Kupima agent kwa scenarios

Day 17 of 30 · AI Jeneretivu 2026: Jenga App na Agent za AI

One-liner: Tumia scenarios halisi kupima tabia za agent kabla ya uzinduzi.
Time: 20 hadi 30 dakika
Deliverable: Seti ya Scenario za Majaribio ya Agent

Learning goal

You will be able to: Kuunda scenarios za majaribio na kupima agent kwa output thabiti.

Success criteria (observable)

Seti ina angalau scenarios 5 za matumizi halisi.
Kila scenario ina input na output inayotarajiwa.
Angalau scenario moja ina edge case.

Output you will produce

Deliverable: Seti ya Scenario za Majaribio ya Agent
Format: Jedwali la scenarios na matokeo
Where saved: Kwenye folda ya kozi ndani ya /generative-ai-2026-build-ai-apps-and-agents-sw/

Who

Primary persona: Digital nomad anayepima agent kabla ya uzinduzi Secondary persona(s): Watumiaji wanaotegemea majibu thabiti Stakeholders (optional): Washirika wa ujenzi

What

What it is

Orodha ya scenarios zinazowakilisha matumizi halisi ya watumiaji. Kila scenario ina input, output inayotarajiwa, na uamuzi wa pass au fail.

What it is not

Si test ya moja kwa moja ya units pekee. Si orodha ya inputs zisizo na muktadha.

2-minute theory

Scenarios halisi huonyesha udhaifu mapema.
Edge cases hupima mipaka ya agent.
Matokeo yaliyoandikwa huongeza uthabiti wa QA.

Key terms

Scenario: Hali ya matumizi inayojumuisha muktadha na lengo.
Edge case: Hali nadra inayoweza kuvunja mfumo.

Where

Applies in

Agent testing
QA kabla ya uzinduzi

Does not apply in

Demos bila watumiaji halisi

Touchpoints

Scenario list
Test logs
Bug backlog

When

Use it when

Unakaribia uzinduzi
Unabadilisha workflow ya agent

Frequency

Kila mabadiliko makubwa

Late signals

Watumiaji wanaripoti errors zisizoeleweka
Agent inashindwa kwenye edge cases

Why it matters

Practical benefits

Failures chache baada ya uzinduzi
Uelewa bora wa mipaka ya agent
Uaminifu wa mtumiaji

Risks of ignoring

Bugs kwenye mazingira ya production
Support tickets nyingi

Expectations

Improves: uthabiti na uaminifu
Does not guarantee: sifuri bugs

How

Step-by-step method

Orodhesha matumizi 5 ya kawaida.
Ongeza edge case moja muhimu.
Andika input na output inayotarajiwa.
Endesha tests na rekodi matokeo.
Tengeneza tiketi kwa errors muhimu.

Do and don't

Tumia data halisi ya mfano
Rekodi pass au fail

Don't

Kuandika scenarios zisizo na muktadha
Kuepuka edge cases

Common mistakes and fixes

Mistake: Scenarios ni za jumla sana. Fix: Ongeza muktadha wa mtumiaji.
Mistake: Hakuna edge cases. Fix: Ongeza angalau moja.

Done when

Scenarios 5 zimeandikwa.
Kila scenario ina output inayotarajiwa.
Matokeo ya tests yameandikwa.

Guided exercise (10 to 15 min)

Inputs

Orodha ya matumizi ya kawaida
Mfano wa mtumiaji

Steps

Andika scenarios 5.
Fafanua output inayotarajiwa kwa kila scenario.
Rekodi matokeo.

Output format

Field	Value
Scenario
Input
Expected output
Pass or fail

Pro tip: Chagua scenario moja yenye data chafu ili kupima mipaka.

Independent exercise (5 to 10 min)

Task

Ongeza scenario moja ya juu ya hatari na ipime.

Output

Scenario mpya na matokeo yake.

Self-check (yes/no)

Je, scenarios zina muktadha halisi?
Je, kuna edge case?
Je, matokeo yameandikwa?
Je, bugs muhimu zimeandikwa?

Baseline metric (recommended)

Score: Scenarios 5 kati ya 6 zimepita
Date: 2026-02-07
Tool used: Notes app

Bibliography (sources used)

Testing AI Systems. Microsoft. 2024-01-01. Read: https://learn.microsoft.com/
Scenario Based Testing. Atlassian. 2024-01-01. Read: https://www.atlassian.com/continuous-delivery/software-testing/scenario-testing