What happens when a professional writer who finds AI-generated content "quite horrific" decides to build an AI writing platform to help people show up more authentically on LinkedIn? With AI as his development team?

Nigel Jay Cooper found himself in exactly that position. He'd spent a decade training professionals and teams in employee advocacy to create authentic LinkedIn content and was determined to fight the flood of ‘beige’ content generated by ChatGPT et al - thoughtless, generic posts lacking personality and spark.

His first instinct was to fight against AI. But when everyone kept using it anyway, he realised he was fighting the wrong battle. The question wasn't whether people would use AI. The question was: can they use it in a way that reflects their personality, experience and humanity? So he set out to build Ghostart, even though he's not a software developer.

Listen to this episode of Curious Business as Nigel shares how he built Ghostart over the past year, and discusses:

• The beige content crisis: Why AI-generated LinkedIn posts are "soul destroying" and how people remove themselves from their own voice
• Building with AI: Using different models for different purposes, learning through trial and error, and managing the "700,000 suggestions you didn't want" that the LLMs tend to spit out
• The technical journey: From two aborted platform attempts to the approach that makes AI development productive and effective
• The Beige-ometer: Training an AI writing coach in rhythm, emotion, and storytelling - not optimising to LinkedIn's algorithm
• Why your beta test should make you feel ashamed: The essential advice that propelled the platform forward

Chapters:

0:01:06 - The Beige Content Problem

0:04:27 - Who Ghost Art Serves

0:09:10 - Building with AI - The Vibe Coding Reality

0:11:47 - Tools and Techniques

0:17:31 - Keeping AI Focused

0:21:20 - The Hardest Parts to Build

0:27:17 - The Beta Launch

0:30:53 - What Users Actually Valued

0:35:47 - Lessons and Success

Links:

More about Ghostart and try it free: https://ghostart.io

Request a demo: https://ghostart.io/teams/demo

Connect with Nigel on LinkedIn: https://www.linkedin.com/in/nigeljaycooper/

-- Thanks for listening.

Curious Business is here to bring you insights, experiences and ideas from founders, leaders and experts to help you think differently about your business and its marketing. My name is Stephen Morris and I help start-ups and scale-ups get momentum in their marketing and pipeline. Like to know more? Book a discovery call: https://www.curiousbusiness.co.uk/discovery/ Or sign-up to 'The Prompt' my monthly newsletter: https://www.curiousbusiness.co.uk/signup/ If you'd like to talk about how podcasts can build reach, enhance authority and fuel your pipeline, visit: www.curiousbusiness.co.uk.

Imagine being told someone you love is dying, and then discovering they have been trapped in a coercive and controlling relationship. You'd turn to a number of agencies to seek help, guidance and justice. But it turns out you often need to know the 'magic words' to unlock that help. In this episode of Curious Business, Stephen Midgley recounts how he navigated those challenges and leaned on Chat GPT and his background as a systems architect, to do something about it - for his mum, and now for others too. What follows is equal parts heartbreaking and baffling, as Stephen moves between police, hospitals, the care system, and other agencies. Each has their own language, arcane forms and reasons to say no, it seems.

Upon being bounced between agencies, the compassion of frontline staff was undermined by broken processes, so Stephen turned to AI. What began as a query on ChatGPT became a practical set of tools that can help others facing the same awful challenges.

Stephen describes balancing his time between building crude, command-line tools with spending precious hours with his mother. The system he’s developing, named LINDA in memory of his mum, is a translator, memory bank, assistant and also a tool to support charities working in the space. He wants to make hidden pathways visible and give people the options they often aren't told about.

What we cover:

• What can happen when you ask for help
• The "password problem": how using the wrong words can lock you out of help
• Three patterns of broken systems
• The LINDA tools: Using AI to translate policy language, surface hidden options, and compare policies against your experiences
• Why AI should augment human support, not replace it
• The 'non-crime' database field and his campaign for more transparency

Links and Resources:

• The LINDA Project: https://www.keec.online/linda/
• Change.org petition for transparency on domestic abuse non-crime closures: https://www.change.org/p/justice-for-linda-and-for-every-abuse-victim-erased-from-the-system
• Stephen's original AI for the Rest of Us presentation: https://youtu.be/CkqwNYQIh6U?si=97nHD52r_49_j2tJ

Charities mentioned:

• AAFDA – supporting families bereaved by domestic abuse: https://aafda.org.uk/

Disclaimer: Stephen Midgley is not a legal professional and nothing in this episode constitutes legal advice.

Thanks to Jari Worsley for sharing Stephen's 'AI For The Rest of Us' talk with me in the first place: https://www.linkedin.com/in/jariworsley/

-- Thanks for listening.

Curious Business is here to bring you insights, experiences and ideas to help you think differently about your business and marketing. My name is Stephen Morris and I help start-ups and scale-ups gain clarity, take action and ignite momentum in their marketing and pipeline. Like to know more? Book a discovery call: https://www.curiousbusiness.co.uk/discovery/ Or sign-up to 'The Prompt' my monthly newsletter: https://www.curiousbusiness.co.uk/signup/ If you'd like to talk about how podcasts can build reach, enhance authority and fuel your pipeline, visit: www.curiousbusiness.co.uk.