Amanoba Logo
Back to Course

Error handling and fallbacks

Day 18 of 30 · Generative AI 2026: Build AI Apps and Agents

One-liner: Add clear recovery paths when the AI fails.
Time: 20 to 30 min
Deliverable: Fallback Logic and Error Responses

Learning goal

You will be able to: Design fallback responses that keep users moving.

Success criteria (observable)

  • At least 3 error types are handled.
  • Each error has a user safe message.
  • A retry path is defined.

Output you will produce

  • Deliverable: Fallback Logic and Error Responses
  • Format: Error table plus messages
  • Where saved: Repo and course folder notes

Who

Primary persona: Digital nomad designing error recovery Secondary persona(s): Users who need clarity Stakeholders (optional): Collaborators

What

What it is

A small set of user friendly responses for common failure cases. It keeps the user on track even when the AI fails.

What it is not

It is not full incident response or SLA planning. It is a practical layer for a small app.

2-minute theory

  • Errors are inevitable, so recovery must be planned.
  • Clear messages reduce frustration and support load.
  • Retry paths help users succeed without leaving.

Key terms

  • Fallback: A safe response when the primary action fails.
  • Retry path: A clear way for users to try again.

Where

Applies in

  • API responses
  • UI error states

Does not apply in

  • Marketing pages

Touchpoints

  • Error screens
  • Logs
  • Support messages

When

Use it when

  • You ship a user facing AI feature
  • You see error logs in testing

Frequency

Once per feature, revise when new errors appear

Late signals

  • Users report being stuck
  • Support tickets mention unclear errors

Why it matters

Practical benefits

  • Better user trust
  • Fewer rage quits
  • Clearer debugging

Risks of ignoring

  • Confused users
  • Higher support load

Expectations

  • Improves: resilience and trust
  • Does not guarantee: zero errors

How

Step-by-step method

  1. List the top 3 error cases.
  2. Write a user safe message for each.
  3. Add a retry path or alternate action.
  4. Log each error with context.

Do and don't

Do

  • Be polite and specific
  • Offer a next step

Don't

  • Show stack traces to users
  • Blame the user

Common mistakes and fixes

  • Mistake: Generic error text. Fix: Add a clear action.
  • Mistake: No retry. Fix: Add a retry button or link.

Done when

  • Three error cases are defined.
  • Messages are clear and polite.
  • Retry paths are in place.

Guided exercise (10 to 15 min)

Inputs

  • Error logs from tests
  • Common failure cases

Steps

  1. Pick three error types.
  2. Write a user message for each.
  3. Add a retry path.

Output format

Field Value
Error type
User message
Retry action
Log field

Pro tip: Short messages with a next action perform best.

Independent exercise (5 to 10 min)

Task

Rewrite one error message to be shorter and clearer.

Output

Revised error message.

Self-check (yes/no)

  • Are top error types listed?
  • Are messages user safe?
  • Is there a retry action?
  • Are errors logged?

Baseline metric (recommended)

  • Score: 3 of 4 checks met
  • Date: 2026-02-06
  • Tool used: Notes app

Bibliography (sources used)

  1. Error Messages Best Practices. NNGroup. 2024-01-01. Read: https://www.nngroup.com/articles/error-message-guidelines/

  2. Reliable Systems. Google SRE. 2024-01-01. Read: https://sre.google/books/

Read more (optional)

  1. Designing Error States Why: Patterns for helpful error messages. Read: https://www.nngroup.com/articles/error-message-guidelines/
Day 18: Error handling and fallbacks | Generative AI 2026: Build AI Apps and Agents | Amanoba