Capabilities of Generative AI
What it takes off your plate
The story: A good office assistant types up the boring forms, notices "sales always dip in March" so you can plan, and handles the fiddly paperwork so you can spend your day on the creative parts of the job.
In AI/AWS terms: Generative AI can automate tedious work like data entry, find patterns and trends to support decisions, and take over complex tasks so people can focus on creative work.
For the exam: Generative AI frees people from repetitive work and supports decisions.
The seven capabilities
The story: Picture the best all-round assistant you could hire:
- They can switch from writing a birthday card to a legal letter.
- They answer the moment you ask.
- You say "make me a poster for the bake sale" and it's done, with no step-by-step instructions.
- They pitch ideas you hadn't thought of.
- They only need a couple of examples of your style, because they already know the basics.
- They remember that you like short emails and your colleague likes long ones.
- They can write 500 product descriptions before lunch.
In AI/AWS terms:
| Assistant trait | Capability | Meaning |
|---|---|---|
| Birthday card to legal letter | Adaptability | Handles many tasks and domains, tailoring content to the context |
| Answers right away | Responsiveness | Generates content in real time, which suits chatbots and customer service |
| "Make me a poster" | Simplicity | Turns complex tasks into simple requests, like writing code or a document from a short prompt |
| New ideas | Creativity and exploration | Produces new ideas, designs, and solutions |
| Needs only a few examples | Data efficiency | Can learn from small amounts of data because it builds on a pre-trained model |
| Knows your preferences | Personalization | Creates content tailored to individual users |
| 500 descriptions before lunch | Scalability | Produces large amounts of content quickly |
For the exam: The seven capabilities are adaptability, responsiveness, simplicity, creativity and exploration, data efficiency, personalization, and scalability.
What is not on the list
The story: That brilliant assistant still sometimes states a wrong date with total confidence. And "how many customers bought something" is how you judge the shop, not a skill the assistant has.
In AI/AWS terms: Conversion rate and cross-domain performance are business metrics, not capabilities. Accuracy is a weak spot of generative AI because of hallucinations.
For the exam: Accuracy is not a generative AI capability. Conversion rate and cross-domain performance are business metrics.