taxfix and veo ai generated ads initial review for pr teams
AI-generated video advertising uses models like Google Veo to produce video in days instead of weeks, so PR and marketing teams can test several creative variants, localize at the same time, and react to competitors quickly. Taxfix's 2025 work with Veo 2 and Veo 3 shows how generative AI compresses production timelines and creates new strategic demands for communications leaders.
Most PR teams are still debating whether AI belongs in their workflow. Taxfix has already rewired its creative pipeline around it.
What did Taxfix do with Google Veo?
The Berlin-based tax platform ran what Google documented as a large-scale experiment with Veo, its generative video model, starting early in 2025 with Veo 2 and moving to Veo 3. According to Google's Think with Google case study, Taxfix ran human-produced stories, AI-generated clips and hybrid variants against each other, including in different languages. The team wanted to know, for example, whether a creative worked better in a customer's native language or second language.
Taxfix CMO Alex Beresford told Google that what used to take six to eight weeks can now be done in a few days with Veo 2, and that the team could make several language versions at the same time. Beresford also said large parts of the audience preferred the AI-first approach, while just as many found hybrid formats the most convincing.
The campaign moved to Veo 3 in mid-2025. Creative Director Nishant John said the model still did things nobody expected, so creative flexibility was mandatory. Taxfix also ran an offline stunt, a van carrying a billboard covered in five-euro notes through Berlin, which German trade outlet Horizont reported cost roughly six thousand euros in cash and carried the message about unclaimed tax refunds.
Why should PR leaders pay attention now?
The headline is not that AI made the ads. The headline is what AI made possible operationally. When production falls from weeks to days, three things change for communications teams.
First, creative variants become testable. You can put several versions into market instead of arguing about one.
Second, localization gets much cheaper. Taxfix tested language versions in parallel, something that would have been far more expensive in traditional production.
Third, reactive PR becomes possible at a scale that was not available before. A competitor makes a misstep on Monday, and your response video can be live by Wednesday.
How does Google Veo work for ad production?
Google Veo is a generative video model that creates video from text prompts. Taxfix used Veo 2 first, then Veo 3, which creates image and sound together from detailed prompts. Beresford described a multi-stage workflow in which ideas are first sketched with Imagen 4 and Veo 2, then produced in Veo 3.
What are the limits and risks of AI-generated video?
Taxfix learned quickly that Veo struggles with physically complex scenes. John said an early script in which someone juggled chainsaws overwhelmed the model. The fix was simpler, surprising motifs based on familiar genres, which the system staged much better. Intricate hand movements and multi-object interactions are the same kind of problem.
Translation for PR teams: AI does specific things extremely fast and other things poorly. The brands winning this phase are the ones mapping that terrain honestly.
Speed also creates reputational risk. When a campaign can go from concept to live in a few days, the usual review steps for legal, compliance and cultural sensitivity can shrink or vanish. Regulated categories such as healthcare and finance face higher stakes. Audiences are also increasingly attuned to AI-generated content, and some react badly to perceived inauthenticity, so passing AI work off as traditional production is a risk if the seams show.
Finally, if every brand in a category adopts the same tools, differentiation moves to strategy and taste, which only humans supply. As John put it in Google's case study, AI is not an autopilot.
What are the three takeaways for comms leaders?
1. The production bottleneck is dissolving. Creative teams that still plan six-week timelines will be outmaneuvered by teams that plan six-day ones.
2. AI does not replace strategy, it exposes the absence of it. When everyone can produce ten variants of a creative, the brands that win are the ones with a clear point of view. Taxfix had one: people would rather do almost anything than their taxes. Its campaign name, "Everything but Taxes," says so directly.
3. Pair digital with physical. The cash van in Berlin gave the AI campaign a news hook the press could photograph.
What is the bottom line for PR teams?
For PR and marketing leaders in 2026, the Taxfix case is about speed, and about what happens to competitive dynamics when one player in a category gets much faster than everyone else. The shift is widening fastest where speed-to-market decides share of voice, such as fintech, consumer apps, e-commerce and fast-moving consumer goods. For AI Communications practitioners, the implication is direct: brands that build generative video into their workflow gain an advantage in earned-media velocity, reactive positioning and localization that slower competitors cannot match by working harder within traditional production.
FAQ: AI-generated ads and PR strategy
Can AI video tools replace traditional production agencies? Not entirely. AI excels at speed and iteration but struggles with complex physical scenes, nuanced performances and high-end cinematography. Hybrid workflows, with AI for rapid testing and human production for hero assets, are emerging as the pragmatic middle ground.
How much does it cost to run a campaign like Taxfix's? Google has not published a per-minute Veo cost for a campaign like this one, and early partners like Taxfix likely worked under pilot terms. Check current Google and competing-tool pricing (Runway, Pika, Synthesia) before budgeting.
What legal issues arise with AI-generated content? Copyright, likeness rights and disclosure requirements vary by jurisdiction. In the EU, the AI Act may require labeling of synthetic media. In the U.S., FTC guidelines on endorsements and testimonials apply if AI-generated personas are used. Brands should involve legal counsel early, and can check the current rules by platform in AI-Generated Content Disclosure: Platform-by-Platform Rules.
Should PR teams learn to prompt AI models directly? Yes. Just as PR professionals learned to write for social platforms, prompting is becoming a core skill. The ability to describe a visual concept clearly, iterate on outputs and spot when a model is failing decides whether AI speeds up or slows down your workflow.
Key takeaways
Taxfix said Veo 2 cut production from six to eight weeks to a few days, with several language versions made at once.
AI video works best on simple, genre-familiar scenarios and fails on physically complex actions like juggling.
Speed without strategy is noise: brands that win with AI creative have a clear point of view and use the tool to amplify it.
Pairing digital AI campaigns with physical activations, like Taxfix's cash van, creates news hooks and tangible PR moments.
PR teams should treat AI video tools as a core skill, similar to social media fluency, and learn prompting and output evaluation.
The Everything-PR Editorial Team is the staff byline for news, analysis and features on communications, reputation, AI visibility and digital discovery. Everything-PR has published since 2009. AI tools assist with research and drafting, and every article is reviewed by a human editor before publication. Coverage follows the Editorial Policy, and substantive corrections are noted on the article under the Corrections Policy.
