top of page
2-8-1 Complex exposure histories __ __Surface exposure dating, erosion rates and burial da

AI meets Media
How we are using far

Reduce Costs

A quicker start - we are creating first drafts of scripts from website urls, existing documents, interviews, images and existing media.


Finding the good bits fast - AI transcription means we find the strongest soundbites and comments sooner - it’s also  collaborative.  We’ll share a page with you to review and highlight your selections and ideas.

Script Writing - we're co-piloting with our own custom 'fine tuned' GPTs meaning you get a polished final draft in an appropriate tone and style quicker.  They also offer more creative choices and decisions - leading to better content.


Existing assets - we interrogate them, capturing valuable knowledge, insights and experience. Efficiently, repackaging them in different formats, styles, lengths and media.


A 40 page pdf can be turned into bite sized learning,  perhaps as 8 x 2min. videos or 4 x 8min. podcasts.


A dense information heavy website can efficiently become a video or a podcast.

We analyze existing media  to identify trends and  themes and generate new content that works for today's audiences. Sometimes known as data-driven content creation.



Turning content into conversation is perhaps the biggest change in the way we have been working.   Creating knowledge bases from existing media assets and turning them into new interactive resources that users can talk to.

    We’ve created 2 so far.

  1. Turning 12 episodes of a podcast into a knowledgeable personality you can quiz. This podcast covered volunteering during the first 4 months of the Covid 19 Pandemic.  This chatbot is now the talking version of the podcast able to discuss all the stories, facts and analysis contained in the 12 x 20 minute episodes.  Get answers to questions like ‘what was the main type of volunteering undertaken’’  uncover stories by saying ‘tell me how older isolated people were supported’  As the chatbot is only trained only on the podcast content all the responses are factual, accurate and reflect the breadth of experience across Scotland in those difficult first four months.


  1. We created an expert teaching assistant from an extensive E-Learning course on Geochronology.  After collating and configuring a knowledge base from all the videos a custom GPT was guided and briefed.  The result - a learning assistant that can produce revision aids, explain concepts, produce downloadable resources and test the student to reveal knowledge gaps.



Got an idea? 


If these techniques has sparked an idea.  Get in touch to have a conversation on how they might work for what you have in mind.


bottom of page