AI, in the terms of your firm
Twenty-two words people use about AI, explained the way an advice firm already works.
What are you actually using?
6 terms
- ModelThe trained prediction engine inside an AI product.
- AI-nativeSoftware built around AI for its main job, rather than adding AI to a few existing features.
- AI-enabledSoftware that adds AI to particular tasks while the product's main work still runs without those features.
- Chat app vs AI inside a productA chat app is a general AI product for conversation; AI inside a product works within software built for a particular firm task.
- Model providerThe company that develops and supplies the AI model used inside a product. It may differ from the product's vendor.
- Where client data goesThe path client information takes through an AI product and its suppliers, including who receives copies and where they are kept.
Why does it get things wrong?
1 of 4 written
- HallucinationA statement an AI makes up and presents as fact. It reads just like the accurate lines around it.
- Training · Non-determinism · Sycophancy
What sits around it?
8 terms
- HarnessThe software around an AI model. It decides what the model is given, what it may do, and which checks run on its work.
- PromptThe request sent to an AI for one piece of work, typed by a person or chosen from a saved list.
- InstructionsStanding directions a product sends the AI with its work. They guide the answer but do not enforce a rule.
- ContextThe material an AI is given for one piece of work. Everything it knows about your client comes from it.
- MemoryInformation an AI product saves from earlier work and gives the AI again later. It does not change the model.
- ToolAn action a product lets its AI request, such as finding a record or adding a task.
- AgentAn AI that chooses and takes steps towards a goal using actions a product gives it.
- PermissionThe access and action limits software enforces for an AI, including when a person must approve an action.
How do you check its work?
2 of 4 written