Same Brand, Different Deal
Welcome back! Last time we looked at what AI can and can't do in agency work. Today's question comes before any of that: which AI tools are you actually allowed to use at work?
It sounds like admin, I know. But it matters more than most people think.
Here's a comparison you'll recognise. Think about a personal Gmail account and your work email. Both send messages, and they might even look alike. But your work email sits under a contract your firm signed. It has security settings, it keeps records, and it gives your firm some control. Your personal account has none of that.
AI tools work the same way. The same brand, such as ChatGPT, Gemini, Claude or Copilot, usually comes in a consumer version for individuals and a business or enterprise version for organisations. They can look nearly identical on screen, yet they deal with your data in very different ways.
What the Help Pages Actually Say
Let's look at four tools. These details come from each vendor's own help pages. Settings do change, though, so always check the current page yourself.
ChatGPT. OpenAI says that on the personal plans (Free, Plus and Pro), your chats can be used to improve its models by default. You can switch that off under Settings, then Data Controls, then "Improve the model for everyone". On ChatGPT Business and Enterprise, OpenAI says it doesn't use your inputs and outputs to train its models by default.
Claude. Since Anthropic changed its consumer terms in 2025, people on the Free, Pro and Max plans choose whether their chats can be used to improve Claude. Those consumer terms don't apply to Claude for Work (the Team and Enterprise plans), and Anthropic says it doesn't use that business data to train its models.
Gemini. With a personal Google account, Gemini Apps conversations can be used to improve Google's products, and some are read by human reviewers. With a qualifying Google Workspace edition, Google says your prompts and content aren't reviewed by people or used to train its models outside your organisation without permission.
Microsoft Copilot. Heads up: Microsoft has recently renamed Microsoft 365 Copilot to simply "Microsoft Copilot", so you may see both names for a while. If you're signed in with your work account, Microsoft's "enterprise data protection" applies. It says your prompts and responses aren't used to train its foundation models.
So what's the pattern? The business versions come with a contract and controls your firm can manage. The consumer versions mostly depend on settings each person picks for themselves. That's why your firm cares which one you're using.
Finding Your Firm's AI Policy
So, does your firm have an AI policy? Lots of agents honestly don't know.
Start with the obvious places: the staff handbook, the intranet, your compliance manual, or an email from your branch manager or IT team. Sometimes it's buried inside an IT or data protection policy rather than standing alone.
You're not alone if your firm is still working this out. In January 2026, Alto's agency trends report, based on a survey of 250 estate and letting professionals, found that about a third of agents felt nervous or unsure about using AI. Plenty of firms are writing their rules as they go.
Once you find the policy, sum it up in a few lines for yourself. Which tools are approved? Which tasks are they approved for? Is anything banned outright? Who do you ask when you're not sure?
No Policy Yet? Ask These Questions
If there's nothing written down, don't take silence as a yes. Take these questions to your manager instead:
- Which AI tools are we allowed to use, and on which accounts: work or personal?
- Do we have a business licence for any of them?
- Which tasks are fine, and which are off-limits?
- Do I have to check or sign off AI output before it goes to a client?
- Who do I ask if I'm unsure?
You'll come across as careful, not awkward. Most managers will be glad somebody asked.
Is This Tool Approved for This Task?
This is the habit to build: check the tool and the task together before you start.
Let's try one. Say your firm has approved Microsoft Copilot on work accounts for drafting marketing copy. Can you use it to draft a social post about a new instruction? Yes, that tool is approved for that task.
Now say you'd like to use your personal ChatGPT Plus account for the same post because you prefer how it writes. Is that allowed? No. It's a fine tool, but it isn't the approved one, and it's on a personal account.
Here's a quick routine. First, name the task. Second, check which tools the policy approves for it. Third, make sure you're signed in to the business account, not a personal one. Only then do you start.
What about a brand-new tool that's not on the list yet? Ask before you use it, not after.
What you can safely type into an approved tool is a separate question, and it's a big one. We'll cover it properly in the next level.
Try It Yourself
This week, find your firm's AI policy, or confirm there isn't one. Write a five-line summary of it, or send your manager the questions above. Then open the help page for any tool you use and check what it says about how your data is handled.
Key Takeaways
- Consumer and business versions of the same AI tool can deal with your data very differently.
- Check each vendor's current help pages, because settings and product names change.
- Find your firm's policy and sum it up: approved tools, approved tasks, bans, and who to ask.
- No policy doesn't mean you can use anything. Ask your manager first.
- Before you start, check that the tool is approved for that task and that you're on the business account.