Example · Test Analyst (manual / UAT)
Tools can now draft test cases. What keeps you is getting the business to test and signing for the result.
23 of your 40 hours still need you
- 15 stay yours
- 8 faster with AI
- 17 tools could take
A typical week at the middle pace. Your plan uses your own tasks.
Is manual testing a dying career?
By our task-by-task count, 23 of a typical 40-hour week as a test analyst still need a person. Tools could already take about 17 hours. Drafting and running cases is what models already do well, and it is 42% of your week. It is also what you were hired to do.
Read more
That work covers cases from a story, the scripted run, the defect write-up and the Friday status report. If you have tried a drafting tool on a story, you have seen it produce thirty cases in a minute. After an hour of tidying, a day of your work is done.
The same trial shows what holds. A few of those cases usually test a rule that no longer exists, and nobody but a person who knows the product notices. The job now has four parts. You know whether a case proves anything, you tell a defect from a business rule working as designed, you get a business user to sit down and look, and you put your name on the report that says the release was tested. Most test analysts already do all four, informally. The title says you write and run cases. Two tasks people often dislike are the completion report and the environment chase. This report helps you take ownership of both. Before the AI part, note that the biggest risk to this seat is older than any model. Manual testing has been moved offshore for twenty years, and many test teams are already half vendor. We score that separately, and it comes first.
The hard truth
Many testers now automate as well as design tests.
In a 2023 developer survey, most people involved in testing said they write automated tests. A role that only runs scripts by hand is the narrowest version of this job, and the hardest to defend.
Sources (2)
- 53% of people involved in testing said the same people design and execute tests at their company. JetBrains State of Developer Ecosystem 2023, 2023
- 58% of those involved in testing write automated tests. JetBrains State of Developer Ecosystem 2023, 2023
What stays yours
"Is this a defect, or is that how the rating engine works?"
This takes context. The answer lives in the product rules, the odd cases a particular record carries, and years of watching what the program broke. A model can read the requirement, but it cannot know that the requirement was wrong and everyone agreed to leave it.
Knowing which records to test with
This takes context, and errors are costly to catch. A wrong premium or payment looks exactly like a right one until a letter goes out. Your test data encodes which records have the odd endorsements and mid-term changes. Nobody wrote that list down.
Getting the business users into the UAT room
This takes presence and trust. Business users test when someone they trust asks them, walks them through it and writes down what they say. A model can send the invite, but it cannot be the reason a claims handler gives up a Thursday afternoon.
Signing the test completion report
You carry this accountability for the employer. The report is the evidence that a release was tested before go-live, and internal audit tests that control. Someone named has to sign. Often, informally, that is the test analyst, beneath the test lead.
Spotting the generated cases that test a rule you removed
This is checking work. A tidy case that proves nothing is the costliest kind of error, because it passes. The person who can read thirty cases and say "two of these are dead" is the constraint on every case-generation tool.
What's next for a manual tester?
Plan A is your best move inside the work you do now. Most sessions in the 13-week plan build toward it.
- A
Become the named UAT coordinator and acceptance author
Why it fits
Why you. You already get the business users into the room and you already draft the report someone else signs. Business users return your calls, and nobody on the vendor side has that standing. The gap between what you do and what your title says is the smallest of the three moves.
The gap. Usually no written acceptance statement exists. The completion report is a template with a traffic light. No business owner signs beside the test team. The UAT schedule lives in a shared mailbox.
The trade-off. Your name is on record when a release goes wrong. The pressure stays the same, but you gain authority. You will write fewer cases for a quarter, and your test lead has a cases-per-analyst number somewhere. The seat is also tied to programs. It moves with you to the next program, but not to a business-as-usual (BAU) support team.
Other moves that fit
These are also good bets in this field. Read them if Plan A doesn't suit you.
Own the quality of machine-written test cases
Why it fits
Why you. Finding the generated cases that test a dead rule is a skill with no name yet, and assistant drafting is about to become the default for the vendor's testers as well as the onshore team.
The gap. There is no written standard for what a case must prove, no record of how many generated cases get rejected and why, and no shared list of retired rules for a tool to check against.
The trade-off. You read more than you write, and some weeks it feels like proofreading for a machine. A vendor can offer the same review as a service line, and an executive can still buy it. The plan makes that deal worse than keeping you, though it does not rule it out.
Back-up plan: domain tester, with a path to business analyst
Why it fits
Why you. Knowing which records break the core engine is worth money at every company in your domain and at every consultancy selling into it. Domain testing is the corner of this occupation where employers hire for context more than scripts, and it is one step from business analysis.
The gap. You have no public write-up and no clear view of what the market pays. Add the ISTQB acceptance testing (CT-AcT) or test analyst (CTAL-TA) certificate only if postings ask for it.
The trade-off. Leaving resets the asset this report says protects you. Your first year anywhere new is your most exposed, because context starts at zero and cases written is all anyone can measure.
What's a plan B for test analysts and UAT testers?
Plan B is a job AI is creating that these skills already fit. One session a week in the plan trains you for it, so it is ready if your job shrinks or you want a change.
BPlan B, picked for this role
Agent simulation scenario designer
Turning real user steps into clear UAT scripts is what agent simulation needs, with customer conversations as the test cases.
Other options
AI evaluation specialist
Signing off on whether the business can accept a release is the same call evaluators make about AI outputs, case by case.
Expert AI trainer (contract)
Grading outputs against expected results is how platforms test trainers, and your domain knowledge sets your pay tier as bridge income.
In your own plan you can switch to any of the 47 roles.
The first two weeks
The first 2 of 13 weeks. Most weeks have one 45-minute, one 60-minute and one 90-minute session, plus one 40-minute Plan B session. In weeks 5 and 9 the third session is the release itself, and week 12 has one slot off. By week 13 you'll have two one-page acceptance statements for real releases, a review checklist for AI-written test cases, and a memo asking to be named UAT coordinator.
- Week 1: Measure your week
What you do
- ALog your time for one week against the task list in this report.
- AFor your last three releases, count cases written, cases run, defects found in UAT, the user acceptance testing stage, and defects found in production within 30 days.
- AOpen a batch of AI-drafted test cases, or draft some from one story. Sort them into four groups: proves something, duplicate, tests a removed rule, or wrong.
- BTake the discernment sections of Anthropic's AI Fluency: Framework & Foundations. Ask Claude for test cases on a made-up login story and mark the weak ones.
- Week 2: Learn the tool and audit rules
What you do
- AAsk your test lead in writing which test environments any AI tool or agent may run in. Ask who made that call.
- AFind who owns the change-management control, the audit rule your completion report feeds. Ask them for last year's written control description.
- ACompare the control text with what actually happens on sign-off day.
- BTake Anthropic's AI Capabilities and Limitations course.
Three courses from the plan
Checked 2026-10-11.
- ISTQB Certified Tester Foundation Level (CTFL) v4.0 (syllabus 4.0.1)ISTQB via ASTQB / AT*SQA (US exam) · $229
Only if a posting you want asks for it. Many bank and insurer HR screens do. It teaches vocabulary more than judgment.
- ISTQB Advanced Level Test Analyst (CTAL-TA) v4.0ISTQB via AT*SQA · $249
It is the one ISTQB exam that describes your durable work. Even so, take it in months 4 to 6, only if postings ask, and only v4.0.
- ISTQB Acceptance Testing (CT-AcT)ISTQB via AT*SQA · $199
Read the free sample exam first. The syllabus predates every tool in this report. Buy only if you want the vocabulary for Plan A on a résumé.
Risks beyond AI: highOffshore and program risk
Offshoring outranks AI here. Manual test execution has been offshored on cost since before today's tools existed, and enterprise test teams are often mostly vendor staff already. Many test analysts are also employed by a program. When the platform goes live, the test team shrinks, with or without an agent. We score this separately because blaming AI for it would get the remedy wrong.
Where this could be wrong
- UAT facilitation may be more exposed than we scored. Our capability score assumes a business user needs a trusted person in the room. If your business users are comfortable with a guided walkthrough tool that records what they do, that task becomes more exposed, and the UAT coordinator move leans harder on the acceptance statement than on being in the room.
- We assumed the completion report is yours informally. If a release manager already signs and you only prepare, the UAT coordinator move is "get your name on it," not "own it," and that task gives you less protection.
- The offshore ratio may already be the ceiling. If your employer has a policy floor on onshore testers for data-handling reasons, the non-AI volatility is lower than "high" and the AI risk moves back to first place.
- Environment chasing may be going away on its own. If the program moves to a cloud platform with self-service environments, a chunk of your lasting work disappears and the split of your week looks worse, though your position does not.
Also matched: test analyst, uat analyst, uat tester. Research checked 2026-10-11. Page updated .