The Matts Workshop

AI in the Workplace Part 2

Workshop Activity · 5 min AI Discovery Sprint Find your next AI experiment: could this be useful?
Start sprint ↗

✓ At Cardiff University: Gemini and NotebookLM are approved for confidential (C1/C2) data, but only when you sign in with your CU account (cardiff.ac.uk).

• For other organisations: treat them as personal learning tools, stick to public, non-confidential work data.

Gemini Prompt Library

Gemini Exercises & Workplace Prompts

Part 1
Click to open Gemini gemini.google.com ↗ Click to open NotebookLM notebooklm.google.com ↗
Click "New Chat" in Gemini between tasks. Use your Cardiff University account for work data.

The coloured tags below match the section markers in the corner of the slides, so you can always see where we are in the session.

01Explain it like I am 5 (ELI5)Part 1
Application: Quickly understand complex topics or even an uploaded document.
"Explain [TOPIC] to me like I am 5 (ELI5)"

Replace [TOPIC] with: "cloud computing", "DNA", "quantum physics", "offside rule in football" etc. Or instead of a topic, upload a document such as your travel policy and ask it to explain that.

02Upload and review grant proposalsPart 2
Application: gather a few past proposals: a funded one, an unsuccessful one, reviewer comments. Upload them to Gemini along with this prompt. No proposals to hand? Any long work documents will do.
Review the uploaded proposal(s) and identify the sections most likely to have concerned or disappointed a reviewer. Look for: vague impact claims, unsubstantiated assertions, weak methodology description, budget items with no justification, or misalignment with likely funder priorities. List each weakness and suggest what a stronger version would include.

Four more review prompts (reusable text, what worked, key phrases, structure) live in the NotebookLM section further down the page.

03Same files, now in NotebookLMPart 2
Application: feel the difference between broad chat and answers grounded in your sources. Create a notebook at notebooklm.google.com, add the same files, and run the same prompt from Example 02. Then ask this follow-up.
Which of your answers came from which document? List each finding with the source it is based on, and flag anything you could not trace back to a specific document.

NotebookLM only answers from the sources you gave it, and cites them.

04Set Deep Research goingPart 3
Application: hand Gemini a research job and get back a sourced report. Switch on Deep Research in the tools menu under the prompt box, paste this, and check the plan it shows you before it starts. You can edit that plan.
I work at a university. Research how UK universities are using generative AI to support staff across teaching, research and professional services, and produce a short, sourced briefing that covers: the five most important developments from the last 12 months; what is changing right now; and three practical things a team like mine should consider trying. Cite your sources throughout so I can verify them.

Two fuller role-based prompts (academic and professional services) are in the Deep Research section further down the page.

05Turn a document into an imagePart 4
Application: get a slide-ready image from material you already have. Pick Create image in the tools menu, upload a document (or stay in your grant chat), and ask.
Using the attached document, create a clean, simple infographic that explains [PICK A SECTION OR TOPIC] to a general audience. White background, one accent colour, large readable labels, no clutter.

More image ideas (editing photos you already have, bilingual posters) are in the Nano Banana section further down the page.

06Build a small tool in CanvasPart 4
Application: finish with the wow: a working web tool from one prompt. Pick Canvas from the tools menu, attach a policy or guidance document, and paste this.
Create an interactive one-page quiz web app that tests staff on the key points of the attached document. Five multiple-choice questions, friendly feedback after each answer, and a score at the end. Use Tailwind CSS and vanilla JavaScript in a single HTML file.

Two more Canvas builds (an expense claim estimator and an interactive concept visualiser) are in the Canvas section further down the page.

The prompts you are likely to use most often, whatever your role, come first. More role-specific ones are tagged and grouped below them: Professional services, then Academic.

Devil's advocate (blind spots)
Application: Stress-test your challenges, new ideas, or decisions before finalising them.
"Act as a constructive devil's advocate. First identify potential weaknesses or blind spots, then suggest alternative perspectives. Be specific and challenging but constructive. My scenario is: [scenario here]"

Add in your own scenario: "I want to move from sales to leadership role", "dealing with a difficult client", provide some useful context as well. You could upload a document instead of typing a scenario e.g. upload your new travel policy. Then experiment with a meeting transcript, an Excel file or even a PowerPoint.

Filter a document to your role
Application: Upload a document and pull out only the tasks, deadlines, and details relevant to your job.
Read this Document from the perspective of a [INSERT ROLE HERE e.g. project manager]. Only pull out the information,action items and deadlines that would impact my role.

Attach your document to Gemini, paste this prompt, and replace the bracketed text with your actual job title to instantly cut through the noise.

Summarise meeting notes
"Please summarize the following meeting transcript. Focus on the key points, decisions made, action items and any important deadlines. Ensure the summary is clear and organised. The summary should be no longer than 200 words: [INSERT TRANSCRIPT HERE]"
Rewrite anything in plain English
Application: paste a policy, a research paper, a report or a meeting transcript and get a version anyone can understand in one read.
You are an expert communicator who rewrites specialist text for a general audience. When I share any text, a policy document, a research paper, a report or a meeting transcript, rewrite it so that a reader with no background in the subject can understand it in one reading. Remove or briefly explain all jargon. Shorten sentences. Put the real-world significance first. Cut hedging and passive voice. After rewriting, note: the three most significant changes you made and why; and any technical terms you chose to keep and why. Do not change the meaning or leave out important content.
Analyse an Excel file
Application: turn a raw spreadsheet into a verified summary, then a chart, in two quick steps.

1. Verified summary

Summarise this Excel file. Use code (Python) to calculate the figures, do not estimate them. Give me: 1. Overview: rows, columns and the date range 2. Key numbers: totals, averages, highs and lows 3. Three insights worth knowing Then show the totals you used and cross-check them before you answer. Use plain business English.

2. Chart the trend

Now chart the most important trend in this data, and explain in two bullets what it means for a manager.

Run step 1, then paste step 2 as a follow-up. Asking the AI to calculate in code and cross-check its totals is what stops it inventing numbers.

Categorise text or feedback
"Analyze the [Name of Feedback column] column. Add a new column named 'Category'. Classify each row into exactly one of these specific categories: 'Compensation', 'Management', 'Work Environment', 'IT/Tools', or 'Work-Life Balance'. If the feedback doesn't fit these, label it 'Other'."
Roleplay for difficult conversations
Application: Useful for learning, or to prepare for an upcoming meeting. Rehearse performance reviews or conflict resolution in a safe space.
"Roleplay: You're Mrs. Grudge, my grumpy neighbour knocking at 2 a.m. to complain about my noisy cats (Knox, Missy, Phoebe). I'm a witty cat owner trying to calm you down and win you over. Stay in character and wait for my replies."
Cut the AI slop (style guide)
Application: Add this to the bottom of any prompt to stop the AI from rambling and make it write clear, direct, and natural-sounding text.
[INSERT AI TASK HERE e.g., Write a report on modern marketing trends] STYLE AND TONE CONSTRAINTS Apply these strict rules to the task above: * Voice: Speak with direct authority. Use active voice exclusively. Omit all meta-commentary (e.g., do not say "Here is the report"). * Banned Words: Zero adverbs. No business jargon. Remove all filler and emphasis words ("very," "extremely," "actually"). * Banned Starters: Never start a sentence with Who, What, Where, When, Why, Which, or How. * Punctuation & Flow: Never use em dashes (—); separate ideas with colons or full stops instead. Avoid rhetorical setups, binary contrasts, and choppy fragments.

Paste this constraint block at the very bottom of your standard prompts to instantly improve the output quality. Ensure you replace the bracketed text at the top with your actual task.

Let AI write the prompt
Application: Get the AI to design an expert prompt for a task in your domain, e.g. 'triage IT helpdesk emails'.
"You are an expert AI prompt designer. Please design the most effective prompt to: [INSERT SPECIFIC TASK AND DOMAIN]. Consider what key elements, criteria, or details matter, what output format is most actionable, and the reasoning steps that are essential."
Reverse-engineer any existing content

One reusable prompt that clones the look, feel, voice and structure of any document you upload.

Analyse the example I have uploaded: it could be a document, image, or website. Identify what makes its style distinctive: its tone of voice, vocabulary, sentence structure, formatting, and layout. Write a single reusable prompt that I can use again with new material to produce new versions with this same look, feel, and structure. At the bottom of that prompt, add the placeholder [PASTE NEW CONTENT HERE] where the new material will go. Output only the reusable prompt, nothing else.

Upload your example, then run this prompt: example in → AI analyses the pattern → reusable prompt out.
Works on anything with a pattern: documents, slide decks (PowerPoint), websites, images, reports, SOPs, transcripts.

Analyse free-text survey commentsProfessional services
Application: read hundreds of open survey comments at once and surface the concerns that come up most.
You are a student support officer reviewing the annual wellbeing survey. Read all the free-text comments below (over 400 responses) and find the concerns students raise most often. Be empathetic and data-driven, focused on student welfare. Give me a bulleted list of the top five concerns. For each one, suggest which policy or support area should review it. [PASTE THE FREE-TEXT COMMENTS]
Summarise a regulator document for your boardProfessional services
Application: turn a dense regulator or compliance document into a board-ready summary with questions worth discussing. Attach the document, and swap in any policy paper.
Persona: Expert policy analyst. Task: Analyse the attached [new regulatory document from Medr] and summarise its central themes and key institutional risks. Then generate 3 strategic questions to encourage critical thinking. Context: The summary and questions will be used by [the University Executive Board] to understand the wider context and prepare for upcoming changes. Format: A bulleted summary followed by a numbered list of the 3 strategic questions.
Redesign a process around its usersProfessional services
Application: rethink any process from the point of view of the people who actually use it. Swap in your own process and profiles.
Persona: [Director of Finance]. Task: Revise the current expense claim process, suggesting changes to the approval workflow, documentation requirements and submission methods. Context: The process needs to support three specific staff profiles: [a new employee, a researcher claiming international travel, and a temporary contractor]. Format: A structured breakdown with a clear heading for each staff profile, using bullet points for the required changes.
Plan a new starter's first three daysProfessional services
Application: a ready-to-adapt onboarding schedule so each new colleague gets the same good start.
Persona: Employee onboarding coordinator. Task: Draft a detailed 3-day schedule and a companion checklist for onboarding a new member of staff, integrating HR, IT and departmental sessions. Context: We need a seamless, comprehensive start for [academics] joining the institution this semester. Format: A numbered daily schedule with time slots, required resources, and the responsible department contact for each item.
Analyse a focus group transcriptAcademic
Application: turn a long focus group recording into the themes and sentiment course leaders need.
You are a marketing assistant writing up a student focus group for course leaders. Read the attached transcript of a 60-minute focus group about the new course. Identify the main themes and the overall sentiment towards the course. Stay objective and analytical, as in market research. Give me a bulleted summary: - The top three positive themes - The top three negative themes [ATTACH OR PASTE THE TRANSCRIPT]
Evidence-based module changesAcademic
Application: turn a student feedback report into the specific changes a lecturer should make next time.
Based on the [student feedback report] what are the three most important changes this lecturer should make before running this module again? Be specific about what exactly needs to change and why.
Compare marking and grading approachesAcademic
Application: an evidence-based comparison to inform an assessment or module redesign.
Persona: [University lecturer] and educational researcher. Task: Compare alternative grading systems in higher education, focusing on [ungrading, mastery-based grading, and contract grading]. Highlight research on outcomes, student engagement and staff workload. Context: I am redesigning my module for the upcoming academic year and need evidence-based practice to inform my assessment strategy. Format: A comparative table followed by a short summary of the research.
Score a grant proposal (UKRI panel)Academic
Application: assess a grant draft against funder criteria, with scores and the weak points most likely to cause rejection.
You are an experienced academic peer reviewer sitting on a UKRI funding panel. You have reviewed hundreds of grant proposals across research councils including ESRC, AHRC, EPSRC, and Innovate UK. When I share a grant proposal section or full draft with you, assess it as a panel reviewer would. Score each of the following criteria out of 5 and give a brief justification for each score: clarity and coherence of the case for support; quality and feasibility of the proposed methodology; strength of the impact and pathways to impact statement; evidence of the team's track record and capability; value for money and budget justification; fit with funder priorities. After scoring, provide: (1) the two or three strongest elements a panel would likely praise; (2) the two or three weakest elements most likely to cause rejection; (3) a suggested priority order for revision if time is limited. Be direct. Do not soften criticism.
Check a proposal against a funder callAcademic
Application: paste a funder call document and your draft, and see where the proposal does and does not match what the funder is asking for.
You are a research funding strategist with deep knowledge of UK and international research funders including UKRI councils, Innovate UK, Horizon Europe, Wellcome Trust, Leverhulme Trust, NIHR, and the Welsh funders Medr and Health and Care Research Wales. When I share two pieces of text, a funder call document and a draft proposal section, analyse how well the proposal aligns with what the funder is actually looking for. Provide: (1) a list of the funder's stated priorities and assessment criteria; (2) for each priority, a judgement on whether the proposal addresses it (clearly / partially / not at all) with a reference to the proposal text; (3) funder language and key phrases from the call document that do not appear in the proposal but should; (4) specific suggestions for where each gap could be addressed. Be precise. Reference the actual words of both documents.
Write better prompts: the 1-2-3 method
The Golden Rule: Better inputs equal better outputs. Follow this 3-step flow for every prompt to get the best results.
1.
Give Context

Provide relevant background information (Role, Data, Reports).

👤 📊 🌐 📄
2.
Make Request

Be specific about exactly what you want the AI to do.

"Review my email for clarity..."
3.
Explain Purpose

Share why you need this (The "So That").

"...to maximize the chances of a quick response."

You may also hear this taught as PTCF: Persona, Task, Context, Format. It is the same idea with different labels: Persona and Context are step 1 (give context), Task is step 2 (make the request), and Format says what shape the answer should take. Four of the Workplace Prompts are written in exactly that shape (they start with Persona:), so you can see it in action.

Treat AI like a teammate, not a vending machine
The Reality: You rarely get a perfect draft from a colleague first time. Don't expect it from AI either. The magic is in the follow-up.
Avoid
The Vending Machine

Insert prompt → expect a perfect deliverable in one shot.

Adopt
The Teammate

Iterate together: guide, correct, refine until it lands.

🤝 The Mindset

Treat Gemini like a clever intern, not a vending machine.

If the output isn't right, don't start over. Just talk to it and guide it.

Try these follow-up phrases:
"Actually, can you make it more formal?"
"That's good, but what if we tried a friendlier approach?"
"Can you expand on that second point? It's a bit too brief."
Context rot: long chats make more mistakes
The Reality: as a chat gets long, the AI starts to drift. Start new chats regularly to keep it sharp.

How a conversation degrades:

Message 5

Sharp, on task.

Message 25

Starts repeating itself.

Message 45

Forgets earlier instructions.

Message 70

Contradicts its own answers.

Three fixes:

  • Start new chats regularly. Don't wait for things to go wrong: a fresh chat keeps the context short and the model sharp.
  • New topic? Start a new chat.
  • Gone off the rails? Run the handover prompt below, then paste the output into a new chat.

Handover prompt (run in the existing chat, then copy the output into a new chat):

Prepare a handover I can paste into a new chat. Rules: 1) five bullets, one line each: decisions, key facts, constraints, open questions, next steps, marking anything unconfirmed; 2) copy anything worth keeping (drafts, wording, lists) verbatim under short headings, no rewriting or merging; 3) end with a two-line brief: the new chat's role and its first task. No sub-bullets, no commentary.
Two heads are better than one: get a second AI opinion
The Reality: an AI rarely spots its own mistakes. It is anchored to its first answer. A second opinion catches what it missed: ask a different tool (Copilot vs Gemini), or even start a new chat in the same tool.
🔀 Best: a different tool (Copilot vs Gemini)

Paste Copilot's answer into Gemini, or Gemini's into Copilot, and ask it to review. They are built on different AI, so they have different blind spots: each catches the other's misses.

🔄 Or: the same tool, a new chat

No second tool? Open a new chat in the same one, paste only the output, and ask for a critical review. With no memory of how it got there, it judges the work, not its own reasoning.

A reusable review prompt:

Act as a critical second opinion. You did not write the text below, so be objective. Check it for factual errors, weak reasoning, missing information, and anything unclear or overstated. List the specific problems you find, then give an improved version. Be direct. [PASTE THE OUTPUT HERE]

Why it works: an AI tends to defend its first answer and rarely self-corrects without an outside nudge. A different tool, or a clean chat, removes that bias.

Keep it safe: only use IT-approved tools, and never paste confidential (C1/C2) work data into a tool that has not been approved.

Chain your tools together
The idea: the output of one tool becomes the input to the next. Each tool does the part it is best at, and you carry the work across them. Experiment, because a short chain often beats one tool trying to do everything.
1️⃣ Research it

Run the Deep Research prompt in Gemini to gather the evidence and write it up as a report.

2️⃣ Ground it

Add that report to NotebookLM as a source, so the next step is anchored to your material, not guesswork.

3️⃣ Shape it

Ask NotebookLM to turn it into a slide deck. One research prompt has become a finished presentation.

Why it works: no single tool is best at everything. Passing the work along lets each do its strongest job, and you stay in control at every step. Another chain to try: run Deep Research, then paste the report into Canvas and shape it into a staff briefing or a quiz.

Stop repeating yourself: keep a MegaDoc
Concept: Stop retyping the same background. Keep a Word doc of your org's context and paste it into prompts.
🔤 Acronyms

e.g. "DCG = Division of Cancer & Genetics"

✒️ Style Guide

"Formal but warm. UK English. Bullet points."

📋 Key Facts

Personnel, boilerplate, standard procedures.

🚀 Project Overview

Goals, status, blockers.

Try it today:

List your top 10 acronyms and one project snapshot. Paste into your next prompt and watch the difference.

"Here is some context from my MegaDoc: [PASTE YOUR ACRONYMS/STYLE GUIDE HERE] Based on this context, draft an update email to the team about [Your Topic]..."
Chat with your own documents (NotebookLM)
The Tool: NotebookLM grounds AI in your own documents. Our curated prompt library helps you analyse them.
🎧 Audio Overviews

Turn dry reports into engaging "Deep Dive" podcasts.

🕵️ Critical Analysis

Identify blind spots, contradictions, and missing data.

🏗️ Structured Output

Convert notes into Slide Decks, Briefs, and SOPs.

Get AI help with code
Two ways to use AI for code:
  • Chat: paste in code, errors, or questions.
  • Coding Agent: connects to your system and can read, write, and run files. Use a local model + decent GPU to keep data private.
Run AI privately on your own device
You Are In Charge: Run AI on your own device. Total privacy, zero fees, no Wi-Fi needed.
🛡️ Privacy & Control
  • Data never leaves your device.
  • Safe for sensitive material.
  • Full IP ownership.
⚡ Cost & Efficiency
  • Zero fees (open-source models).
  • Works offline.
  • Lower energy use.
🎓 Private AI Workshop

A 2-hour session to get you started on your own laptop.

Contact us for details →
NotebookLM

Understand anything with NotebookLM

Part 2
Click to open NotebookLM notebooklm.google.com ↗
Sign in with your Google account, then create a notebook and add your sources. Use your Cardiff University account for work data.
1 Upload your sources
2 Deepen your understanding
3 Ask and create
🎧 Audio Overviews

Turn dry reports into engaging "Deep Dive" podcasts.

🕵️ Critical Analysis

Identify blind spots, contradictions, and missing data.

🏗️ Structured Output

Convert notes into Slide Decks, Briefs, and SOPs.

Hands-on: review grant proposals (upload and compare)
The exercise: gather a few past proposals: a funded one, an unsuccessful one, reviewer comments. Upload them to Gemini and run the prompts below. Then add the same files to a NotebookLM notebook and run them again. Compare: NotebookLM sticks to your documents and cites them. No proposals to hand? Any long work documents will do.
1. Pull out the reusable text: find the boilerplate you can lift into your next bid.
I have uploaded one or more previous grant proposals. Please identify all sections that contain reusable institutional boilerplate: for example, descriptions of the university's research environment, facilities statements, EDI statements, or standard methodology descriptions. For each section, copy the text and label which part of a proposal it came from.
2. Find what worked: the passages most likely to have impressed a reviewer.
Looking at the uploaded proposal(s), identify the writing that is most likely to have impressed a reviewer. Look for: a clear and compelling case for need, strong impact framing, well-justified methodology, and confident team track record. Pull out the strongest paragraph or passage from each of these areas and explain briefly why it is effective.
3. Find the weaknesses: what a reviewer would have marked down.
Review the uploaded proposal(s) and identify the sections most likely to have concerned or disappointed a reviewer. Look for: vague impact claims, unsubstantiated assertions, weak methodology description, budget items with no justification, or misalignment with likely funder priorities. List each weakness and suggest what a stronger version would include.
4. Extract the key phrases: build a reference bank for future bids.
From the uploaded proposal(s), extract the key phrases, terminology, and framing language used. Group them under: research significance, methodology, impact, team capability, and budget justification. I want to use these as a reference when writing future bids.
5. Summarise the structure: see the skeleton and spot the gaps.
Summarise the structure of the uploaded proposal(s). For each section, note: the heading used, the approximate word count, the main argument being made, and how well it connects to the section before and after it. Flag any structural gaps or sections that feel underdeveloped.

Try these two prompts

Upload your sources to a notebook first, then paste a prompt into the chat box.

1. Executive Briefing: turn a pile of documents into a manager-ready summary.
"Act as a senior analyst. Read all of my uploaded sources and produce an executive briefing for a busy manager. Include: a three-sentence situation summary; five key findings ranked by importance, each with a source citation; the main options with their trade-offs; and two recommended next actions. Flag any important questions the sources do not answer."
2. Critical Reviewer: surface contradictions, gaps and unsupported claims across your sources.
"Review all of my uploaded sources together and act as a critical reviewer. Identify any contradictions or disagreements between the sources, note important gaps or missing evidence, and list the assumptions that are not supported by the documents. Cite the specific source for each point."
Open NotebookLM Prompt Library ↗
Gemini Deep Research

Now AI can do your Google searches for you

Part 3

It does not skim a few links and guess. It reads hundreds of pages properly while you get on with other work, then hands you one organised report with a link to every source.

1 Ask one clear question
2 It reads hundreds of pages
3 Get one sourced report

You stay in charge: before it starts, Deep Research shows you its research plan, and that plan is editable. Cut the topics you do not care about, add specifics, then set it going.

For academic staff: scope a field before a grant or curriculum review

Open Gemini, switch on Deep Research, and paste this. Swap in your own field and institution.

I am an academic at Cardiff University preparing to write a research proposal and update my module curriculum. I need a structured overview of the current state of research into AI tools in higher education pedagogy. Please research and produce a report that covers: 1. The key debates in the field: what are researchers currently disagreeing about or actively investigating? 2. The most significant published findings from the last three years on how generative AI is changing teaching, learning, and assessment in universities. This field moves quickly, so weight your answer towards work from the past 12 months. Include specific studies where possible. 3. The main gaps in the current evidence base: what questions remain unanswered that a new research proposal could address? 4. The policy context in the UK: what have Medr, UKRI, or the Quality Assurance Agency published on AI in higher education that an academic should be aware of? 5. Three to five specific research questions that are currently underfunded or underexplored, which could form the basis of a competitive grant application. 6. The practical implications for my module curriculum: based on these findings, what should I consider when updating my course content, assessment methods, and teaching practice to reflect how AI is changing higher education? Cite your sources throughout so I can verify and follow up on the most relevant ones.

For professional services: an international student market briefing

Same steps. This one scopes a market for strategic planning rather than a research field.

I work in professional services at Cardiff University, supporting international student recruitment and strategic planning. I need a market briefing to inform our recruitment strategy for the next intake. Please research and produce a report that covers: 1. The current trends in international student recruitment to UK universities: which source countries are growing, which are declining, and why. 2. The factors shaping demand for Welsh universities specifically: how Wales is positioned compared with the rest of the UK, including any devolved policy, visa, or funding differences that affect international students. 3. The competitive landscape: how Cardiff and other Welsh universities compare with similar UK institutions on international student numbers, fees, and subject strengths. 4. The policy and regulatory context: recent UK Government and Welsh Government decisions on student visas, the Graduate Route, and international education strategy that could affect recruitment. 5. Three to five practical opportunities or risks the briefing suggests we should plan for, with the evidence behind each. Cite your sources throughout so I can verify and follow up on the most relevant ones.

Then reuse the report

🎙️ Audio

Turn it into a two-person podcast to listen to on the go.

🌐 Web page

Generate a clean, shareable page for clients or stakeholders.

📊 Infographic

Convert dense figures and deadlines into a scannable visual.

Nano Banana

Work images that actually work

Part 4

Nano Banana (yes, that really is the name) is the image tool built into Gemini. It fixes the two things image tools always fumbled: it can put accurate, readable text inside an image, and it can edit a picture you already have without wrecking the rest of it.

📈 Notes to diagram

Paste rough bullets, get a clean visual for a slide, poster or report.

✂️ Edit what you have

Swap a background, remove clutter, or correct the text on an existing image.

🏴󠁧󠁢󠁷󠁬󠁳󠁿 Bilingual posters

Ask for a Welsh version of a poster: it translates the text inside the image.

🖼️ Made-to-purpose images

A header image for a report or intranet page, without the stock photo hunt.

Try these two

The first needs nothing but your notes. For the second, upload an image (a poster, a slide, a photo) along with the prompt.

Create a clean, simple diagram I can drop straight into a PowerPoint slide. White background, one accent colour, large readable labels, no clutter. It should show this process: [PASTE YOUR ROUGH BULLET POINTS HERE]
Here is our event poster. Keep the layout, images and style exactly as they are, but change the date to [NEW DATE] and correct any spelling mistakes. Then create a second version with all the text translated into Welsh.

One habit before anything goes out the door: read every word in an AI image carefully, and take extra care with pictures of real people and places.

Gemini Canvas

Draft it, then shape it

Part 4

Canvas turns a chat into a finished document, one you would actually send. Find it in the tools menu under the box where you type. Your draft opens beside the chat: highlight any sentence, say what to change, and only that part is rewritten.

1 Open Canvas, paste your notes
2 Get a full first draft
3 Highlight and refine, line by line
Turn my rough notes below into a one-page briefing for staff. Plain English, a clear title, three short sections with headings, and a "what you need to do" list at the end. [PASTE YOUR NOTES OR A POLICY DOCUMENT HERE]

Then try the magic bit: highlight one paragraph and type "make this friendlier". When the draft is right, ask "turn this into a five-question quiz to test my team" and watch it build one.

Canvas can also build small working tools

Describe the tool you want and Canvas writes it and previews it live. You never need to read the code. Two to try:

For professional services: an expense claim estimator.
Act as an expert web developer. I need a single-page HTML application using Tailwind CSS and vanilla JavaScript. Build an 'Expense Claim Estimator' for university staff. It should have inputs for travel zones (e.g. UK, Europe, International), number of nights, and a toggle for 'with receipts' vs 'flat rate per diem'. It should dynamically calculate the estimated total claim based on standard mock university rates as inputs change. Include a button to 'Generate Summary for Finance'. Ensure all JavaScript is contained within the single HTML file. Make it look clean and modern with a professional blue and white colour scheme.
For teaching and learning: an interactive concept visualiser.
Act as an expert web developer and instructional designer. Create a single-page HTML web app using Tailwind CSS and vanilla JavaScript to teach psychology students about the Normal Distribution (bell curve). Include interactive sliders allowing the user to adjust the 'Mean' and 'Standard Deviation'. Include a canvas element that dynamically redraws the bell curve based on those slider inputs. Add short text descriptions explaining what standard deviation actually means in plain English, updating dynamically if possible. The design should be clean, engaging, accessible, and require no external assets other than Tailwind via CDN.
Before you go

Responsible AI & Your AI Road Map

🤝 Human-Centric
  • Augment, don't replace.
  • Protect critical skills.
  • Target repetitive tasks first.
✅ Practical Checks
  • Always double-check the output.
  • Be transparent: "I used AI for this."
  • Use IT-approved tools only.
⚖️ Responsibility
  • Watch for bias and hallucinations.
  • You own the result, not the AI.
  • Use AI in line with your values.
No one knows everything about AI. The AI Community Approach. 📄 Read our paper ↗
🧠 Mindset

Be curious. Ask: "Could this be useful?"

💬 Conversations

Shared learning, wins, and fails.

🧪 Experiments

Learn by doing. Build expertise.

🤝 Community

Share solutions and stay agile, together.

Take it Home

Bring 'AI in the Workplace' to your team

Two ways to share what you've learnt today.

Workshop Flyer Share with your manager One-pager covering everything they need to book us on-site.
Train the Trainer We train your trainers Equip your internal champions to deliver this workshop.

Questions?

Want us to deliver the workshop at your site or train your trainers?

thematts@cardiff.ac.uk