staff and over EUR 25 million in annual recurring revenue. Chief Product Officer Moses Velasco is the named spokesperson.

What: A suite of AI capabilities called Creative Intelligence, organised into four stages: image generation and plain-language feed setup (Create), predictive attention scoring with heatmaps and clarity scores (Evaluate), element-level performance analysis through Creative Insights (Understand), and feed-driven dynamic ads (Act and scale). No availability dates, pricing or validation data were disclosed.

When: Today, October 12, 2026. The survey cited in support of the launch was first published on March 26, 2026.

Where: Inside Bannerflow's platform, which serves campaigns across display, social, video, CTV and DOOH. The survey cited covered marketers in Europe, the US and Canada.

Why: Bannerflow argues that marketing teams hold more performance data than ever but struggle to turn it into creative changes, because insight and production sit in separate systems. The suite aims to bring prediction and analysis into the place where creative is built, while competitors and the major ad platforms are building similar tools of their own.