KEY POINTS: - We solve the AI implementation gap - transforming teams from AI-curious to strategically confident in days, not months - Personal trainer approach for AI adoption - hands-on practice with your real work, not theoretical training - People+AI methodology - amplifying human value rather than replacing teams - Three-tier service model: from single workshops (immediate impact), to workflow and process redesign, to full department transformation (3-15 months) - Proven results with 1,000+ people globally, 10/10 satisfaction scores, working with approved enterprise tools

What is a GPT and why does it matter?

Do you know what GPT actually stands for? Thought so. It's the abbreviation for 'Generative Pre-trained Transformer'. Are you any the wiser now? Thought so. Let's break it down...

📰 TLDR;

  • 1/ Generative: It can create or 'generate' new content like text, stories, images, visuals, audio, or answers to questions. Which is pretty cool. These things have always been represented through movies and shows of our imagination, from Knightrider, to Short Circuit, Star Trek, and so much more. And in real life, predictive text, computer translations from one language to another, things like Siri, etc. This action has been around for a while.

  • 2/ Pre-trained: Before you even start chatting with something like ChatGPT, it's already been 'taught' a lot by 'reading/understanding' loads of information (known as Large Language Models or LLM for short), kind of like how you learn things in school before you do your homework. These LLMs are humongous data dumps. People have then spent the time to 'train' the Ai to make the 'correct' predictions. It's just like a giant game of Guess Who, where the Ai pretty much guesses it correctly. However, because it's created by us people it's prone to a few issues here and there from biases and other things. Also, the data set might be corrupt in some ways such as skewed masses of certain samey stuff, which then means the output spews that out. For example, if you ask an Ai to generate a beautiful person image, more likely it'll generate a slim, young, white woman, usually a bit undressed and buxom. And just to make it a bit more fun, sometimes the Ai if it can't find the answer, will just output the wrong answer, these are called hallucinations. All important stuff to be aware of.

  • 3/ Transformer: Not Optimus Prime or Megatron (yet). This is a special way it learns and understands language, helping it make really good guesses about what to say next based on inputs > which are decoded > then encoded and transformed > into the output related to the ask, based on the best guess.

So, that's what a GPT in a nutshell. As ChatGPT told me, a GPT is like a well-taught student who's really good at making up stories and answers based on many things it has already learned! I hope you found this useful. I'll start sharing more Ai hints, tips, insights and updates regularly on LinkedIn and connect them to what they could mean to our world of communications. Stay tuned.

Frank Dias 8-bit profile image wearing glasses

Frank Dias

An employee comms pro with 15+ years of exp across diverse industries.💗s crafting impactful strategies, narratives, & experiences aligned with business priorities. Specialises in🌍IC &engagement, change mgt, & business partnering.

Follow Frank on ➡️LINKEDIN

At Ai x (by) Comms Lab, we help teams practically master strategic AI use for business impact. We teach you how to fish🎣 with AI🤖, so you can fish to bring your strategic use of AI to life with guidance and support on your journey.

Are you ready to transform your team into strategic AI masters? Get in touch ↖️

LinkedIn icon button to visit the Ai by Comms Lab company page

➡️ Subscribe to my newsletter /REMIXING comms

Practical AI news, insights, learning & more curated & filtered for your inbox.

© 2025 Ai x Comms Lab