Tickets vary widely, and each type needs its own instructions. Which pattern keeps prompts focused?
Why B is the answer
Routing
Routing classifies first, then uses a path built for that type, so instructions for one type do not crowd out the others.
In detail
- Routing classifies the input first, then sends it down a path built for that type.
- Each path holds only the instructions it needs, so one ticket type does not crowd out another.
- It also lets you tune or test each path on its own.
- A cheaper model can often do the classification, while the specialist path uses what it needs.
Use case
An illustrative scenario, not a real incident.
A help desk receives billing, technical and account tickets. One prompt covering all three grows long and inconsistent. A small classifier routes each ticket to a short prompt with the right instructions and tools.
See it in different ways
flowchart TD
A{"Are the steps known?"} -->|Yes| B{"Does the input type change the path?"}
A -->|No| C["Orchestrator-workers or agent"]
B -->|Yes| D["Routing"]
B -->|No| E["Prompt chaining"]mindmap
root((Workflow patterns))
Fixed path
Prompt chain
One of several paths
Routing
Same task many times
Parallel
Path found at run time
Orchestrator workers
Quality loop
Evaluator optimizerflowchart TD
Q{"What type of ticket?"}
Q -->|"Billing"| A0["Billing prompt and tools"]
Q -->|"Technical"| A1["Technical prompt and tools"]
Q -->|"Account"| A2["Account prompt and tools"]Key point: Route first, then use a focused prompt.
Why the other options are wrong
- A. Evaluator-optimizer
- Evaluator-optimizer improves one output through feedback rounds. It does not separate different types of input.
- C. Prompt chaining through one shared prompt that lists every ticket type
- One shared prompt with every ticket type is the crowding problem the question describes.
- D. A single agent holding all of the instructions in its system prompt
- A single agent with all the instructions has the same problem, and every request pays for all of that text.