Plain-English guides, tips and news on using AI tools like Claude, ChatGPT, Gemini, Perplexity and Microsoft Copilot to save time at work and in your business. Written by Harry Lang and the team. No jargon, no hype, just things you can use today. Explore practical guidance for business owners, directors, startup founders, civil servants and public sector teams.
Same brain, different hands. Claude thinks, writes and researches with you. Claude Code is the same model working directly on your files: it builds websites, tools and teams of agents. If the output is something you read, use Claude. If it runs, use Claude Code.
Inside: what each does, what you can build, five signs it’s time to upskill, a four-week route up and 16 tips.
How To Use Claude Code & Agents: learn it, train your team and put AI agents to work.
A five-part series by Harry Lang for business owners and teams who want Claude Code and AI agents doing real work, with no coding background needed. Three guides cover how to learn Claude Code in four weeks, what agentic AI means for a business and how to brief your first AI agent. Two more help you choose who should train your UK team in Claude Code and how to score any Claude Code trainer in London or the UK.
How do you start using Claude Code and AI agents at work? Pick one small, repeat job and run Claude Code on a copy of your files. Use plan mode so you approve each step, write a CLAUDE.md so it remembers your rules and turn the method into a skill. Only then let an agent run it on a schedule, with a check and a stop.
The free AI Foundations training under the AI Skills Boost is open to every UK adult. Harry Lang on what it covers, what funding there is for SME training, where the free course stops short for a small business, and an honest way to choose between free and a hands-on day.
Three guides by Harry Lang covering the latest features, practical prompts and workflows for Claude, ChatGPT and Perplexity. Each includes official sources, rollout details and answers to common questions.
How To Use AI: get ChatGPT, Copilot and Claude doing real work, then train your team.
A five-part series by Harry Lang on how to use AI at work properly. Three guides show how to use ChatGPT, Microsoft Copilot and Claude Projects effectively, with prompts to copy, comparison tables and the settings worth changing. Two more help you decide who should train your team in AI and how to judge any UK AI training course before you pay for it.
How do you use AI effectively at work? Brief it like a capable new colleague: say who you are, what you need, who it is for and what format you want, and paste in the real material. Give recurring work its own Project or Notebook, point the tool at the right file, treat the first answer as a draft and check any fact that matters. Then train the team on their own tasks, in short hands-on sessions with a follow-up.
What AI Training Covers: the best AI training for your business, your sector and your area.
A five-part series by Harry Lang on what good AI training actually teaches and how to pick the right course. Each guide comes with an infographic showing what is covered: a half-day session for small businesses mapped minute by minute, the best AI training matched to six business problems, the routes to AI training for lawyers in the UK, how to find the best AI teacher for Oxford businesses and which AI training to recommend at each level in England.
What AI training is best for a business? The best AI training is short, live and hands-on, built around your own documents and your most expensive problem. It should cover which tool to use for what, how to brief it, how to check the answers and what must never be pasted in. Start with a free government-backed course or a 90-minute team introduction, then book a practical half-day for the people who use AI most.
Employee AI Literacy Training: give every employee the judgement to use AI safely.
A six-part series for UK employers, HR and L&D leads on employee AI literacy training: what AI literacy means, what the EU AI Act literacy duty asks of UK businesses, how to build a programme in 90 days, the eight competencies every employee needs, how to measure and evidence AI literacy, and how to choose a course.
What is AI literacy training for employees? AI literacy is the working knowledge and judgement employees need to use AI tools usefully and safely: understanding what AI is doing, asking it for useful work, checking what it produces, protecting data and knowing when not to use it. Train everyone to a foundation level, go deeper for regular users, managers and high-risk roles, and keep a record of who reached which level.
Our new specialist job board lists AI, machine learning and generative AI jobs in London and across the UK, in every industry, with guides to each role, the skills employers ask for and what the work pays.
Where are the AI jobs in the UK? London has the largest concentration, centred on King's Cross, Shoreditch, the City and Canary Wharf. Cambridge, Oxford, Manchester, Edinburgh and Bristol also have strong AI job markets, and many roles are hybrid or remote. The most advertised roles are machine learning engineer, AI and LLM engineer, data scientist, MLOps engineer and AI product manager.
AI Prompting Masterclass: thirteen free guides to prompts that actually work.
Thirteen plain-English guides to better prompts for ChatGPT, Claude, Microsoft Copilot, Custom GPTs and AI image tools. Each one stands on its own, takes about ten minutes to read and is packed with copy-and-paste prompts, with the reason each one works. See the whole masterclass.
What makes a good AI prompt? Brief the AI the way you would brief a capable new colleague: who it should be, the context it cannot guess, one clear task and the format you want back. Then ask it to check its own answer. Every guide in the masterclass is built on that habit.
What Are The Best Prompts For... CEOs, directors, managers and their teams?
A eight-part series of copy-ready AI prompts for the people who run UK businesses and the teams they lead: CEOs and managing directors, directors and boards, line managers, strategy, HR, sales, and the prompts to use when you train your whole team to use AI. Every prompt works in ChatGPT, Claude, Microsoft Copilot and Gemini, and every guide explains why it works and how to check the answer.
What are the best AI prompts for business leaders? The best AI prompts for business leaders give the AI a role, the real material, the decision you need to make, the format you want back and the limits to respect. Then you check the answer before you act on it. A good prompt saves an hour; a team that shares good prompts, rules and habits saves that hour every day.
Prompts are the start, not the finish. If you want your leadership team and staff using AI the same safe, effective way, our Basic AI Training gets everyone prompting properly on their own work in a day, the AI Policy Workshop sets the rules first, and the free AI Capability Assessment shows where your team stands today. Want to learn the method behind the prompts? Read How To Prompt AI next.
New: How To Prompt AI
How To Prompt AI: write prompts that get finished work back.
A seven-part series on how to prompt AI properly, from the Five Rivets method and ten prompt engineering tips to the specific habits that work in ChatGPT, Microsoft Copilot, AI agents and AI video tools. Every guide has prompts you can copy, worked examples and plain-English answers.
How do you prompt AI? Tell it what you want, who it is for and what good looks like. Paste the real material rather than describing it, ask for a specific format and set a few limits. Then treat the first answer as a draft and steer it with short follow-ups. The tool changes; that habit does not.
The new features of ChatGPT, Claude, Microsoft Copilot, Google Gemini and Perplexity, checked against official release notes. Each guide covers what launched, what is good, what is not, and where a UK business should actually use it.
Compare ten AI books for beginners, work and business, including Co-Intelligence, The Coming Wave and Learn To Fly With AI. Plus four top-five charts for advanced readers, business, beginners and creatives, a guide to which AI book to read first and ten answered questions.
Fourteen AI tools I pay for and use most days, ranked honestly: Claude, ChatGPT, Perplexity, Granola, ElevenLabs, Xero, Close, Gemini, Netlify, Lovable, Upwork, monday.com, Mailchimp and Transpond. Scores, prices, free plans, the best tool for each job and what each one gets wrong.
The thirteen AI tools I use every day to grow The Oxford AI School: Claude, ChatGPT, Perplexity, Gemini, ElevenLabs, Xero, Close, Granola, monday.com, Upwork, Mailchimp, Lovable and Netlify. What each one does, what it costs and what drives me up the wall.
Which AI tools are UK businesses using, and how do you choose one that earns its keep? Six charts, the research behind them and a practical buying guide to ChatGPT, Gemini and Copilot at work.
Businesses are buying the tools. Whether their people know what to do with them is another question. Harry’s report brings together the latest ONS, Deloitte and international research on how many UK businesses use AI, which sectors lead, what shadow AI looks like and where training and AI policy have fallen behind.
35%of UK firms with 10+ staff use AI (ONS)
10%of AI users describe extensive use (ONS)
31%of GenAI users work without employer knowing (Deloitte)
Four new guides for UK business owners, managers and L&D leads who have got past the basics. Measure what your team can already do with a free AI skills matrix, decide what advanced AI training should come next, learn how non-technical staff can build and supervise AI agents safely, and check the latest UK AI training statistics before you quote them.
Assess, upskill and evidence your team’s AI skills
Free government courses, best-course guides & 8 industry guides
UK Industry AI Training
Start with the free AI training the UK government now offers every adult, then go deeper with the guide written for your profession. Each one covers the tools that genuinely suit the job, the regulator and data risks that come with it, and three practical frameworks for training a team on real work. Free courses, funded Skills Bootcamps, apprenticeships and adoption grants, all in plain English.
Fourteen practical guides: eight for leaders choosing the right AI workshop, trainer or team learning plan, and six beginner guides for specific teams, with prompts to try and the rules that keep you out of trouble.
First-person write-ups from Harry Lang on things actually built with AI at The Oxford AI School. What worked, what it cost, what went wrong, and the parts worth copying. No case studies, no hype, just the working out.
Danny Denhard and I sit down with Mark Ritson on Marketing Unfiltered #108. Why most agentic AI talk is hype, why trained generalists who use AI will beat narrow specialists, why distinctiveness beats differentiation, and a few takes too unfiltered for anywhere else. Watch or read, free.
The 30 slides from my talk to final year marketing students at Oxford Brookes University. Seven career mistakes, what AI is doing to the work, and how to get hired when a machine can write the first draft. The short answer is yes.
What this note covers
All 30 slides in a swipeable carousel
Ten predictions for AI and marketing, with odds
The CRAFT prompt: a brief in five parts
Four skills AI can't do for you
Marketing job titles and skills for 2031
How to get hired as a marketing graduate
See the slides30 slides · 9 min read · Published 12 November 2026
Pick Claude, ChatGPT or Perplexity, tell it your job or what you want to make, and get the ten add-ons worth installing first, each with setup steps. 382 connectors, skills, plugins, MCP servers, GPTs and Spaces, all from official directories, ranked by quality and fit. No sign-up.
What AI training actually costs a UK business in 2026, from free government courses to university programmes. Real published prices, per-head maths for a team of eight, the hidden costs nobody puts on the invoice, funding options and our own prices on the table.
What this note covers
The AI training price ladder, free to £7,000
Cost per person for a team of eight
What drives the price up or down
Four hidden costs of AI training
Funding through the Growth and Skills Levy
Is AI training worth the money?
Read the note12 min read · Published 2 November 2026
A free, ten-minute AI skills assessment for workers and teams. Eight practical skills scored 0 to 4, a total out of 32, five bands from Grounded to Flying solo and one safety rule. The scoring is fully transparent, there's a team mode, and there's no email wall.
Thirty questions to ask an AI training provider, in six stages and in the order they're worth asking. Why each question matters, what a good answer sounds like, the red flags worth walking away from and a weighted scorecard to compare quotes. Tick as you go, then print it.
You may have earned the subscribers, but LinkedIn's AI algorithm decides how few of them actually read what you publish. Why LinkedIn newsletters don't work anymore, from the relevance filter on email sends to the golden-hour test and the AI slop crackdown, and what to do about it.
Three completely different kinds of book get sold as "an AI book", and most people buy the wrong one. The five tests a beginner-friendly AI book has to pass, a shortlist grouped by what you actually want from it, and the honest answer on learning ChatGPT, Claude and Copilot from a book. One conflict of interest, declared up front.
What this note covers
The three kinds of AI book, and who each is for
Five tests to apply before you buy
A shortlist grouped by goal, not ranked
Best books for ChatGPT, Claude and Copilot
How to read one so it actually sticks
Book, course or team training?
Read the note18 min read · Published 19 September 2026
I kept meeting clever, busy people who were quietly convinced they had missed the AI boat, so I wrote the book I wanted to hand them. No code, no maths, and honest about the downsides most AI books skip. What is in it, who it is for, and how it actually got made with Claude as co-author.
What this note covers
Who the book is for, and who it is not
The three parts, chapter by chapter
The honest bit about AI's real costs
Working with AI on something book-length
What stayed human, and why
Whether you need the book or the training
Read the note10 min read · Published 15 September 2026
We launched the Oxford AI School shop over a weekend, with no stock, no designer and no warehouse. The whole eight-step process, from generating the medieval crest to the product pages that make individual items findable in Google and AI search.
What this note covers
Prompting an image model for print-ready artwork
Choosing a print-on-demand partner
Writing product copy with AI, then cutting it
Product schema and a Merchant Centre feed
What it cost versus the old way
The four mistakes worth avoiding
Read the note11 min read · Published 7 September 2026
New: 27 practical guides
AI for business, answered in plain English.
Clear guidance for business owners, directors and teams choosing AI tools, training staff, protecting data, writing policy and putting AI to work.
Explore the complete AI for Business collection
Browse all 27 guides by topic, from first steps and team skills to marketing, policy, research and implementation.
ChatGPT, answered properly: prompting, alternatives, Atlas and cancelling.
A four-part series on the questions people actually ask about ChatGPT. Written for people using it at work, in plain English, with the steps you can follow today rather than a list of features.
Perplexity AI, answered properly: how to use it, what's new and how it compares.
A five-part series of Perplexity field guides on the questions people actually search for: how to use Perplexity AI, the new Perplexity features, whether Perplexity is better than ChatGPT or Claude, and what Perplexity is good at. Each guide gives a short answer first, prompts to copy, an infographic and FAQs.
Business AI field guides: research, prompts, prototypes, pitfalls and local training.
Five practical guides to using AI at work: Perplexity vs Claude for research, the best Perplexity prompts for business data, OpenAI Canvas vs Claude Artifacts, the top three AI implementation mistakes to avoid in 2026, and who teaches AI for business in Oxford. Each guide opens with a short answer, then gives prompts or a test to copy, an infographic and FAQs.
How to cancel your AI subscription: ChatGPT, Claude, Copilot, Perplexity and Grok.
A five-part series on cancelling the AI subscriptions people search for most. One golden rule runs through all of them: cancel where you pay. Each guide covers the web, iPhone and Android routes, how to check it worked and what to try before you go.
Five everyday AI tools, explained without the jargon.
Start with the tool your team already has. These practical guides explain what ChatGPT, Claude, Microsoft Copilot, Google Gemini and Perplexity do well, where they go wrong and how to use them safely at work.
Twenty straight answers to the questions people actually ask us in the room – what AI is, how it works, whether it is safe, how to use it at work, and how to learn it properly. No jargon, no hype, no maths.
Machine learning is the method behind almost every AI tool you have heard of, and almost nobody has ever had it explained properly. This three-part series fixes that: what machine learning is, what a model is, and what an algorithm is – in plain English, for business owners and teams, with no maths and no coding.
Read the series, then learn it on your own real work. Our Basic AI Training and AI Basics Deep Dive cover this hands-on, in person across the UK or live online.
The questions UK businesses and individuals are actually typing into Google and ChatGPT right now, answered in full. Whether your business is ready for AI at all, what AI training covers and costs, how to roll it out across a team, where a beginner should start, how to learn ChatGPT and Claude properly for work, and how to get found in AI search.
The Most Common UK Prompt Searches About AI Training
These are the phrases UK business owners, managers and marketers are actually typing when they go looking for AI help, whether that is into Google or straight into ChatGPT, Claude, Gemini, Copilot or Perplexity. Each one gets a full, plain-English answer: what the thing really is, when you need it, what it costs you in time and money, and how to tell a good option from a bad one.
Searched one of these and want the short version? Every guide ends with the same honest answer: the tools are the easy part, and the training is what makes them pay. Start with our Basic AI Training or check where your team stands with the free AI skills assessment.
AI Training Questions Business Leaders Actually Ask
Direct, plain-English answers for owners, directors and team leaders choosing AI training, building safe adoption and measuring the result. Browse by the question you would ask ChatGPT, Claude, Gemini, Copilot or Perplexity.
Plain-English answers for business owners, directors and teams deciding where AI helps, how to start safely, what skills to build and when practical training pays off.
Plain-English answers to the questions UK business owners are asking now: what AI agents can genuinely do and whether they are safe, what an AI policy needs to say and who is liable when an output is wrong, and how to get your business cited by ChatGPT, Gemini, Claude and Perplexity.
How to start using AI, even if you’re completely new to it. A day one guide to the three AI assistants most people use at work: what ChatGPT, Claude and Microsoft Copilot are, where everything is on screen, the first prompts to try, and the habits that get good answers fast – with a free infographic for each tool.
You may have earned the subscribers, but LinkedIn decides how few of them actually read what you publish. Why LinkedIn newsletters don’t work anymore: the AI algorithm’s relevance filter on email sends, the sixty-minute golden-hour test, depth scores and the AI slop crackdown, plus what to do about it.
78% of recruitment agencies have adopted AI, but only 6% have embedded it into how the desk works – and agencies with proper training are fifteen times more likely to see a strong return. The Oxford AI School launches ‘AI for Recruiters’, half-day AI training for recruitment agencies, executive search firms and in-house talent teams, as 9 out of 10 UK job seekers now use AI on their applications.
What an AI readiness assessment really measures, a ten-question check you can score in about ten minutes, and the right first move for each band of results.
A step-by-step plan for training your team to use AI at work, from choosing one tool and three real tasks per team to measuring the hours it gives back.
A free, one-page AI usage policy template for a small UK business. Seven plain-English sections you can copy, adapt and put in front of your team this week.
If an employee acts on a wrong AI output and a customer is harmed, where does UK liability sit? What Air Canada and Mata v Avianca established, and the five safeguards that matter.
A chatbot answers, an agent acts. A side-by-side comparison of autonomy, tool access and risk, and how to tell which one a vendor is actually selling you.
Ten real jobs an agent can take on today, from inbox triage and invoice chasing to CRM updates and content repurposing, plus the tasks to keep a human on.
The five real risks, including prompt injection and excessive access, each paired with the control that manages it, and a checklist for setting an agent up safely.
A free, five-step method for auditing your visibility across ChatGPT, Google AI answers, Gemini, Claude and Perplexity, and what to do if you are missing.
Around 84% of AI citations come from earned media, not your website. How PR coverage turns into ChatGPT and Perplexity citations, why paid media barely registers, and how PR teams can win them.
A practical nine-step guide to GEO and AEO, covering prompt research, answer-first content, authority, technical discoverability, schema and AI visibility measurement.
How startups actually use AI to move faster with fewer people, from product and marketing to support and fundraising. A practical guide for founders, no hype.
Founders are busy and AI seems intuitive, so is training worth it? An honest look at what AI training gives a startup team, and when to skip it and just get on.
Practical AI training for businesses in Manchester and Birmingham, in person and live online, built around your real work. For owners, founders and teams.
Practical AI training for government departments, councils and public sector teams. Real skills, taught safely, with proper attention to data and governance.
A practical guide to how civil servants and government staff can build AI skills safely, what to learn first, and how departments can train teams responsibly.
PR already owns the hardest part of generative engine optimisation: earning credible third-party authority. So why are agencies allowing GEO to be framed as an SEO add-on? A practical argument for training the team, building a cross-functional operating model and piloting the service before putting it in the pitch deck.
Thousands of marketing emails, most of them the same five tips in a new jacket. Here are the ones worth your inbox – 40+ ranked across four tables: the best free marketing newsletters in the UK, the best marketing blogs in Europe, the best AI marketing newsletters, and the best in the world. Plus an FAQ on starting your own.
Every single published review of The Oxford AI School is a five-star review. The launch announcement in full – why a marketing veteran started a school to teach basic AI to farmers, solicitors, publicans and small business owners, what the training costs, and the verified feedback from the people who have already been through it.
Your team is sitting on the most powerful tool since the internet, and most of them are using it to write slightly better emails. We handed the workforce a Ferrari and watched them drive it to Lidl in second gear – 75% of AI-using businesses report better productivity, but only 12% see any more revenue. Why most corporate AI training is theatre, why ‘basic’ AI training is actually the whole game, what the best AI training for a business really looks like, and what it should cost across London and the South East.
Your team is already using AI. The only real question is whether they are following your rules or making up their own. The fifteen questions we get asked in the room, answered in full – do I need one, is it legally required, what should it include, how do I stop staff pasting client data into ChatGPT – plus a free AI policy template you can download as a Word document and complete this afternoon.
Most AI policies I read were written to be filed, not followed. There is no single UK AI Act to comply with, which confuses people into either doing nothing or paying for a document nobody reads. This is the five decisions a governance framework actually has to make – approved tools, the data red line, which calls stay human, who owns it, and how you would prove any of it – plus how to get it written in a day.
Lawful, ethical and robust is the definition of trustworthy AI. Govern, Map, Measure and Manage is the method the NIST AI Risk Management Framework gives you to get there. Here is what both mean in a business with actual work to do on Monday, including how to measure whether a tool is reliable enough to use, and why alignment failures in real companies are far more mundane than the debate suggests.
The question that catches businesses out is never whether they are allowed to use AI. It is ‘show me what happened’, eight months later, from somebody who is not pleased. Compliance is proving it, auditing is checking it, and explainability is being able to say why. The records that make this survivable fit on a spreadsheet, and the audit takes ninety minutes twice a year.
Most people pick the first AI tool they come across and learn as they go – then wonder why AI feels like more effort than it saves. Answer ten quick questions about your actual job and this free tool hands you a short, sensible stack: the tools worth your time, whether the free tier will do, and which model to reach for when the work matters. Plus the five layers every good AI stack has, and the data rules to follow if your work is confidential.
Twenty minutes from the UK’s first AI Growth Zone at Culham, ordinary Oxfordshire businesses are quietly stuck in 2019. AI in Oxford is booming; AI for the ordinary Oxfordshire business is barely happening. The gap is not about skill – it is about access. Here are the real blockers, and a simple six-step plan to get any small business AI enabled with basic AI training in Oxfordshire.
The 40 questions people ask most about learning AI for work in 2026 – answered plainly. Will AI take your job? Is it too late to start? How do you train a team without breaking the budget? Straight, jargon-free answers on job security, salaries, choosing a course, and getting your small business using AI without wasting money.
Ask anyone in marketing how to get visible in ChatGPT and they will tell you to post on Reddit. So we checked. Reddit matters far more than most people realise, and far less than the loudest voices claim – and it depends entirely on which AI you ask. A data-led look at ChatGPT, Perplexity and Gemini, and what it means for getting your brand found by AI.
Claude isn’t one product anymore – it’s four. A short field guide to the best way to use Chat, Projects, Cowork and Code, so the right tool picks itself. Chat for a fast answer, Projects for recurring work, Cowork to hand off a whole task, and Code to build something that actually works.
When a constant raft of low-cost AI tools meets ever-simpler usability, everyone can work bigger, better, smarter and faster. Twin that with a global power race, and we’ve got a problem. A no-nonsense reality check on the real value and the real cost of the AI gold rush.
From unhappy CMO to start-up teacher in 24 hours: Harry explains the career pivot, pressure and purpose behind launching practical AI training for local SMEs.
5 AI Tasks Every Small Business Should Automate First
When people ask where to start with AI, the honest answer is: start where it saves you the most boring time, fastest. These are the five jobs we help small businesses hand to AI first — in roughly the order they pay off.
1. First drafts of emails and replies
Most teams lose an hour or more a day to the inbox. Paste the message you received, tell the AI the outcome you want and the tone to use, and let it produce a first draft. You stay in control — you read, tweak and send — but you skip the blank-page part that eats the time.
2. Turning notes into finished documents
Meeting notes, voice memos and rough bullet points become clean proposals, summaries and reports in seconds. The trick is to give the AI your raw material and a clear instruction about the format you need, rather than asking it to invent facts.
3. Repurposing one piece of content into many
A single blog post or recorded talk can become a newsletter, five social posts and a short video script. AI is at its best when it is reshaping content you already have, so your marketing output multiplies without multiplying your hours.
4. Researching and summarising
Long PDFs, competitor websites and dense reports can be summarised into the three things you actually need to know. Tools like Claude and Perplexity are excellent here — just remember to sense-check anything you will rely on.
5. Cleaning up and analysing data
Spreadsheets, customer lists and survey responses can be tidied, categorised and analysed in plain English — no formulas required. This is often the task that surprises people most, because it used to need a specialist.
The takeaway: pick one of these, try it this week, and build the habit before adding the next. Most of our clients save 10+ hours per person per week once these five are in place. Want help getting started? See our AI training for business or book a free intro call.
Tools5 June 2026·By Harry Lang·5 min read
ChatGPT vs Claude vs Gemini: Which AI Should You Actually Use?
It is the question we are asked in almost every session. The honest answer is that the three leading assistants are all excellent, and the best one is usually the one you will actually open. But here is how we help people choose.
ChatGPT (OpenAI)
The best-known and the most versatile all-rounder. Brilliant for brainstorming, images, voice conversations and a huge ecosystem of add-ons. If you want one tool that does a bit of everything and the widest range of features, this is a safe first choice.
Claude (Anthropic)
Our go-to for writing, careful analysis and working with long documents. Claude tends to produce natural, thoughtful prose and is excellent when the quality and tone of the words really matter — proposals, reports, sensitive emails and detailed reading.
Gemini (Google)
The natural pick if you live in Gmail, Docs and the rest of Google Workspace, because it is built right into the tools you already use. Strong for research and for anyone who wants AI woven into their existing day rather than in a separate tab.
So which one?
For most individuals and small teams we suggest starting with one tool, learning it properly, and only adding a second once the first is a habit. Don't agonise over the choice — the gap between the tools is far smaller than the gap between using AI well and not using it at all. We teach all three, and help you match the tool to your actual work.
Explore our programmes to learn Claude, ChatGPT, Gemini, Perplexity and Copilot hands-on.
Beginner28 May 2026·By Harry Lang·4 min read
How to Write Better AI Prompts: A Beginner's Guide
If your AI answers feel generic, the problem usually isn't the tool — it's the instructions. Here is the simple four-part recipe we teach in every session for getting genuinely useful results out of any AI.
Role — tell the AI who to be. "You are an experienced bookkeeper" or "Act as a friendly copywriter" instantly sharpens the answer.
Task — say exactly what you want it to do, in one clear sentence. Vague in, vague out.
Context — give it the raw material and the background: who it's for, the tone, any facts it must use. This is the step most people skip, and the one that makes the biggest difference.
Format — describe what good looks like: "a 150-word email", "five bullet points", "a table with three columns".
Put together, a weak prompt like "write a marketing email" becomes: "You are a friendly copywriter for a small Oxfordshire bakery. Write a 120-word email announcing our new sourdough, warm and not salesy, ending with a clear call to visit this weekend." The difference in quality is night and day.
One more tip: treat it as a conversation. If the first answer isn't quite right, tell the AI what to change rather than starting again. That back-and-forth is where the real magic happens.
The Oxford AI School is registered with the Information Commissioner's Office, the UK regulator for data protection. See the certificate, what registration means for the people who train with us, and how to check it on the public ICO register.
Where to get hands-on Claude and Microsoft Copilot training, practical AI training for London professional services firms, new county pages for Essex and Sussex, and a plain answer to how The Oxford AI School relates to the University of Oxford.
Get better results from the AI tools you already have. How to use ChatGPT, Claude and Microsoft Copilot effectively, plus a practical method for using AI effectively at work: clearer briefs, the right source material, copyable prompts and a simple way to check the output before you rely on it.
How To Use Claude Like A Pro: Projects, prompts, context, Skills, MCPs and Connectors.
A six-part series for people who already use Claude and want it to do proper work. Each guide covers one feature, with where to find it, a decision flowchart, prompts you can copy, a practical exercise and eight plain-English FAQs. Start with Projects and prompts, then add context, Skills and connected apps when you have a job that needs them.
How do you use Claude like a pro? Give each piece of recurring work its own Project, brief Claude with the task, audience, sources and the output you want, and supply only the context the job needs. Save a tested method as a Skill. Connect apps through Connectors or MCP only when a real task needs live data, and check the result against the source.
How to choose an AI course that fits your starting point and goals: beginner lessons in London, generative AI courses for beginners, and what to check before booking an intensive AI bootcamp. Each guide includes five copyable practice prompts and a plain-English FAQ.
One honest guide to choosing who teaches you and your team: the best AI training and AI trainers in the UK ranked by need and budget, a four-step provider checklist, a trainer scorecard, and how to brief corporate AI training that changes the way people work.
Straight answers to three questions business owners ask before they spend money on AI: how to choose executive AI coaching, what Claude and Claude Code training actually teaches and how to optimise a website for AEO, GEO and AI search. Each guide by Harry Lang opens with a short answer, a four-step checklist diagram and a plain-English FAQ.