At We Know Video, we use AI every day. It’s a brilliant assistant, but a terrible creative director. The trick is knowing what to hand it and what to hold onto. Here’s where we draw the line.
AI & Production
This isn’t yet another piece about AI killing creativity, or production companies defending their turf, at We Know Video, we’re AI friendly around these parts. We use it every day across scripting, concepting, image generation, editing support and motion, not tied to one tool, but across a mix depending on the job.
It might be Claude helping to better structure ideas and pressure test directions before we lock a script. Runway helping us solve something in the edit. Nano Banana or sometimes even ChatGPT generating images for concept frames. Or more recently Freepik’s node-based workflows when we need more control over outputs.
The actual tools constantly shift but the role they play doesn’t. They’re there to speed up the mechanical parts of production and open up more time for the decisions that actually make the work better. But we’ve also watched brands and businesses spend big money on video that doesn’t land, and not because the production didn’t cut it, but because AI was used in the wrong place.
That distinction matters more than almost anything else right now and people still aren’t talking about it enough. Here’s where we’ve landed: the question isn’t whether to use AI. It’s knowing exactly where to aim it.
The problem isn’t AI. It’s the brief.
Over 60% of video marketers are now using AI tools to create or edit content and that number is moving fast. And we get it. The tools are pretty impressive; audio transcriptions that used to take hours now takes seconds. Storyboard frames that once required a specific designer get generated in minutes. Rough cuts that took days get assembled automatically.
For certain parts of production, this is a huge step forward, and leaves us creatives more time to inject energy into the fun stuff. For others, it’s a shortcut that costs you more than it saves.
The issue isn’t the tools themselves. It’s that most teams are increasingly treating AI as a full replacement for the production process rather than a precision instrument for specific tasks. Essentially, if AI is doing all the thinking, it shows. If it’s doing the heavy lifting around the thinking, it works.
If it’s the former, what you end up with is content that feels instantly recognisable as AI-generated. Synthetic voice, generic visuals, no resemblance of human instinct in the edit whatsoever, and audiences pick it up faster than most brands think.
And there’s a growing disconnect between what brands think is landing and what actually is. New research from the Interactive Advertising Bureau shows 82% of ad execs believe Gen Z and Millennials feel positive about AI-generated ads. In reality? It’s just 45%. That perception gap is getting wider, not smaller. And Gen Z are the most sceptical of the lot, with 39% reporting negative sentiment toward AI ads. But ultimately, the takeaway here isn’t that AI is bad. It’s that audiences are more sensitive to how it’s used than brands think.
The question that changes everything
After working across hundreds of campaigns, some where AI was a genuine advantage and some where it was a quiet liability, we’ve landed on a bit of a filtering framework here at WKV.
The WKV Framework
Is this video’s job to build trust or to move fast?
Build trust → Keep humans at the helm
- Customer testimonials
- Brand films and hero content
- Thought leadership
- Executive communications
- Case studies with real people
- High-stakes campaign launches
Move fast → AI is your friend
- Transcription and captions
- Rough cuts and first edits
- Storyboard and concept frames
- Script research and structure
- B-roll and asset generation
- Repurposing long-form content
It sounds simple. But most briefs don’t come in with this framing. And that’s where the expensive mistakes happen.
Where we draw the line on AI
A new client came to us wanting AI-generated voiceover for a customer testimonial series. As the budget was tight, they felt like the ElevenLabs output sounded convincing enough (they’d used it before on another testimonial). But we pushed back.
Why? Because a testimonial’s entire job is to transfer trust. And trust doesn’t transfer through a synthetic voice reading a script, no matter how natural it sounds. It’ll waste everyone’s time, budget and worse, it’ll do nothing positive for your reputation (or ours even).
The audience might not always consciously clock it as AI. But something will inevitably feel off, and that feeling is the exact opposite of what a testimonial needs to do.
Instead:
- AI helped us organise interviews, shape the narrative, generate supporting assets and speed up the edit.
But the core of it stayed human. Real people stayed on camera, with their real voices, sharing their real experiences, and most importantly in the case of a testimonial; in their own words.
The result: 44% higher engagement than their previous AI voiceover attempt on Meta. Similar timeline and budget. We’d just been more deliberate about where AI sat in the process.
Where AI saves you money
Here’s what most people get wrong about AI and production budgets: the saving isn’t in replacing human talent. It’s in eliminating the wasted time around it.
Every production has a whole heap of billable hours of non-creative work:
- Transcription (woah what a drag this was pre-AI)
- Admin cuts
- Rough assembly
- Asset resizing
- Caption generation
- Storyboard iteration
These tasks used to require copious amounts of skilled time. Now they don’t. When AI handles the mechanical work, more of your budget goes toward actual creative decisions and resources that move the needle. Your director spends less time on admin and more time on the calls that make a film good.
It’s not about how to make a cheaper version. It’s how to make a better version with the same budget.
What to ask before your next brief
- What is this video’s primary job? Awareness, leads, trust, conversion. Define it clearly.
- Who’s watching, and what do they need to feel? If it involves trust or credibility, be verrryyyy deliberate about the use of AI.
- Where is the time and cost actually going? If it’s admin-heavy tasks, AI should be doing that, always.
- What needs a human decision? Creative direction and ideation (use AI to pressure test ideas but never to formulate them). Casting. Performance. Narrative instinct. This is the stuff that sets apart generative AI and top tier creative work these days.
We don’t look at these as tasks, but a judgement lens.
To rehash
AI is not going to make great video for you all by itself. At least not for a veryyyy long time. But it will make the process of creating great video faster and cheaper, without losing credibility, if you know where to aim it.
The brands getting the most out of it aren’t replacing human creativity. They’re protecting it.
Every tool has a job. And for anything that needs to build trust, we keep humans at the helm every time. Not as a philosophy, but because the audience can feel the difference.
We Know Video is a leading Australian video production company. We care deeply about the gap between content that gets made and content that means something. If you’re a brand trying to figure out where you sit in all of this, chat with us via the chat box on the right hand side of the website, or shoot us an email at info@weknowvideo.com.au.
