Working with AI Assistants
By the end of this lesson you will know the main AI assistants, understand what your account plan means for confidentiality, and be able to write prompts that produce usable work.
The tools
ChatGPT (OpenAI), Gemini (Google) and Claude (Anthropic) are all AI assistants built on large language models. All of them write, summarise, explain, translate and draft code. They differ in how they behave, what they are good at, and — most importantly for you — how they treat what you type into them.
At First Leadership, the approved assistant is Claude. Use it for your work, and nothing else.
Account plans — the part that actually matters
| Plan | What it means for your data |
|---|---|
| Free / standard | A consumer account on consumer terms. Your conversations may be used to improve the provider's models. |
| Paid individual | More capacity and better models — but still a personal consumer account. |
| Organisation / business / enterprise | A company-controlled account with contractual terms that your data is not used for training, plus administrator controls and audit records. |
The difference between these is not the price. It is who is responsible for the data, and what the provider is permitted to do with it.
A personal paid account is not a company account. Paying more does not make it safe for client data.
This is the reason behind the rule you already acknowledged in the Non-Disclosure Agreement: never put client or company data into a public assistant. Which account you are signed into decides whether anything you type stays private.
Prompt basics
- Say what you want, who it is for, and what form the answer should take.
- Give it the context it needs — and only context you are permitted to share.
- Ask for a draft, not a final answer.
- Iterate. Your second prompt matters more than your first.
Weak: write about inventory
Strong: Draft a 200-word summary of stock turnover for a manager with no accounting background. Plain language, no jargon.
Checking the output
This is the step people skip. Assistants produce fluent, confident text that is sometimes wrong.
- Verify every fact, number, name and reference against a real source.
- Never submit output you have not read in full.
- You are responsible for what you hand in. “The AI wrote it” is not a defence, and a mentor spots unchecked output immediately.
What AI is bad at
- Anything needing current internal knowledge it does not have
- Arithmetic on your data — check it yourself
- Knowing what it does not know
- Deciding what matters
Checklist
- Working in Claude, not another assistant
- Nothing confidential typed into any public account
- Prompt states the audience and the output format
- Every fact in the output verified
- Read in full before submitting
The main assistants, what your account plan means for confidentiality, and how to prompt and check output.
لا توجد تعليقات حالياً.