Fundamentals of GenAI · Topic 3 of 7
GenAI Use Cases
Identify common business use cases for generative AI such as summarisation, assistants, search and code generation, and match them to tasks.
- 4 min
- GenAI use cases include text, image, audio, video and code generation, summarisation, translation, search and assistants.
- Customer service agents and chatbots are among the most common examples.
- Summarisation and search help people handle large volumes of text.
What it is
GenAI use cases are the business tasks where creating or transforming content brings value. The exam gives a short scenario and asks you to spot the GenAI use case, or to see that GenAI is not needed.
Most use cases follow one of a few patterns: create, summarise, convert, answer or assist.
How it works
Main use case groups
| Use case | What the model does | Example |
|---|---|---|
| Text generation | Writes drafts, emails, product descriptions | Marketing copy |
| Summarisation | Shortens long content into key points | Meeting notes, long reports |
| Question answering | Answers questions in natural language | Staff help desk |
| AI assistants and chatbots | Holds a conversation and helps with tasks | Customer service agent |
| Translation | Converts text between languages | Multi language support |
| Code generation | Writes, explains and fixes code | Developer helper |
| Image generation | Creates pictures from a description | Ad concepts |
| Video and audio generation | Creates clips, voices or music | Training videos |
| Search | Finds answers by meaning, not just keywords | Enterprise knowledge search |
| Recommendation | Suggests items, with explanations | Product suggestions |
| Data extraction | Pulls structured fields from messy text | Reading contracts |
Customer service
GenAI chat agents can answer common questions at any hour, summarise a conversation for a human agent, and suggest replies. The exam often pairs this with a human hand off for hard or sensitive cases.
Content creation
Marketing teams use GenAI to produce first drafts and variants for different audiences. A person still reviews the result for accuracy, tone and legal fit. The model speeds up the draft but does not remove the need for review.
Knowledge work
Summarisation and search help staff cope with large document sets. A model that reads long reports and returns the key points saves time. Linking it to company documents makes answers relevant, and a later chapter explains RAG, which does this.
Software development
Code assistants suggest code, write tests, explain unfamiliar code and help with documentation. Amazon Q Developer is the AWS assistant for these tasks.
Fit and misfit
GenAI fits tasks with open ended or language based output, where a draft is useful and a person can check it. It fits poorly where a precise number or guaranteed fact is needed, or where a simple rule works.
| Good fit | Poor fit |
|---|---|
| Draft a reply to a customer | Calculate an exact tax amount |
| Summarise a long document | Provide a legally binding statement without review |
| Generate marketing images | Decisions that need full explanation and no error |
Choosing a service for the use case
For text tasks through a managed API, Amazon Bedrock is the main AWS answer. For a ready assistant over company data, Amazon Q Business is the answer.
For translation only, Amazon Translate is a focused option. A task specific service is often enough when you do not need open ended generation.
Combining with other AI services
GenAI often works with the services you saw earlier. Transcribe can turn a call into text, and a model can summarise it. Textract can read an invoice, and a model can explain the findings.
What to remember for the exam
- GenAI use cases include text, image, audio, video and code generation, summarisation, translation, search and assistants.
- Customer service agents and chatbots are among the most common examples.
- Summarisation and search help people handle large volumes of text.
- Code generation use cases include writing, explaining and fixing code, supported by Amazon Q Developer.
- GenAI suits language based, open ended tasks where a person can check the output.
- GenAI is a poor fit for exact calculations or tasks needing guaranteed accuracy.
- Amazon Bedrock gives managed access to models, and Amazon Q Business serves company knowledge assistants.
- Human review remains important for high stakes content.
Look for the verb. Draft, summarise, converse, translate, generate or explain point to GenAI, while predict a value or classify a record usually points to traditional ML.
Common mistakes
Using GenAI for exact calculations. Models predict text, so arithmetic and strict rules are better handled by normal software.
Skipping human review for public content. Draft output can contain errors or unwanted wording.
Treating summarisation as search. Summarisation shortens content you already have. Search finds the right content in the first place.
Choosing GenAI when a managed AI service is enough. Translation or speech conversion may be simpler with a task specific service.
Expecting an assistant to know private company data on its own. It needs a connection to that data.
Check yourself
Name two GenAI use cases that handle long documents.
Summarisation and question answering over the documents, often with search.
Why is GenAI a poor choice for a tax calculation?
Tax needs an exact result from fixed rules, and a model can produce plausible but wrong numbers.
Which AWS assistant helps developers write and explain code?
Amazon Q Developer.
Topic 3 of 7 in Fundamentals of GenAI
Test yourself on Fundamentals of GenAI
The full AWS AIF-C01 question bank has 849 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by AWS. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.