Our Blog.

Most marketing teams measure video by views. Here's how to measure the ROI of a brand film in a way your CMO and CFO will take seriously, with the metrics that matter for each goal.

How brand films get made in 2027, step by step, from the brief to final delivery, and how we use Abby, an AI producer's agent, to run the logistics of every production.

A clear video brief gets you a better film and a faster, more accurate proposal. Here's what to include, with a template you can copy for your next brand or launch film.

Met investors or buyers at TechCrunch Disrupt or LA Tech Week? Here's how to write a follow-up email that gets replies, and why adding a video makes it work better.

GPT-6 now builds a new interface for every answer. Here is what generative UI means for product demo videos, and what AI and SaaS companies should show instead.

Quick answer: Use a launch film when you need attention and belief at a specific moment, like a funding announcement, a public launch, or a rebrand. Use an explainer when buyers do not yet understand what your product does, and use a product demo when they already understand it and want to see it work. Most teams come to us knowing they need a video and not knowing which one. The three formats get used interchangeably in briefs, and they do very different jobs. Picking the wrong one is the most common reason a video underperforms, and it usually has nothing to do with how well the video was made. Here is how to tell them apart.

TechCrunch Disrupt is in San Francisco from October 13 to 15, and more than 10,000 founders, investors, and operators will spend three days meeting each other at Moscone West. Most of those conversations will be short. What happens after depends on what an investor finds when they look a company up the following week, and for most early-stage startups that means a website and a deck. We make brand and launch films for startups, so we clearly have a stake in this. But you don't have to take our word for why video matters in fundraising. Researchers have studied it for over a decade, and the findings are consistent.

Something has shifted in how the public feels about the technology industry, and the numbers are worse than most founders realize. In 2026, only about 19 percent of users say they feel excited about AI. Two years ago that figure was around 50 percent. Roughly 54 percent of Americans report outright AI fatigue. The Ipsos AI Monitor, which surveyed more than 23,000 people across 32 countries this spring, found that only 27 percent believe AI will reduce the amount of disinformation online, while 45 percent expect it to make things worse. As one Wall Street Journal writer put it, the only thing growing faster than the AI industry may be Americans' negative feelings about it. If you are building an AI company, it is tempting to read this as a cultural mood that does not touch your business. It does. Two findings should stop any founder in their tracks. Nearly half of consumers now say they prefer brands that avoid using generative AI in customer-facing content. And around 52 percent say they reduce their engagement when they suspect content was AI-generated. That is not a vibe. That is a conversion problem sitting in your funnel.

Here is a pattern we watch play out constantly. A founder invests real money and real attention in a launch film, gets something they are proud of, posts it once on launch day, and waits. A few hundred views trickle in from people who already follow the company. The film was good. The launch was quiet. And the conclusion the founder draws, wrongly, is that video did not work. The film was never the problem. The distribution was. A launch film is not a deliverable you publish, it is source material you deploy, and the founders who understand that difference get an order of magnitude more out of the same asset.

Polaris is a creative companion for people who build imaginary worlds. Novelists, screenwriters, songwriters, concept artists, indie game developers. You talk to it like a collaborator at 2am, and it listens, asks the questions you missed, and remembers your characters, your lore, your voice. Whatever you make stays yours, end-to-end encrypted, never sold, never used to train AI. When the team came to us for a launch film, we hit the problem that defines most software launches. Almost none of that is filmable. There is no interface worth shooting, no feature that photographs, no moment where the product does something a camera can capture. The thing Polaris actually sells is a feeling: privacy, and the trust that comes with it. Our job was to make that feeling legible, and to make it for an audience of working creatives who can spot a generic SaaS ad from across the room. Here is how we did it, and why the approach matters for any founder thinking about how to launch.

Almost every founder we film is compelling in a meeting and stiff in the first take. The same person who can hold a room for an hour, unscripted, goes rigid the moment a lens is pointed at them. Their vocabulary changes. They start using words they would never say out loud to another human being. Somewhere around take four, they relax, and the good material finally arrives. This is so common that it is basically the default. It is also worth understanding, because the cause is not nerves and the fix is not confidence.

The most powerful video a startup can make in 2026 is also one of the simplest. It is the founder, on camera, telling the truth about why the company exists. In a year when nearly everything else about video has gotten more automated, more abundant, and more synthetic, that one thing has quietly become the hardest to fake and the most valuable to own. The market is making this obvious. Look at what is actually working right now and a clear pattern emerges. Audiences have stopped rewarding polish for its own sake and started rewarding real people with something real at stake. The trend reports coming out this month keep landing on the same blunt advice for founders: put an actual person, with actual conviction, on camera. As feeds fill with generated content and buyers increasingly form opinions inside AI answers before they ever reach your site, human authorship has turned into a competitive signal. Trust has become a distribution factor. People share, cite, and buy from companies they believe, and belief is far easier to build from a face than from a logo.

Meta has rolled out expanded disclosure labels for advertising on Facebook and Instagram, and while the change looks small on the surface, the direction it points is worth every founder's attention. Here is what actually happened. Ads created or significantly edited with generative AI now carry a disclosure inside the "About this ad" panel, the menu users reach by tapping the three dots on a promoted post. Meta applies the label automatically when advertisers use its own AI features, like background generation, image generation, and animation, and it detects third-party tools such as Photoshop and DALL-E through the C2PA metadata standard. It extends the AI labeling Meta already uses on organic posts into paid advertising, and it comes with a disclosure control in Ads Manager that advertisers are expected to use when their creative contains AI-generated or AI-altered content. Taken alone, a note buried in an info panel is not dramatic. Taken as a signal of where the whole system is heading, it is.

There is more venture money in the market than almost any point in history, and for most founders it has never felt harder to reach. Both things are true at once, and understanding why is the difference between a raise that closes and one that stalls. The headlines make the abundance obvious. This month alone, the video-AI company Twelve Labs raised $100 million with Amazon and Nvidia behind it, and that barely registered against the mega-rounds now setting the tone: OpenAI reportedly closing a raise near $110 billion, xAI pulling in $20 billion, Baseten taking $1.5 billion in its fourth round in 18 months. The money is real. The problem is where it goes. By recent Crunchbase figures, close to 88 percent of AI startup funding in 2026 went to US-based companies, and an outsized share of that flowed to a small circle of giant names. If you are not one of them, you are not raising in that market. You are raising in a much more selective one, and it runs on different rules.

The 2026 marketing playbook is easy to summarize. Use AI to make ten times the video at a tenth of the cost, post daily, test everything, and let volume win. The tools are real, the savings are real, and the pressure to do this is real. It is also, for a growing number of brands, starting to backfire. The interesting thing is that the entertainment industry, which has every incentive to chase the technology, is quietly arriving at a more nuanced answer. This year's major trend reports have converged on a single theme. Dentsu titled its 2026 media outlook "Human Truths in the Algorithmic Era." EY's framing centers on authenticity and experience. One widely cited report called 2026 the year the industry stopped chasing technology and started redefining meaning. None of these are anti-AI. The same studios and agencies writing them are signing AI deals as fast as they can. A major Hollywood agency just partnered with an AI firm to help its clients tell new stories, and even Getty, which once sued over AI, is now licensing hundreds of millions of assets to OpenAI. The message is not reject the tools. It is stop letting the tools set the strategy. For a founder deciding how to use AI video, that distinction is the whole game.

Two stories landed within days of each other this past week, and read together they map the fork the entire video industry is now standing at. In one, Google's DeepMind invested about $75 million in A24, the independent studio behind Moonlight, Everything Everywhere All at Once, and the recent hit Backrooms. In the other, ByteDance unveiled Seedance 2.5, a model that generates a 30-second commercial in native 4K for nearly nothing. One company paid a fortune for access to how great films get made. The other made the making of them look almost free. The distance between those two bets is the most useful thing a founder can think about right now when deciding how to make their own video.

This week the advertising industry gathered on the French Riviera for Cannes Lions, the festival that has always worked as an early read on where the business is heading. For two years running, the conversation there was dominated by what AI could do. This year the mood shifted, and the shift is worth paying attention to if you are a company deciding how to make your next video. The headline from Cannes 2026 is not that AI has taken over creative work. It is that the industry has decided where AI belongs, and where it does not. As generative tools make video faster, cheaper, and more abundant than ever, the thing that has become scarce, and therefore valuable, is the human judgment and authenticity that no model can produce. The festival gave that idea a name it did not have before. Call it the authenticity premium.

Startups spend months, sometimes years, building a product. Founders refine the features, teams polish the website, investors pore over the deck. Then, a few days out from launch, someone in a meeting asks the question nobody planned for: do we have a launch video? What follows is a scramble. One of the most important assets the company will ever publish gets built in a week, under deadline pressure, with no time to think. We see it constantly. The video usually gets finished and the date usually gets hit. What suffers is everything that actually makes a launch video work.

The AI video space moves fast enough that a ranking written in March is usually wrong by June. Models leapfrog each other monthly, pricing shifts, and the leaderboards that crown a winner one week reshuffle the next. So treat this as a snapshot, accurate as of mid-June 2026, of where the major tools stand and how the market is reacting to the newest arrival. That arrival is Google's Gemini Omni, which reset the conversation when it rolled out more broadly this month. It is worth understanding what it actually does, why the reaction has been so loud, and how it stacks up against the established players: Runway, Kling, Higgsfield, Veo, and Sora.