In A(i) Nutshell

Andrew Miles Davis breaks down a productivity system he has used and taught for years but never fully explained in detail, built around four decisions for anything that lands on a to-do list: do, delay, delegate, or delete. He draws on Stephen Covey's urgent and important matrix to explain how items within the do category still require prioritisation, using his own test of what would not let him sleep at night. He argues delay is often mistaken for procrastination but works as a genuine productivity tool when paired with a system that guarantees the task actually resurfaces later. Delegation gets a modern update, covering both handing work to people and increasingly to AI, with Andrew encouraging listeners to practise identifying which steps of a workflow genuinely need human involvement now, ahead of a future where AI agents handle far more autonomously. He closes on delete as the hardest and most disciplined choice, sharing his own decision years ago to deliberately drop Google AdWords from his skillset entirely rather than trying to do everything. Subscribe to In AI Nutshell for daily ten-minute episodes covering how to actually think about productivity in an AI-driven world.

Andrew Miles Davis covers a genuinely massive week in AI news, opening with OpenAI's double release of ChatGPT 6 Astra, already being called the most intelligent and aligned model yet for coding, scientific research, and complex agent tasks, alongside ChatGPT Images 2.5, which promises 50% faster generation, sharper detail, and significantly improved multi-turn editing that no longer distorts faces across repeated prompts. He covers Apple's AI comeback attempt with a major overhaul of Siri into a dedicated chatbot-style assistant capable of generating images, tempered by the now-familiar pattern of a limited free tier designed to push users toward a paid subscription. Other stories include 11 Labs signing its first major label deal with Universal Music Group to build a licensed AI remix platform, the Seattle Times and Newsday suing OpenAI and Microsoft over alleged unauthorised training on paywalled journalism, an expanded class action lawsuit against Anthropic alleging its $100 and $200 subscription tiers were marketed with usage multipliers that mean far less in practice due to five-hour windows and weekly caps, a widely reported claim from an Anthropic figure about long-term existential AI risk, and Amazon Prime Video launching AI-powered lip-sync dubbing technology starting with a German romance series. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.

Andrew Miles Davis answers eleven randomly generated questions in his favourite monthly format, covering the most memorable being asked what he would never trust AI to choose for him. His answer centres on major strategic life decisions, explaining that when a human decision goes wrong there is at least someone to hold accountable, even if that someone is yourself, whereas an AI carries no such responsibility and leaves you with nowhere to direct the frustration when things go badly. He also touches on how he uses AI as an honesty check for his own goals, asking it to be bluntly honest about gaps between his stated ambitions and his actual behaviour. Beyond AI, the episode covers lighter ground including the most ridiculous thing he has ever queued for, involving a five hour London traffic ordeal caused by closed bridges, memorable adverts he still remembers years later, and the first website he ever regularly visited, an early UK garage and drum and bass forum tied to the London record shop Uptown Records. Subscribe to In AI Nutshell for one of these unscripted, unfiltered episodes every month alongside daily ten-minute AI insight.

Andrew Miles Davis returns to prompt hacks with a shift in focus, arguing that as prompting fundamentals become second nature for most people, the real differentiator is what gets added to the end of a prompt rather than the prompt itself. He shares three foundational endings aimed at clarifying message and tone: asking AI to explain something as if you are a complete beginner with no prior knowledge, which forces the model to build from a genuine foundation rather than assuming existing understanding; instructing AI to adopt a specific tone such as professional, empathetic, and reassuring, which shifts focus toward business outcomes and away from purely technical detail, illustrated with a real example of drafting a complaint email calmly despite personal frustration; and asking AI to avoid industry jargon and acronyms in favour of plain, accessible language, preventing specialist terminology from undermining clarity. He also points listeners to his continuously updated prompt endings guide on his website for further examples. Subscribe to In AI Nutshell for prompt hacks and daily ten-minute AI insight built for marketers who want sharper, more usable answers.

Andrew Miles Davis marks 78 weeks of Cool Tools Tuesday with three tools recommended directly by listeners rather than sourced by him, with no affiliate or sponsorship ties to any of them. Sidelong.app promises to combine research, writing, and publishing into a single workflow, and Andrew flags a personal use case, uploading nearly three years of podcast transcripts to potentially generate that same length of social media content. Olva.ai offers bot-free real-time meeting transcription that records audio directly from a device without needing to formally join a call, positioned as an alternative for people uncomfortable inviting a visible AI assistant like Fathom or Otter into meetings while still wanting the transcript benefits. Jobseeker.com rounds out the three, an AI-powered platform for building CVs, cover letters, and tracking applications that a listener credited with directly helping them land a job, though Andrew notes it currently appears geared toward US and EU users with the UK notably absent from its country selector. Subscribe to In AI Nutshell for three new tools every Tuesday, sourced from what's actually working for real listeners.

Andrew Miles Davis argues that one of the biggest mistakes marketers can make is assuming AI simply lets them do the same job faster, when in fact it is fundamentally shifting where marketing value sits. Drawing on his ten years in the music industry, he compares the emerging marketer role to a music producer, someone whose value lies not in playing every instrument but in knowing what good sounds like and directing the process toward it. He explores what happens to value when production becomes cheap, using music's shift from paid cassettes to free streaming as a parallel, and argues that judgement, taste, and the ability to spot technically competent but boring or off-brand work will matter more than the ability to prompt well. He warns against mistaking polished-looking AI output for genuinely good thinking, and closes with a practical challenge: pick one task you produce repeatedly, like a weekly report or email copy, and start building a process where AI handles more of the production while you focus entirely on the decisions. Subscribe to In AI Nutshell for daily ten-minute episodes exploring where AI is genuinely changing the shape of marketing work.

Andrew Miles Davis covers a week defined by AI facing pressure on multiple fronts, starting with OpenAI's advertising business reaching a billion dollar annualised run rate in under 200 days as self-service ad buying expands across India, Europe, the Middle East, and North Africa, a development Andrew is openly hoping fails given his long-held concerns about ads eroding trust in AI-generated answers. He covers Australia's music industry body barring fully AI-generated tracks from official charts following controversy over an AI-assisted Madonna cover, Sony and Warner suing Anthropic for allegedly torrenting and scraping songs to train Claude and seeking up to $150,000 per song, and the US government weighing in on the New York Times' lawsuit against OpenAI in favour of fair use, framed around not falling behind the rest of the world on AI development. Other stories include Instagram replacing its optional AI creator label with a blunter AI-generated tag that throttles reach for undisclosed AI content, a survey suggesting older content creators are becoming more valuable to brands seeking credibility in an AI-saturated market, Time magazine's fourth annual Time 100 AI list sparking debate over the inclusion of celebrities like Paris Hilton and Ben Affleck alongside traditional tech figures, and a study showing AI can measure heart fat in under two seconds compared to fifteen minutes manually, with significant implications for cardiovascular risk assessment. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.

Andrew Miles Davis breaks down a practical framework for moving beyond surface-level AI summaries into genuine analysis, using a fictional set of 300 customer comments as a working example. Rather than simply asking AI to summarise, he demonstrates asking it to identify themes with supporting evidence, then challenging its own conclusions to surface minority patterns that frequency alone would bury, uncovering a small but significant cancellation and refund issue hiding beneath larger but less urgent themes. He connects this three-question approach, what is this, so what, and what now, to his broader AI Essentials framework, and uses three stages of his Story Lens method, arrival, exposure, and decision, to explain where genuine human value now sits in a workflow AI can execute in minutes rather than hours. The core argument is that human judgement does not disappear as AI takes over the mechanical work, it simply moves to questioning and deciding what the analysis actually means. Subscribe to In AI Nutshell for daily ten-minute AI insight and to the YouTube channel for the full demonstration with live prompts and spreadsheets.

Andrew Miles Davis opens September with an honest review of August, where the YouTube channel finally launched with two videos live, and ChatGPT voice conversations during daily dog walks became a genuine part of his working process for developing ideas before reaching his desk. For September he is bringing his backlog of thousands of training questions accumulated over fifteen years of face-to-face sessions to LinkedIn video, alongside a return to a full training and keynote schedule after a deliberately quieter August. The centrepiece of the month is the public launch of his Story Lens Method, a framework built around the idea that everything shared online tells a story whether intentionally or not, drawing on one of seven core story structures to help people navigate change, judgement, and decision-making in an AI-driven world. Having delivered variations of this system with organisations for years, Andrew describes finally building it out fully over the summer as genuinely exciting, a rare reaction for him. Subscribe to In AI Nutshell for daily ten-minute AI insight and to follow the Story Lens Method rollout across podcast, YouTube, and LinkedIn.

Andrew Miles Davis covers three tools on this week's Cool Tools Tuesday, leading with a genuine change of opinion on Copilot's integration with Excel, which he had previously dismissed after a poor experience with a similar feature in Gemini a year earlier. He walks through five practical use cases including turning messy spreadsheets into instant insights, generating charts and pivot tables from plain language requests, writing formulas without needing to know the syntax, automatically cleaning data and standardising UK date formats, and summarising what has changed across workbook versions, concluding that the native Copilot integration outperforms both ChatGPT and Claude's Excel connectors. He also covers Umind.com's prompt library, a free daily-updated resource of trending prompts across images, video, and text, and IdeaVideo.ai, a free video generation aggregator giving access to Seedance, Veo, Kling, and other major models, positioned as a low-stakes starting point for marketers wanting to build video skills without committing to a paid platform. Subscribe to In AI Nutshell for three new tools every Tuesday with honest verdicts from someone testing them in real work.