In A(i) Nutshell

Andrew Miles Davis takes you back 20 years to MySpace UK, when people asked whether social media was a fad and why he left the BBC to join a website. He explains why he has almost identical conversations about generative AI today, and introduces the Story Lens method, which he uses to understand change, make sense of complicated situations and make better decisions. He compares the hero's journey, the story most of us know from Disney films, The Hunger Games and The Devil Wears Prada, with the stranger comes to town story, using Shawshank Redemption, Ted Lasso and the Banks family in The Fresh Prince of Bel Air as examples. Andrew argues that social media and AI are both strangers coming to town. When the rules change, people divide, we discover how we really work, mistakes are exposed, and decisions have to be made. The people who benefit most are the ones who experiment, adapt and build a new normal. Andrew leaves you with two questions. What has changed in your working world because of AI, and what decisions have you been putting off? Subscribe to In AI Nutshell for practical AI ideas for marketers and content creators, every day in ten minutes.  

Andrew Miles Davis covers a week of AI news that is mostly about work. He starts with ChatGPT's new image update that turns prompts or photos into animated GIFs. Then he looks at four studies on AI at work. Nearly one in five workers laid off because of AI have been asked to return, 32% of organisations that cut staff because of AI had to rehire, and nearly half of those workers want a significant pay rise. Atlassian research found that 82% of AI users say it has taken on tasks they actually enjoy, and another study found 76% of workers have asked AI a question they were too embarrassed to ask a colleague. Andrew asks whether we are asking what AI can do rather than what we want it to do. He also covers Apple's new 18+ age rating for dating apps and AI companions with romantic conversations, Jermaine Dupri challenging Spotify's planned AI remix tool, Grok users finding their usage limits cut, and Netflix launching an AI feature to help you choose what to watch. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.  

Andrew Miles Davis builds this episode around a line from a workshop attendee, who said she uses AI every day but does not feel she is good at it. He breaks down five mistakes he sees, which come down to searching, accepting, producing, blaming and misusing. First is using AI like Google, with one chat for unrelated questions, and why that confuses the model and can use up credits. Second is accepting the first answer, and Andrew explains his AI sandwich, where you add the first 15%, AI does the middle 70% and you finish the last 15%. Third is using AI mainly to produce content, and how his AI Essentials show there is far more to do with AI, with the best results coming from people who already know what great looks like. Fourth is blaming the tool when it does not understand you, and why taking responsibility for explaining yourself gets better results. Fifth is knowing how to use AI but not where to use it, and how to hand your admin to AI so you can spend more of your day on what you love. Subscribe to In AI Nutshell for practical AI ideas for marketers and content creators, every day in ten minutes.  

Andrew Miles Davis answers three questions from his training sessions in this fortnightly FAQ episode. First, does AI make us lazy? Andrew says yes, but not automatically. Using AI to remove repetitive admin is just saving time, but outsourcing your thinking is different, like a calculator you can no longer do without. Second, will AI steal my job? Andrew says AI will take out a lot of jobs, as technology always has, but right now it feels like it is taking tasks and changing jobs. He explains why the best AI results come from people who already know what good looks like, and why you should bring your experience to the table. Third, how do I write better prompts? Andrew compares prompting to martial arts belts and shares GCSE, a simple framework of goal, context, sources and expectations. He also explains why he now thinks prompt endings matter more, and where to find his list at AndrewMilesDavis.com/prompt-endings. Subscribe to In AI Nutshell for the answers to the questions your team is asking, every other week in ten minutes.  

Andrew Miles Davis shares three cool tools for Cool Tools Tuesday. First is a big update to Granola, the note taker that does not need an invite to your call. Granola has now added voice dictation, which takes it head to head with Whisper Flow, the tool Andrew uses more than any other because of his RSI. Andrew explains why he might have to cancel his Whisper Flow subscription if Granola does everything he needs. Second is a blast from the past, Ideogram, which has launched a vector feature. You can create SVG files from text, edit one specific area of an image or logo without changing the rest, and the quality stays sharp when you zoom in. Andrew explains why this matters for logos, why other image tools often change things you did not ask for, and how long he thinks the feature will stay unique. Third is Athena, a tool that helps you stay focused on one goal by asking questions, giving morning briefings and keeping a to-do list, and why it works best when you focus on one project. Subscribe to In AI Nutshell for three cool tools every Tuesday in ten minutes.  

Andrew Miles Davis shares what he is looking forward to in November. First is the new Grand Theft Auto, which he has waited over a decade for. Andrew explains where he thinks AI fits in, with a theory that the long wait was about getting to a point where AI can help players build and share their own worlds, like a full version of South London, for years to come. Second, and his top pick from a work point of view, is the personal agent space, which he thinks will shape his 2027. He shares how he has been using Muse, how his use has changed to a collective brain approach, and why knowing what he wants to do makes it easier to make better decisions. Andrew also explains why he sees personal agents as the next evolution of generative AI, rather than a replacement for large language models. Finally, he looks forward to November being the last full month of work until February or March, and why he prefers a full month as a creature of habit. Subscribe to In AI Nutshell for what is coming next in AI, every day in ten minutes.  

Andrew Miles Davis covers a quiet week of AI releases with some big stories underneath. He starts with the AI graveyard, a running list of AI startups that have failed, including Relay and Character AI, and why a product that is only one feature struggles once ChatGPT or Google adds it for free. He covers O'Reilly research showing employee AI use is moving faster than training and governance, with 39% of people saying they acted on an AI answer that turned out to be wrong and only 24% always checking first. He also covers a report that 59% of marketers say admin has increased in the last three years. Andrew explains Anthropic's new Smart Reports feature for Claude, which lets employers analyse how staff use Claude at work, and his simple rule that if it is a work account, assume work can see it. Then he looks at Bank of America's email asking customers not to share banking passwords with AI assistants, and why responsibility stays with you. He finishes with the growing list of sites, including Yelp, airlines, banks and Amazon, that are blocking AI agents, and what that means for the personal agent race. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.  

Andrew Miles Davis shares the AI tools he used the most in October. He leaves out ChatGPT and Whisper Flow because he uses them every month. First is Claude, which he says has been a better tool than ChatGPT this month for strategic work, Claude Design and building artifacts, although he finds the credit limits really annoying compared with ChatGPT. Second is Meta Muse, a personal agent he has been using to research subjects and clients, with six or seven use cases lined up. He explains why you need to know what you want from a personal agent before you spend credits and time, and that he plans to try Grokbot for a month. Third are two note takers, Fathom and Granola, which he has connected to ChatGPT and Claude to find patterns across 200 to 300 sales and training calls. His plan is a report on what is really happening in the room when it comes to AI at work. Subscribe to In AI Nutshell for the AI tools that matter to marketers and content creators, in ten minutes.  

Andrew Miles Davis shares three prompt endings, the short lines you add at the end of a prompt to get better answers from AI. First is "turn this into a fill-in-the-blank template I can reuse", which turns a one-off answer into a tool, like a meeting follow-up email or a proposal structure. Second is "focus only on" one angle and "ignore everything else", which stops answers wandering, whether you use it on a customer survey or to see an argument from the other person's side. Third is "summarise the core argument in one sentence before giving the details", which tells you straight away whether a long report is worth reading. Andrew also explains why he now thinks the ending of a prompt matters more than the beginning, and how it takes you from white belt to black belt in prompting. You can find his full list of prompt endings at AndrewMilesDavis.com/prompt-endings. Subscribe to In AI Nutshell for prompt tips you can use today, in ten minutes.  

Andrew Miles Davis shares three cool tools in week 85 of Cool Tools Tuesday. First is Editly, a writing tool that corrects, edits, rewrites and translates your text, with a focus on humanising AI writing. It shows you what it changed and why, gives you a clean version to copy, and even lets you send your text to a human editor. Andrew explains why he still thinks the human side matters in writing and creativity. Second is UseToby, which prepares you for a job interview from a link to the role and your CV. It researches the company, checks recent news and reported interview questions, and gives likely questions with a difficulty rating. Andrew connects it to media training and the idea of having three things you always want to get across. Third is Zight, the screen recorder Andrew has used for around ten years, which has AI features including transcription you can ask questions of. He explains why he prefers it to Loom, and mentions Screen Studio for tutorials. Subscribe to In AI Nutshell for three cool tools every Tuesday in ten minutes.