Practical Uses of AI: Part I - Short Midjourney Product Video Process

In the first part of our new series on practical uses of AI we share how we’ve been using Midjourney to turn static product images into short-form video content. It’s a technique we’ve used both for our own ventures and on several projects for paying clients - and it’s proved both a cost-effective way to increase engagement and a fiscally shrewd means of extracting additional value from existing libraries of image assets.

These ‘moving pictures’ have really caught on…

As anyone managing company social media accounts in 2025 can attest, the quest for attention online has become something of a thankless grind. The markets for consumers’ are saturated and highly competitive. Putting out content that algorithms will surface ‘organically’ is tougher than ever, with Big Tech pushing brands ever more aggressively towards their paid acquisition channels. Add to that the inexorable march of video across everyone’s feeds and you’ll understand why communications professionals pull their hair out when asked to create engaging content armed only with Colin’s grainy JPEGs from a trade show in Milton Keynes and some boilerplate platitudes from ChatGPT…

This year at Blackdown we’re seen video content gaining on average 4X the engagement as static image content across platforms. The message - from both audiences and algorithms - is clear; people like to watch videos and the feeds of content they consume will continue to be weighted heavily in video’s favour for the foreseeable.

Video ain’t cheap, or…is it?

There’s no getting around the fact that good video content is rarely inexpensive to produce. For SMEs in particular - whether operating in B2C or B2B markets - it’s often unfeasible to regularly pump out enough new video content to keep algorithms / audiences satiated. Some small teams might be blessed with members both gifted (and willing) to get in front of a smartphone camera on a regular basis, and in some verticals that lo-fi aesthetic can work well if you’ve lucked out on in-house talent. For the rest of us - particularly those who sell to other businesses - it’s increasingly hard to avoid the curse of pumping out staid and scripted corporate content that serves little purpose beyond box ticking.

When professionals are used - whether that be script writers, actors, camera operators, editors, colourists or one of the myriad other professionals you’ll find in abundance here in the UK - even a short shoot will quickly run to a price tag of many thousands of pounds.

To be clear, we still think that this is money that can be very well spent - but recommend brands save their powder for delivering the most important messages with maximum impact.

AI video hype

The high cost barrier to generating professional quality video has driven billions in investment in AI companies specialising in video generation recently, with investors desperate to back software-as-a-service (SaaS) starts who might be able to offer a more cost effective, ‘synthetic’ alternative to the traditional processes of video production.

London’s Synthesia have raised $330 million of investment for their text-to-talking-heads (or ‘avatar’ as they prefer) business, aimed squarely at the corporate market (think How-Tos and FAQs from text prompts). New York’s Runway ML have pulled in over $535 million and have a broader offering, with everything from text-to-image generators, image-to-video functionality and even an array of AI enhanced video editing tools.

But what can you actually use any of this for in a business context? These generative tools - and a plethora of others like ChatGPT’s Sora 2 and Google’s Nano Banana - are undoubtedly impressive in their ability to conjure moving images seemingly from nowhere. But if you’re not the coolest creative agency in Soho with free rein to experiment on a brand building project lavished with cash - what can most businesses and agencies actually use generative video tools for, today?

An image-to-video workflow, that (mostly) works

Recently we’ve found a generative video technique that’s actually proved useful, and we’ve come back to it several times for both in-house and client projects. It’s best suited for companies with a plethora of image assets to hand but next to no budget for shooting fresh video content.

We use a Midjourney Standard plan @ $24 USD (c. £21 GBP) / month to transform images like this…

…into short product videos like this:

Here’s another example from one of our other products from that same range - organic cotton vegetable bags. We spent an awful lot of time in our home studio staging and laying out these flatlays…

…so getting some extra mileage out of these assets by turning them into short videos like this was particularly rewarding. Many companies don’t retain 24/7/365 access to studios kitted out for high quality product photography / videography, so being able to reuse assets they might’ve commissioned third parties to shoot, at considerable expense, can help stretch marketing budgets that little bit further:

It’s worth noting a few things when creating your own product videos using this method:

• It rarely works on the first try - expect to go through a few iterations, but be mindful of credit usage within your plan

• Set your product image to be the ‘Starting Frame’

• Keep your text prompt nice and simple - “slow zoom out” for example

• Stick to product shots - where people are involved the model still tends to get a little overly ‘creative’

• Keep ‘Motion’ set to ‘Low’

If you want advice, or for us to handle similar requirements for you, please don’t hesitate to get in touch.

Images and videos created on Midjourney remain the property of the account holder even after a subscription has ended, and are (with a few exceptions) licenced for commercial use. We’d recommend checking Midjourney’s full Terms of Service for applicability to your own use case.

For each creation you’ll be presented with four options by default (this can be changed):

Simply choose which you like best or if none of them are quite right try a) rerunning the same creation - results can differ dramatically from one run to the next or b) tweaking your prompt if the results aren’t quite what was expected.

You’ll also find options to extend beyond the standard c. 5-second video duration. At the moment we usually trim them down a bit to 3-4 seconds, and are experimenting with longer form content created from editing multiple clips together.