CGC Guides

AI and the Business of Video Production

By Dario Nouri and Kyrill Lazarov | Updated July 14, 2026

Will AI replace video production companies? Across two years of on-the-record conversations, from the week OpenAI’s Sora demo landed in February 2024 to working 2026 studio workflows, the archive’s answer has stayed remarkably consistent: no, but it is re-sorting the business. AI today lives overwhelmingly in pre-production (storyboards, concept art, mood boards, scratch voiceovers), it is squeezing stock footage, animation, and the cheap end of the market hardest, and the work it struggles most to touch is the reason clients hire a crew at all: real people, real moments, and the strategy that decides what to make.

This guide pulls the AI playbook out of eight episodes: the hosts’ own reckoning the month Sora dropped, a sixteen-year Toronto agency founder who calls this the biggest shift since the cell phone and builds custom GPTs, a Miami owner who keeps AI strictly in pre-production, a New York area studio that lists AI consulting partly as an SEO play, and the practical wins, warnings, and client-side constraints in between. The through line is not fear or hype. It is that adapting is a business decision, and the owners doing it well treat AI like every tool shift before it: learn it first, keep a human hand on the story, and charge for the judgment the machine does not have.

The Sora Moment: Are We Next?

Start where the industry’s stomach dropped. In February 2024, OpenAI unveiled Sora, a model that generates video from a written prompt, and within weeks Dario and Kyrill sat down without a guest to work through what it meant on Sora and the future of video production. The opening feeling was the one every creative had: the wave that had already reached writers, graphic designers, and photographers had finally hit video’s shore.

Then the frame flips, and the flip is the thesis of this whole guide. Looked at as a business rather than a threat, these are tools, and tools lower cost and expand what a small team can take on: B-roll you did not have time to grab, a second interview angle generated from a single camera, small efficiencies that compound across an entire production process. The tension between threat and tool never fully resolves in the episode, and that is honest. It has not fully resolved in the industry either.

The Biggest Shift Since the Cell Phone

For scale, take the read of someone with sixteen years and 1,500 projects behind him. Chris Stasiuk of Signature Video Group joined the show the same season, on navigating the future of content creation, and described the moment a friend sent him the famous Sora clip of a woman on a train with no context attached. His heart skipped when he learned it was AI, and he admits he lost sleep before the fear steadily turned into excitement.

His sober map of the fallout has held up: stock footage libraries and animators face the heaviest pressure, live events are relatively safe because someone still has to be in the room, and the durable human work is understanding a client’s business problem and shaping the story that solves it. Dario’s Photoshop analogy from that conversation is the useful precedent: the software democratized image editing, yet most people still cannot use it well, so trained professionals stayed essential. The strategy stays human; the trust it earns is a moat too, the same one at the center of our guide on client retention.

Where AI Actually Lives: Pre-Production

Ask working owners where AI sits in their pipeline today and the answer is strikingly consistent: pre-production, almost exclusively. Manuel Izquierdo of Mi Media Productions in Miami is the cleanest example, from balancing creativity, business, and AI: he uses ChatGPT to spell-check copy, to build detailed storyboards by feeding in talent, locations, and products, and to generate scratch voiceovers so a client can react to a script before a real voice artist is booked. Low risk, real speed, and the craft untouched.

The hosts’ own excitement in the Sora episode pointed the same direction: feed a model a client’s brand guidelines and generate a unique storyboard in motion, paired with music, so the client feels the tone before anyone shows up to shoot. Chris Stasiuk’s version is concept art in hours instead of days of paid illustration, which recently gave his team the confidence to pitch a Dune-inspired sci-fi concept they would never have risked before. The consensus prediction is that animation transforms first and hardest, because it is built from scratch rather than filmed, and an animation project is already roughly eighty percent pre-production. Cheaper concepting does not shrink the work; it moves the budget toward thinking.

For proof this is adaptation rather than fashion, take the longest lens in the archive. Dave Lindsay of Avalanche Studios has edited across three decades, watching the gear go from hundred-thousand-dollar Avid systems to a phone in your pocket, and on lessons in longevity he describes folding AI into storyboarding and voiceovers the same way he absorbed every tool shift before it: the technology serves the story, never the other way around.

The New Skill: Directing the Machine

Kyrill’s sharpest prediction from the Sora conversation is about jobs, and it starts with a precedent: autofocus quietly killed the focus puller on low-budget corporate shoots. Every leap in technology removes a role and creates a new one, and this time the new role is the prompter.

Animators especially, he argues, will rebuild their skill set around directing a model, the way the tooling itself has already moved from clipped, comma-separated Midjourney prompts toward free-flowing description. For owners, the practical implication is that prompting is a hireable, trainable craft skill now, worth the same deliberateness you would bring to any crew role and its rate: someone on the team should own it, track the tools, and spread what works.

The Economics: Cheap Content Floods the Market

Now the uncomfortable part. The biggest negative the hosts land on is not quality, it is economics: if anyone can pick up a camera and let AI stitch a story together, content gets cheap, work moves to bulk, and bulk is brutally hard to monetize in a service business. Lower barriers also mean more competition, and many clients cannot tell good video from bad, or simply buy on budget, which makes it harder than ever to explain what a video actually costs. Holding a price in that market takes the discipline from our guide on pricing video production work, and a clear story about why a full production company beats a commodity shooter.

The counterweight came two years later from the same Chris Stasiuk, back on the show for New Year’s goals, growth, and Gen AI. By 2026 he is using tools like Higgsfield AI for storyboarding and spec work, he expects AI to gut the traditional stock footage business, and yet his prediction for where the value moves is almost reassuring.

Behind-the-scenes reality, genuine stories, and the visible craft of how something was made are hard to fake convincingly. Scarcity is pricing power, and authenticity is about to be scarce.

The Client Side: Regulation, Ownership, and Disclosure

Whatever the tools can do, the buying side moves slower, and owners who sell to corporate clients need to price that in. Sora arrived in an election year, and the hosts expected exactly what followed: disclosure rules, watermark debates (a corner mark is easy to crop, metadata is easy to strip with a screenshot), and an unresolved question of who even owns fully generated content. Plenty of corporate clients already ban ChatGPT over proprietary data and treat AI video the same way, a caution that runs strongest in finance and other regulated industries, where legal teams, not editors, decide what tools touch the brand.

Platforms add another layer: TikTok already asks creators to disclose AI content and is training its own detector on those labels, and AI-generated material may not perform as well organically. None of this kills AI in client work. It just means the studio that asks about a client’s AI policy in pre-production, and can explain exactly where the tools did and did not touch the deliverable, is offering a kind of certainty the cheap option cannot.

AI as a Business Play, Not Just a Workflow

The most entrepreneurial thread in the archive is owners turning AI from a cost saver into a selling point. Gabe Nazario of Offbeat Creative lists AI consulting on his site, and on winning on trust in a crowded market he is refreshingly honest about why: yes, the team genuinely uses the tools (ChatGPT, Claude, Munch for social cuts, and Gamma, which turns a voice memo and a ChatGPT draft into a polished proposal deck), but the listing itself tells clients they are current, not an older shop avoiding the technology.

It works on both levels, the same way strong video SEO works: the capability is real and the signal wins the click, one more lead tentacle in the system from our guide on how video production companies get clients. Chris Stasiuk’s version is the custom GPT: instead of prepping a Vegas talk, he spent a week building Hook Hound, a purpose-trained model for writing hooks that wildly outperformed generic ChatGPT, proof that training the tool on the exact job is where the leverage is. And Dario’s forward bet from the Sora episode is a custom AI preset per client, loaded with their brand guidelines and look, so generated content stays consistent across a project. Consistency, not novelty, is what justifies a premium.

Adapt Without Losing Your Head

The hosts’ own operating stance, laid out in the milestone retrospective looking back and moving forward, is pragmatic rather than panicked: AI is here, it is a tool, and refusing to learn it is the only losing move. They point to the set roles most exposed (makeup, through AI cleanup in post, and audio operation) while insisting that strategy, client relationships, and taste remain human work. And they practice it: AI editing cut the podcast’s own post-production from four hours to about ninety minutes.

The caution comes from the other end of the archive, all the way back on finding work in your local community, and it has aged well. Braeden King’s early ChatGPT test drives read like Wikipedia articles with invisible references, and Kyrill flagged the strategic risk that still stands: if search engines start downranking AI-written content, why gamble a site you spent two years ranking? The wins that hold up are the small, smart ones, rephrasing titles, filling a carnet spreadsheet with lens specs on command, and Braeden’s closing worry is the right one to end on: protect your own thinking, use the tools, and stay the operator. Which is also the honest pitch to clients, who should be choosing a video partner that is visibly adapting rather than standing still, and the fastest way to test ours is a free competitive quote.

The AI Playbook

Pulled from eight conversations, in the order a studio should actually adopt it.

1. Start in pre-production: storyboards, concept art, mood boards, and scratch voiceovers, where the risk is low and the speed is real.
2. Keep a human hand on the story and the final cut. The tools draft; you direct.
3. Treat prompting as a craft skill: assign an owner, track the tools, and spread what works across the team.
4. Build one custom GPT for one repeatable job (hooks, proposals, briefs) and test it against the generic model.
5. Put your AI capability on your site, honestly. It is a positioning signal and a search play at the same time.
6. Ask corporate clients about their AI policies in pre-production, and document exactly where AI touched the deliverable.
7. Do not gamble ranked pages on AI-written content. Use it to assist your writing, not replace your judgment.
8. Bank the efficiency: when AI cuts your post time, take some of it as margin instead of racing your own price down.
9. Double down on real people and process-driven content. As the feed floods, authenticity is the premium product.
10. Budget time every week to keep up. Half the job now is knowing what tomorrow looks like.

Frequently Asked Questions

Source Episodes

Every perspective in this guide comes from an on-the-record conversation. Go deeper with the full episodes:

The Hosts

About

Creatives Grab Coffee is a podcast about the business behind video production: sales, strategy, pricing, team building, and everything that happens off camera. New episodes every week on YouTube, Spotify, and Apple Podcasts.

Lapse Productions is a Toronto-based video production company serving tech, finance, healthcare, and manufacturing clients with corporate, promotional, event, and testimonial video. New to commissioning video? Start with our guide to the types of corporate video.