Modern hiring has developed a cute little habit: asking candidates to create brand-new work for free, then acting confused when people call it free work.
“It should only take two hours,” says the recruiter, attaching a prompt that requires market research, customer segmentation, a 90-day roadmap, a dashboard mockup, and apparently a small séance with their churn data.
The trick is not always to refuse. Sometimes refusing is right. Sometimes doing it is strategic. But there is a third option candidates forget because the process trains you to behave like a grateful intern at the gates of employment heaven:
Swap the assessment.
Not “I won’t prove myself.”
Not “Here is my weekend, please use it responsibly.”
More like: “I’m happy to demonstrate the skill. Here are cleaner ways to evaluate it without turning my candidate work product into unpaid consulting.”
That sentence should be on a mug, but HR would probably make you complete a case study to earn it.
The assignment swap is the missing middle
Asha, a senior lifecycle marketer, made it through a recruiter call, hiring manager interview, AI interview screen, and panel.
Then came the “small exercise.”
The company wanted her to:
- audit their onboarding emails,
- identify conversion leaks,
- rewrite three emails,
- propose experiments for the next quarter,
- present the work to the team,
- and share editable files “so everyone can review async.”
Translation: “Please perform a free lifecycle strategy sprint while we decide if you are worth health insurance.”
Asha had a good instinct: she didn’t want to torch the opportunity by sounding difficult. The hiring market has already taught candidates that boundaries are a luxury item. But she also knew the request was not a skills test anymore. It was a free consulting interview task wearing a Patagonia vest.
So she offered three alternatives.
- A redacted past work sample with metrics and decision logic.
- A 45-minute live working session using a fictional scenario.
- A paid work trial if they wanted company-specific recommendations.
That is the whole move.
You are not dodging evaluation. You are forcing the evaluation to become honest.
Compare the five ways to handle “one more sample”
Not every assignment is evil. Some are reasonable. Some are clumsy but salvageable. Some are just a company trying to outsource its backlog through finalists.
Use this matrix before you open the doc.
| Option | Best when | Candidate risk | What it proves | Boundary line |
|---|---|---|---|---|
| Do the take-home as requested | The scope is tiny, fictional, time-capped, and tied to clear work trial evaluation criteria | Medium | You can follow the prompt and deliver under constraints | “I’ll spend up to 90 minutes and submit a PDF summary.” |
| Cap and clarify the take-home | The prompt is vague but the company seems serious | Medium-low | You can prioritize and communicate scope | “Which two outputs matter most for the scorecard?” |
| Swap in past proof | You already have similar work, redacted and ready | Low | You have done the job before | “I can walk you through a comparable project and the decisions behind it.” |
| Propose a live working session | They need to see thinking, not harvest output | Low-medium | You can reason in real time | “Let’s use a fictional scenario or sanitized data.” |
| Request a paid work trial | They want real company analysis or usable deliverables | Low if paid, high if unpaid | You can solve their actual problem | “For company-specific recommendations, I treat that as paid project work.” |
| Decline | The ask is large, proprietary, unpaid, or arrives after too many rounds | Low for dignity, possible loss of role | You protect your time and IP | “I’m going to withdraw rather than complete unpaid strategic work.” |
The goal is not to pick the toughest option. The goal is to pick the option that matches the risk.
A fictional two-hour exercise with a rubric? Fine, maybe.
A real company data take-home asking you to diagnose their revenue leak? That needs a Data Room Boundary Packet, a paid scope, or a polite exit.
A “quick revision” after they already saw your first version? That is how unpaid revision loops are born. They start as feedback and end as product management with no badge.
Option 1: Do it, but only if it passes the clean test
A clean unpaid take-home assignment has four traits:
- Fictional or sanitized context — not their real backlog.
- Time cap — written, not vibes-based.
- Evaluation criteria — you know what good means before you start.
- Non-reusable output — they cannot quietly ship it, sell it, or feed it into Monday’s roadmap meeting.
If the task passes those four tests, doing it may be reasonable.
But still cap it.
Send this:
Happy to complete this. To make sure I’m focusing on the right signal, I’ll spend up to 90 minutes and submit a concise PDF covering assumptions, approach, tradeoffs, and final recommendation. Can you confirm the top 2-3 criteria the team will use to evaluate it?
Notice what this does.
It asks for the hidden interview scorecard without sounding like you’ve been radicalized by spreadsheets. It also makes your boundary boring, which is the highest form of professional power.
Option 2: Cap and clarify before the vague prompt eats you
Vague assignments are blank checks with homework font.
“Tell us how you’d improve activation.”
What does that mean? A teardown? A research plan? Copy? SQL? A board deck? A prophecy?
Do not guess silently. Guessing is how candidates turn a two-hour assignment into a nine-hour apology.
Ask:
I can take this a few directions: funnel diagnosis, experiment plan, messaging rewrite, or executive summary. Which output is closest to what the role will actually do in the first 30 days?
Then:
Is the team evaluating depth of analysis, prioritization, communication, or tactical execution? I want to make sure the work sample matches the scorecard.
This does two useful things.
First, it exposes whether they know what they are hiring for.
Second, it gives you permission to stop. If they say “just do what you think is best,” that is not empowerment. That is scorecard fog.
Option 3: Swap in past proof when the work already exists
Hiring teams often ask for new work because they are bad at reading old proof.
That is not your fault, but it is now your problem for exactly one email.
Offer a past proof package:
- redacted artifact,
- short context note,
- your role,
- constraints,
- decision tradeoffs,
- results,
- what you would do differently now.
Call it a proof walkthrough, not a portfolio dump.
Script:
Instead of creating new company-specific work, I can share a redacted example of a similar project and walk through the problem, constraints, decisions, stakeholder management, and results. That should give you a cleaner view of how I operate without creating new unpaid work product.
This is especially strong if the role already put you through an automated hiring screen or one-way video interview where your proof had to be compressed into bot-readable answers. If you need help turning messy experience into clear proof blocks before a live walkthrough, NoSweatKing is an AI interview copilot that decodes questions and helps you answer in your own voice.
The important part: your past work proves competence without handing them fresh strategy.
Option 4: Offer a live working session when they claim they need “thinking”
Some teams say they need a take-home because they want to see how you think.
Great. Then watch me think.
A live working session can be fair if it is structured correctly:
- 45 to 60 minutes,
- fictional scenario,
- no real confidential data,
- clear prompt,
- interviewer participates,
- output stays conversational,
- no editable files after.
A live working session is not automatically better. It can become theater. It can punish people who think quietly. It can turn into a culture fit interview with a whiteboard costume.
But compared with a weekend take-home, it limits theft.
Script:
If the goal is to evaluate my approach and tradeoffs, I’d be happy to do a 45-minute live working session on a fictional scenario. That may be more useful than a polished take-home because you’ll see how I clarify assumptions, prioritize, and communicate decisions in real time.
This makes the company choose: do they want your judgment, or do they want your deliverable?
Those are not the same thing.
Option 5: Ask for a paid work trial when the output is real
Here is the cleanest rule in the whole miserable circus:
If they can use it, they should pay for it.
If the prompt uses their real data, real customers, real roadmap, real funnel, real competitors, or real team conflict, you are not doing an assessment. You are doing work.
That does not mean you have to become hostile. You can be extremely calm while refusing to be farm equipment.
Script:
I’m happy to work on a company-specific problem. Because that would create usable recommendations for the business, I’d treat it as a paid work trial with a defined scope, timeline, and usage terms. If unpaid is required, I can instead use a fictional scenario or walk through a comparable past project.
This is not entitled. This is normal.
The weird part is a profitable company asking strangers to donate strategy in exchange for maybe getting ghosted by Thursday.
How to choose your move in under five minutes
Use this quick decision path.
If the assignment is fictional, short, and scored clearly
Do it or cap it.
Ask for criteria. Submit a PDF. Avoid editable files unless there is a reason.
If the assignment is vague but possibly legitimate
Clarify first.
Ask what output matters. Ask how it will be evaluated. Ask how much time they expect you to spend.
If they cannot answer, your risk just went up.
If the assignment uses real company information
Swap it.
Offer a live working session on a fictional version, a redacted past proof walkthrough, or a paid work trial.
If they ask for revisions after revisions
Stop the loop.
Say:
I’m happy to answer questions about the submission. Additional revisions would go beyond the agreed assessment scope, so I’d suggest either moving to the next interview step or defining a paid trial scope.
Coachability is not infinite free labor.
If this arrives after five rounds
Your leverage is not imaginary.
Ask for the process map:
Before I invest additional work, can you confirm the remaining steps, decision timeline, and whether this is the final evaluation before offer-stage discussions?
Endless interview rounds thrive when every step feels too small to challenge.
The red flags that mean “swap or walk”
You do not need a law degree to smell exploitation. Watch for these phrases:
- “This should be quick” with a giant prompt attached.
- “Use our current product/site/data.”
- “We’d love your fresh eyes.”
- “Just send the editable deck.”
- “The team may use this as a discussion artifact.”
- “We don’t have a rubric, we just want to see how you think.”
- “Can you add one more version?”
- “We’re still calibrating the role.”
“Still calibrating” is recruiter-speak for “the job is a moving target and you may be asked to audition for several versions of it.”
Proceed accordingly.
What to send when you want to sound firm but not radioactive
Here is the complete email template.
Hi [Name],
Thanks for sending this over. I’m happy to demonstrate how I approach this type of work.
Since the prompt appears to involve company-specific analysis and potentially usable recommendations, I’d like to suggest one of three assessment formats:
- I can walk through a redacted past project with similar scope, including decisions, tradeoffs, stakeholder management, and results.
- We can do a 45-minute live working session using a fictional or sanitized scenario.
- If the team wants recommendations on the actual business problem, I’m open to a paid work trial with defined scope, timeline, and usage.
My goal is to give the team strong signal while keeping the assessment fair and bounded. Let me know which option is most useful.
Best, [You]
This is not a manifesto. It is a menu.
Menus are powerful because they make you look collaborative while refusing to become unpaid infrastructure.
The next step: build your swap assets before you need them
Do not wait until the recruiter sends the cursed Google Doc.
Prepare three assets now:
- A redacted proof walkthrough — one past project, six slides max, focused on decisions and results.
- A boundary script — your version of the email above, in your voice.
- A paid trial line — one sentence that names your requirement without apology.
Your proof walkthrough should include:
- the business problem,
- constraints,
- your role,
- options considered,
- tradeoffs,
- final recommendation,
- outcome,
- what you learned.
That structure works for behavioral interview answers, portfolio presentations, stakeholder management interviews, and take-home alternatives because it translates your work into evaluation signal.
The hiring system will keep asking for more because “more signal” sounds nicer than “we have decision debt and no one wants to own the hire.”
You do not have to feed the machine unlimited proof.
You can prove skill without donating strategy.
You can be cooperative without being extractable.
You can make the process fairer by forcing the work sample to confess what it really is.
