A Brutally Honest Review of Runway ML for Small Agencies

You’re drowning in content requests. Every client wants more video for their social feeds, their ads, their website. But traditional video production is a time and money sink. A simple 30-second social clip can mean hours of shooting, editing, and finding stock footage, blowing your retainer budget for the month. Runway ML promises to fix this by letting you generate video from text or images. This isn’t about replacing your creative team; it’s about giving them a tool to stop wasting time on low-value, high-volume video work so they can focus on the projects that actually matter.

What Most Small Agencies Get Wrong About AI Video

The first mistake everyone makes is thinking these tools are magic. You type “cinematic shot of a coffee shop” and expect a masterpiece. That’s not how it works. You get a weird, wobbly, slightly-off video of a coffee cup melting into the table. The hype cycle has sold a dream that the technology can’t quite deliver on yet.

It’s a tool, not a magic wand.

Treating it like a one-click solution is the fastest way to get frustrated and cancel your subscription. AI video generation is a new skill. It requires learning how to write effective prompts, how to chain different tools together, and most importantly, how to edit and clean up the output. You can’t just hand it off to an intern with no guidance and expect good results. The agencies seeing success are the ones that invest a few hours in building a new workflow around the tool. They don’t see it as a replacement for an editor, but as a force multiplier for a creative who understands composition and storytelling. It’s for generating unique b-roll, creating animated versions of static assets, or quickly visualizing a concept for a client before you commit to a full-scale shoot. It’s not for shooting your client’s next Super Bowl commercial.

Another major misstep is ignoring the “AI look.” Let’s be honest, much of the output has a distinct, slightly surreal aesthetic. It’s getting better every month, but you can still spot it. Trying to pass off a raw AI-generated video as a real-life shot will make your client look cheap. The key is to either lean into the aesthetic for creative projects or use the tool for clips where realism isn’t the primary goal. Think abstract backgrounds, animated logos, or stylized product shots. Don’t try to make a realistic video of a person speaking. It will look creepy, and your client will hate it.

A Real-World Workflow: Idea to Ad in 30 Minutes

Let’s get practical. Say a local real estate client needs a quick 15-second Instagram Story ad for a new luxury condo listing. Before, this would mean a trip to the property, a few hours of shooting, and then editing. Now, you can do it from your desk. Here’s a workflow that actually works:

1. Source Your Static Image. You already have high-quality photos of the property from the photographer. Pick the best one, maybe a wide shot of the living room with a great view. This is your foundation.
2. Animate with Image to Video. Upload that photo to Runway’s Gen-2 model. Instead of a text prompt, you’re using an image prompt. You can use the motion brush to specify which parts of the image should move. For example, you could highlight the clouds outside the window and prompt for “subtle cloud movement, gentle sunlight shifts.” This creates a cinemagraph effect that’s far more engaging than a static photo. Generate a 4-second clip.
3. Create a Panning Shot. Now, use Runway’s ‘Infinite Image’ tool. This is my favorite feature. You upload the same photo and literally drag a frame across it to create a smooth pan from left to right. It’s like having a camera on a dolly. Generate another 4-second clip of this pan.
4. Generate Abstract B-Roll. You need a third clip for texture. Use a text prompt in Gen-2 like “close up on luxury marble countertop, soft focus, warm lighting.” This gives you a high-end, abstract shot that establishes a mood without needing to be photorealistic.
5. Assemble and Edit. Drag all three clips into Runway’s built-in video editor. It’s basic, but it’s enough for this. Add simple cross-dissolves between the clips. Use the text tool to overlay the property address and the agent’s contact info. Add a royalty-free music track from their library.
6. Export and Deliver. Export the final 12-second video in the 9:16 aspect ratio for Instagram Stories. The whole process, from uploading the first photo to exporting the final video, can take less than 30 minutes. You’ve just delivered a valuable asset to your client in the time it takes to drink a cup of coffee.

This is the kind of practice automation that makes a difference. It’s not about firing your video team; it’s about letting them produce five of these small assets in the time it used to take to produce one, freeing them up for the bigger, more strategic projects.

So, Is the Output Actually Usable for Client Work?

Yes, but with major caveats. You are not going to create a feature film with this. You are not going to create a hyper-realistic product demo where every detail needs to be perfect. The current state of the tech is best for short-form content, social media ads, and conceptual storyboarding.

My biggest gripe is the consistency between generations. You can use the exact same prompt and seed number (a setting meant to reproduce a result) and get slightly different outputs. This makes creating a longer, coherent scene from multiple short clips a massive headache. It’s frustrating when you get a perfect 4-second clip and the 4-second extension you generate has the main object subtly change shape. This is a real problem for anything beyond simple, disconnected shots.

However, the ‘Infinite Image’ feature I mentioned is brilliant. Being able to expand a static image into a wide, panning shot has saved me from so many boring client-provided photos. It’s perfect for creating dynamic motion for websites or social posts from assets that would otherwise be completely static. This one feature alone is almost worth the price of admission for an agency that handles a lot of social media content. It’s a concrete love of mine because it solves a real, everyday problem instantly.

The output is most usable when you’re not aiming for perfect realism. Animated logos, data visualizations, abstract background textures, and dreamlike sequences are where it excels. If a client sells something like artisanal candles or craft beer, the slightly surreal, artistic vibe of AI video can actually be a perfect fit for their brand. If your client is a law firm, maybe stick to traditional video for now.

How Does Runway ML Pricing Stack Up?

Let’s talk money. Runway has a few tiers, but for any serious agency work, you can ignore the free one. The free plan gives you 125 credits, which sounds like a lot until you realize a single 4-second HD video clip can cost 5 credits. You’ll burn through that in 20 minutes of experimenting. Honestly, the free plan is a joke for anything other than just seeing what the interface looks like.

The real starting point is the Standard plan, which is $15 per month, or $12/user/month if you pay annually. This gets you 625 credits per month, which is enough to produce a decent number of short social videos. For what you get, $12 a month is an absolute steal. A single video clip from a premium stock site can cost hundreds of dollars. This plan lets you generate hundreds of unique clips for the price of a sandwich. It’s a no-brainer for any agency that wants to start experimenting with AI video.

The Pro plan is $35 per month ($28/user/month annually) and gives you 2250 credits. I think this is overkill unless you have a dedicated staff member who is spending several hours a day in the platform. The biggest jump here isn’t just the credits, but the access to more advanced features and faster processing times. If you find yourself constantly waiting in a queue to generate your clips on the Standard plan, it might be time to upgrade. For most small agencies, starting with Standard is the right move. You can always upgrade if the ROI is there.

Copyright, Ownership, and Client Transparency

This is the part everyone skips, and it’s the part that can get you in trouble. Who owns the video you create? According to Runway’s terms of service (as of early 2026, and you should always check the latest), users own the content they generate. This is great news. It means you can use it for commercial projects for your clients without worrying about Runway coming after you for a cut. However, this comes with a huge asterisk.

The AI was trained on a massive dataset of images and videos from the internet. There is an ongoing legal debate about whether this constitutes copyright infringement. While the current consensus and legal precedent seem to be leaning in favor of the AI companies, it’s a gray area. My advice: be transparent with your clients. Let them know you are using AI tools to create certain assets. Frame it as a benefit, a way to produce more content for them efficiently. Don’t try to pass it off as traditional video production. This protects you and manages their expectations. Also, I would be very cautious about using it to generate images of celebrities, recognizable characters, or specific artist’s styles. That’s just asking for a cease and desist letter.

The Bottom Line: Real Hours Saved

So, does this actually save your agency time and money? Let’s do some simple math. A junior designer or video editor at your agency might have a blended rate of $50/hour. A simple task like creating three unique video backgrounds for a client’s social media campaign could easily take two hours using traditional methods (searching stock sites, simple editing, formatting). That’s $100 in labor costs.

With Runway, a creative who knows the platform can generate, edit, and export those same three backgrounds in about 20 minutes. The credit cost would be negligible, maybe 30 credits, which on the Standard plan costs about $0.58. You’ve replaced a $100 task with a $0.58 task that takes one-sixth of the time. Do that just once a month, and you’ve already paid for your annual subscription five times over. In reality, you’ll find dozens of these small use cases every week. It’s a tool that directly reduces your cost of goods sold and frees up your most valuable resource: your team’s creative time.

This isn’t about futuristic AI hype. This is about practical, front-office automation for your creative work. It’s a tool that helps you serve your clients better and more profitably. It has its flaws, and it requires a new way of thinking, but for a small agency looking for an edge, it’s one of the most valuable subscriptions you can buy.

If you want the deep cut on this, deeper coverage of AI agent platforms.

Install it yourself in an afternoon if you’re comfortable running commands, or add the $1,500 install-support tier and we set it up for you via screenshare. Full package at deepusecase.com/vault.