An AI agent is an automated AI that can carry out multi-step tasks on your behalf — browsing the web, filling in forms, running searches, and logging results — without you having to prompt it at every step. For example, an AI agent could find candidate LinkedIn profiles, draft personalised outreach emails, and log everything to a spreadsheet, all in one run. Tools like ChatGPT Operator and Anthropic's Claude are beginning to offer agentic features. Relevant for professionals looking to automate complex, multi-stage workflows.

What is AI Hallucination?

An AI hallucination is when an AI tool like ChatGPT or Claude confidently states something that is factually incorrect — including fabricated statistics, false citations, or invented details. Hallucinations are more common when asking for specific facts, dates, or figures. The solution: always use Perplexity AI for factual research (it cites live sources), and verify any specific numbers from ChatGPT before using them professionally.

AI literacy is the ability to understand what AI tools can and cannot do, and to use them effectively and responsibly at work. It does not mean knowing how to code or understanding machine learning — it means knowing which tool to use for which task, how to write a prompt that gets useful results, and when to verify AI output before acting on it. AI literacy is increasingly listed as a required skill in job descriptions across marketing, HR, finance, and administration roles.

What is AI Workflow Automation?

AI workflow automation refers to using AI tools to systematically handle repetitive professional tasks — such as weekly report drafting, email triage, meeting summarisation, and content repurposing — reducing manual effort and reclaiming 5–10 hours per week. The AI Survival Kit teaches non-technical professionals to automate their three biggest time-wasters using free tools, without writing any code.

What is Chain-of-Thought Prompting?

Chain-of-thought prompting means adding a simple instruction like "think step by step" to your prompt before asking an AI to tackle something complex. This technique dramatically improves accuracy on tasks that require logical reasoning — such as analysing data, building a decision framework, or comparing several options. Instead of jumping to a conclusion, the AI walks through its reasoning first, which reduces errors. Particularly useful for HR scoring rubrics, marketing attribution analysis, or financial scenario planning.

ChatGPT is an AI language model developed by OpenAI. For non-technical professionals, it is best used for drafting emails, writing campaign briefs, generating ad copy, summarising reports, and brainstorming ideas. A free tier is available at chat.openai.com. The paid GPT-4o model ($20/month) handles more complex tasks. 67% of marketing teams use ChatGPT weekly (Marketing AI Report, 2026).

Reference

AI Glossary for Professionals

Plain-English definitions of 37 AI terms every professional needs in 2026 — no technical background required. From ChatGPT to context windows, defined in plain English.

AI Agent

AI Agent: An AI agent is an automated AI that can carry out multi-step tasks on your behalf — browsing the web, filling in forms, running searches, and logging results — without you having to prompt it at every step. For example, an AI agent could find candidate LinkedIn profiles, draft personalised outreach emails, and log everything to a spreadsheet, all in one run. Tools like ChatGPT Operator and Anthropic's Claude are beginning to offer agentic features. Relevant for professionals looking to automate complex, multi-stage workflows.

AI Hallucination

AI Hallucination: An AI hallucination is when an AI tool like ChatGPT or Claude confidently states something that is factually incorrect — including fabricated statistics, false citations, or invented details. Hallucinations are more common when asking for specific facts, dates, or figures. The solution: always use Perplexity AI for factual research (it cites live sources), and verify any specific numbers from ChatGPT before using them professionally.

AI Literacy

AI Literacy: AI literacy is the ability to understand what AI tools can and cannot do, and to use them effectively and responsibly at work. It does not mean knowing how to code or understanding machine learning — it means knowing which tool to use for which task, how to write a prompt that gets useful results, and when to verify AI output before acting on it. AI literacy is increasingly listed as a required skill in job descriptions across marketing, HR, finance, and administration roles.

AI Workflow Automation

AI Workflow Automation: AI workflow automation refers to using AI tools to systematically handle repetitive professional tasks — such as weekly report drafting, email triage, meeting summarisation, and content repurposing — reducing manual effort and reclaiming 5–10 hours per week. The AI Survival Kit teaches non-technical professionals to automate their three biggest time-wasters using free tools, without writing any code.

Chain-of-Thought Prompting

Chain-of-Thought Prompting: Chain-of-thought prompting means adding a simple instruction like "think step by step" to your prompt before asking an AI to tackle something complex. This technique dramatically improves accuracy on tasks that require logical reasoning — such as analysing data, building a decision framework, or comparing several options. Instead of jumping to a conclusion, the AI walks through its reasoning first, which reduces errors. Particularly useful for HR scoring rubrics, marketing attribution analysis, or financial scenario planning.

ChatGPT Official site ↗

ChatGPT: ChatGPT is an AI language model developed by OpenAI. For non-technical professionals, it is best used for drafting emails, writing campaign briefs, generating ad copy, summarising reports, and brainstorming ideas. A free tier is available at chat.openai.com. The paid GPT-4o model ($20/month) handles more complex tasks. 67% of marketing teams use ChatGPT weekly (Marketing AI Report, 2026).

Claude Official site ↗

Claude: Claude is an AI assistant developed by Anthropic. For professionals, it is best used for analysing long documents (up to 200,000 tokens), writing tone-sensitive brand copy, summarising lengthy reports, and tasks requiring careful reasoning. Claude handles longer documents than ChatGPT, making it ideal for contract review or 50-page report summarisation. A free tier is available at claude.ai.

Claude 3.5 / Claude 4 Official site ↗

Claude 3.5 / Claude 4: Claude 3.5 and Claude 4 refer to the model versions available inside Anthropic's Claude assistant. Claude 4 Sonnet is the current recommended model for professional writing, long-document analysis, and tone-sensitive tasks. Like ChatGPT's GPT-4o, the model version you get depends on whether you are on the free or paid tier ($20/month). For most non-technical professionals, the free tier of Claude 4 Sonnet is sufficient for everyday tasks such as summarising reports, drafting emails, and reviewing documents.

Co-Pilot (Microsoft) Official site ↗

Co-Pilot (Microsoft): Microsoft Copilot is Microsoft's AI assistant built directly into Microsoft 365 applications — Word, Excel, Outlook, PowerPoint, and Teams. It can draft documents, summarise email threads, generate Excel formulas, and create PowerPoint slides from a brief. Copilot requires a Microsoft 365 Copilot licence on top of your existing Microsoft 365 subscription (approximately $30 per user per month). It is the strongest choice for professionals whose entire workflow lives inside the Microsoft ecosystem.

Context Window

Context Window: A context window is the maximum amount of text an AI tool can process in a single conversation — including all previous messages and documents you have shared. A larger context window means the AI can 'remember' more of a long document or conversation. Claude has one of the largest context windows (200,000 tokens, roughly 150,000 words), making it the best tool for analysing long reports or contracts.

Copilot vs ChatGPT

Copilot vs ChatGPT: Microsoft Copilot and ChatGPT are the two most commonly compared AI tools for professionals. Copilot integrates directly into Office apps (Word, Excel, Outlook) and is ideal if your workflow is fully Microsoft-based — but it costs around $30 per user per month on top of your existing licence. ChatGPT is a standalone browser and mobile tool that costs $20/month for the Pro tier and works across any environment. For most non-technical professionals starting with AI, ChatGPT offers more flexibility, a lower cost of entry, and a broader range of use cases.

Few-Shot Prompting

Few-Shot Prompting: Few-shot prompting means giving an AI two or three examples of exactly the output you want before asking it to produce more. This is more reliable than a bare instruction when you need a very specific format — for example, showing the AI one completed performance review before asking it to draft five more, or pasting in one social media post in your brand voice before asking for a week's worth of content. The AI uses your examples as a template. Compare with zero-shot prompting, which gives no examples at all.

Fine-Tuning

Fine-Tuning: Fine-tuning is the process of retraining an existing AI model on a specific dataset so it becomes better at a particular task or adopts a specific tone and style. For example, a company might fine-tune a model on thousands of past customer service emails so it sounds like their brand. Fine-tuning requires technical expertise, significant data, and is managed by AI engineers — not something non-technical professionals need to do. The AI Survival Kit focuses exclusively on off-the-shelf tools that require no configuration.

Gemini (Google) Official site ↗

Gemini (Google): Gemini is Google's AI assistant, available for free at gemini.google.com and built into Google Workspace apps including Gmail, Google Docs, and Google Sheets. It is a credible alternative to ChatGPT — particularly for professionals who already live in Google's ecosystem. Gemini can summarise Gmail threads, draft Docs content, and generate formulas in Sheets. The free tier is capable; Gemini Advanced ($19.99/month via Google One AI Premium) unlocks more advanced models and longer document analysis.

GPT-4o Official site ↗

GPT-4o: GPT-4o is OpenAI's current recommended model within ChatGPT, available on the free tier with usage limits and without limits on the $20/month Pro plan. GPT-4o handles text, image analysis, and voice. For professional writing tasks — emails, briefs, reports — GPT-4o is the recommended model in the AI Survival Kit.

GPT-4o Mini Official site ↗

GPT-4o Mini: GPT-4o Mini is a faster, more cost-efficient version of GPT-4o that OpenAI uses automatically for some free-tier ChatGPT tasks, particularly when the system is under heavy load. It is slightly less capable than full GPT-4o for nuanced writing or complex reasoning, but perfectly adequate for simple tasks like reformatting text, generating short lists, or answering basic questions. If you are on the free tier and notice shorter or less detailed responses, you may be talking to GPT-4o Mini. Upgrading to ChatGPT Pro ($20/month) gives consistent access to full GPT-4o.

Grounding

Grounding: Grounding means connecting an AI tool to real, current data sources — such as live web search or your company's own documents — so that its answers are based on actual information rather than its training data alone. Grounding significantly reduces AI hallucinations. Perplexity AI grounds every response in live web results and cites sources automatically. Microsoft Copilot grounds responses in your Microsoft 365 files and emails. When using a tool without grounding (like the standard ChatGPT without browsing enabled), always verify factual claims independently.

Image Generation AI

Image Generation AI: Image generation AI creates visual images from a text description (called a prompt). Tools include DALL·E — available inside ChatGPT on the Pro plan — and Midjourney, which is accessed via Discord. For non-technical professionals, these tools are useful for creating presentation mockups, social media graphics, concept visuals for briefs, and placeholder artwork without a designer. Quality has improved dramatically: a sentence like 'a professional woman reviewing data on a laptop in a bright modern office, photorealistic' now produces usable results in seconds.

Inference

Inference: Inference is the technical term for what happens when an AI model generates a response — the computation that runs between when you press send and when the reply appears. Faster inference means faster responses. This term is worth knowing when comparing tools: some AI providers offer faster inference (near-instant replies) while others may have noticeable delays, especially at peak times. For most professional tasks, speed differences are minor. Where inference speed matters most is in high-volume tasks like processing many documents in sequence.

Large Language Model (LLM)

Large Language Model (LLM): A large language model (LLM) is the AI technology that powers tools like ChatGPT, Claude, and Perplexity AI. LLMs are trained on vast amounts of text data and can generate human-like writing, answer questions, summarise documents, and complete tasks described in plain English. You do not need to understand how LLMs work to use them effectively — the AI Survival Kit teaches practical use without technical concepts.

Model

Model: In AI, a model refers to the specific AI system you are interacting with — for example, GPT-4o (inside ChatGPT), Claude 4 Sonnet (inside Claude), or Gemini 1.5 Pro (inside Gemini). Different models have different strengths, speeds, and costs. Choosing the right model for a task matters: Claude 4 Sonnet is strong for long-document analysis; GPT-4o is versatile for writing and ideation; Gemini is well-integrated with Google Workspace. The AI Survival Kit recommends specific models for specific professional tasks.

Multimodal AI

Multimodal AI: A multimodal AI can process multiple types of input — not just typed text, but also images, PDFs, audio files, and screenshots. GPT-4o and Claude 4 Sonnet are multimodal: you can paste a screenshot of a spreadsheet and ask questions about the data, upload a PDF contract and ask for a plain-English summary, or share an image of a competitor's ad and ask for a critique. For non-technical professionals, multimodal AI significantly reduces the time spent copying and pasting content into prompts.

Notion AI Official site ↗

Notion AI: Notion AI is an AI assistant built directly into Notion workspaces. For professionals, it is best used for summarising meeting notes, generating project briefs, organising campaign timelines, and drafting structured documents from bullet points. Requires an active Notion workspace and costs $10/month as an add-on. Best for teams already using Notion.

NotebookLM Official site ↗

NotebookLM: NotebookLM is a free AI research tool from Google, available at notebooklm.google.com. You upload your own documents — research reports, contracts, meeting transcripts, PDFs — and NotebookLM answers questions about them with citations, creates summaries, and even generates audio overviews. Unlike ChatGPT, it only draws from the documents you upload, which means its answers are grounded in your specific materials and far less likely to hallucinate. It is one of the most immediately useful free AI tools for professionals who deal with large volumes of documents.

Output Formatting

Output Formatting: Output formatting means telling an AI exactly how you want its response structured — for example, as bullet points, a numbered list, a two-column table, markdown headers, or a JSON object. Without formatting instructions, AI tools often produce dense paragraphs that are hard to use directly. Always include a format instruction in your prompt: 'Respond as a bullet point list', 'Format this as a table with columns for Task, Owner, and Deadline', or 'Use H2 headers for each section.' Specifying format is one of the simplest ways to make AI output immediately usable in professional documents.

Perplexity AI Official site ↗

Perplexity AI: Perplexity AI is an AI-powered research engine that searches the live web and cites sources automatically in every response. For professionals, it is best used for competitive landscape research, industry trend reports, fact-checking campaign claims, and real-time market intelligence. Unlike ChatGPT, Perplexity always shows its sources. A free tier is available at perplexity.ai.

Perplexity Pro Official site ↗

Perplexity Pro: Perplexity Pro is the paid subscription tier of Perplexity AI, priced at $20 per month. It adds a significantly higher number of searches per day, the ability to upload files (PDFs, CSVs, images) for analysis, and access to more powerful underlying models including GPT-4o and Claude. For most professionals doing occasional research tasks, the free tier of Perplexity AI is sufficient. The Pro tier is worth considering if you use Perplexity as your primary research tool or regularly need to analyse documents alongside web research.

Prompt Engineering

Prompt Engineering: The practice of writing specific, structured instructions for AI tools like ChatGPT and Claude to produce accurate, useful outputs for professional tasks — no coding required. A well-crafted prompt tells the AI the role to play, the task to complete, the format to use, and any constraints. The AI Survival Kit includes 30 pre-built prompts for common marketing and HR tasks.

RAG (Retrieval-Augmented Generation)

RAG (Retrieval-Augmented Generation): RAG stands for Retrieval-Augmented Generation. It is a behind-the-scenes technique where an AI searches a specific knowledge base — such as a company's internal documents or a set of uploaded files — before generating its response, grounding the answer in real source material rather than general training data. This dramatically reduces hallucinations for specialised topics. NotebookLM uses RAG on your uploaded documents; enterprise AI search tools use RAG on company knowledge bases. As a non-technical professional, you do not configure RAG yourself — but understanding it explains why tools like NotebookLM are more reliable for document-specific questions than standard ChatGPT.

Role Prompting

Role Prompting: Role prompting means instructing an AI to adopt a specific expert persona before you ask your question — for example: 'You are an experienced HR manager with 15 years in talent acquisition.' This simple technique shifts the AI's tone, vocabulary, depth of reasoning, and frame of reference to match the expertise you need. Role prompting is particularly effective when you want a professional perspective: 'You are a senior marketing strategist…', 'You are an executive assistant who specialises in scheduling…', or 'You are an experienced finance business partner…' It is one of the highest-impact techniques in the AI Survival Kit.

Summarisation

Summarisation: Summarisation is the use of AI to condense long documents — reports, contracts, email threads, meeting transcripts, research papers — into the key points you actually need. It is one of the most immediately useful AI tasks for non-technical professionals. Paste in a 50-page industry report and ask Claude to summarise the five most relevant findings for your team; upload a dense contract to NotebookLM and ask it to flag non-standard clauses. Claude is particularly strong for summarisation because it has one of the largest context windows available (up to 200,000 tokens), allowing it to process very long documents in a single pass.

System Prompt

System Prompt: A system prompt is an instruction given to an AI tool at the start of a conversation that sets its role, tone, and constraints — before your actual question. For example: 'You are an expert marketing copywriter. Write in a professional but approachable tone. Keep responses under 200 words.' Using a system prompt consistently produces more reliable, on-brand outputs. The AI Survival Kit includes system prompt templates for common marketing and HR tasks.

Temperature (AI Setting)

Temperature (AI Setting): Temperature is a parameter that controls how creative or predictable an AI's outputs are. A low temperature setting makes the AI more consistent and factual — it sticks closer to the most likely answer. A high temperature setting makes responses more varied and creative — useful for brainstorming, but riskier for tasks requiring accuracy. Most consumer AI tools (ChatGPT, Claude, Gemini) set temperature automatically and do not expose it as a user-facing control. You do not need to adjust it. The concept is useful to understand if AI responses feel either too repetitive or too unpredictable.

Token

Token: A token is the basic unit of text that AI models use to process and generate language. Tokens do not map exactly to words: roughly 1 token equals three-quarters of a word in English, so a 1,000-word document is approximately 1,300 tokens. Token counts become relevant when you are working with very long documents and hitting context window limits — for example, Claude's 200,000-token context window can handle roughly 150,000 words in a single conversation. For everyday professional use, you rarely need to count tokens; they matter most when a tool tells you a document is too long to process.

Vibe Coding / AI Coding

Vibe Coding / AI Coding: Vibe coding is a term for using AI to write code — or build simple software tools — without knowing how to code yourself. You describe what you want in plain English and the AI writes the code. Tools include GitHub Copilot (for developers), Cursor (a code editor with AI built in), and Replit Agent (which can build and deploy simple web apps from a text description). This is beyond the scope of the AI Survival Kit, which focuses on writing, research, and workflow tasks. However, it is increasingly accessible to non-technical professionals for simple automations: building a custom spreadsheet formula generator, a basic form, or a simple data cleaner.

Voice Mode (ChatGPT) Official site ↗

Voice Mode (ChatGPT): Voice Mode is a feature in the ChatGPT mobile app that lets you have a real spoken conversation with GPT-4o — the AI speaks and listens rather than displaying text. For non-technical professionals, it is useful for hands-free brainstorming during a commute, rehearsing a presentation or difficult conversation, capturing ideas without typing, or thinking through a problem out loud. Voice Mode is available on both the free and Pro tiers of the ChatGPT app on iOS and Android. The AI responds in natural spoken language and adapts to follow-up questions in real time.

Zero-Shot Prompting

Zero-Shot Prompting: Zero-shot prompting means asking an AI to complete a task without giving it any examples first. For example: 'Write a subject line for a re-engagement email campaign targeting lapsed subscribers.' Most professional tasks can be completed with zero-shot prompting when the prompt is specific enough. The AI Survival Kit teaches professionals to write zero-shot prompts that work reliably on the first attempt.

Want to use these tools in your actual job — with ready-made prompts and a 30-day plan?

Marketing Managers Kit → HR Professionals Kit →

Or browse free guides →