Describe your idea. Get curated feedback. See what works and what doesn't.
Works for startups, books, campaigns, products, anything you're creating.
Takes 2 minutes to submit • Results in 10 minutes*
No complicated dashboards. Just clear, useful feedback.
People discuss your idea. Some love it. Some find problems. Just like real life.
See at a glance what's working and what needs improvement. No guesswork.
Not vague advice. Actual questions you can answer to make your idea better.
Anyone creating something people will have opinions about
Test startup ideas before building. Find problems early.
Test book concepts, character arcs, plot ideas.
Test campaigns, messaging, positioning before launch.
Test video ideas, courses, content before creating.
Test product concepts, features, user experience.
Test a shower thought or an invention.
Does this ad resonate with people, is the message clear.
Test anything people might have opinions about.
We tested the same concepts with ChatGPT, Gemini, and Friction Genie.
Here's what happened:
Pitch: Most meal kit services assume you're cooking for 2-4 people, which leaves solo households wasting food or eating the same thing for days. We're building a meal kit specifically for people living alone - single-serving portions, recipes that don't require advance planning, and pricing that actually makes sense when you're just feeding yourself. The market is huge (36% of US households are single-person) but completely underserved. We'd charge £6-8 per meal, delivered twice a week, with zero commitment. The key differentiator is "impulse cooking" - you can order tonight's dinner at 2pm and get it by 6pm in major cities.
We submitted the same startup concept to all three tools. Here's how they responded:
ChatGPT gave us general business questions:
How will you make the unit economics work with single-serving portions and fast same-day delivery at £6–8 per meal?
Generic business 101What makes this different from simply ordering takeaway or ready meals, especially if customers want dinner within a few hours?
Surface-levelHow will you manage inventory and reduce food waste while supporting "impulse" same-day orders?
Generic operationsWhy have existing meal kit companies not targeted solo households yet, is there a structural reason that makes this market difficult?
Lazy questionThese are generic "have you thought about costs?" questions. Could apply to any food delivery business. Nothing specific about your approach or market.
ChatGPT 5.3 on free tier
Gemini gave more specific pushback on logistics:
How do you maintain a £6-8 price point while absorbing the high last-mile delivery costs associated with a 4-hour "impulse" turnaround?
More specificSince packaging and shipping costs for a single-serving box are nearly identical to a family-sized box, how do you prevent your profit margins from being eroded by the lower order value?
Smart economicsHow will you manage inventory and food waste (shrinkage) in local hubs when you are moving away from the predictable, week-in-advance subscription model to on-demand ordering?
Operational detailWhat is the specific "hook" that prevents a solo diner from simply walking to a local supermarket for a high-end ready meal that is cheaper and available instantly?
Competition angleThese show understanding of logistics challenges, but feel like business school case study questions rather than real market concerns.
Gemini 3 Flash on free tier
Friction Genie generated a full forum with multiple personas debating the idea. Here are the questions that emerged:
How do you achieve profitable unit economics with £6-8 orders when Gorillas and similar services failed with much higher order values, and what specific delivery cost structure makes this work?
References real failure casesWhat percentage of your target market (solo households) actually wants to cook on impulse at 2pm versus just ordering takeaway, and how did you validate this behavior exists at scale?
Questions core assumptionCan you achieve the 2-4 hour delivery window required for fresh ingredients at density without building expensive micro-fulfilment centres in every neighborhood, or would partnering with existing delivery networks compromise freshness?
Forces real trade-offHow do you solve the inventory and waste problem when you need variety to appeal to impulse buyers but lower order frequency means harder utilization of pre-portioned ingredients?
Specific operational tensionWhat is your customer retention strategy given that solo diners have more irregular schedules than families, and how does that affect LTV compared to traditional meal kit subscribers?
Cohort-specific economicsReferences specific failed companies (Gorillas), questions core behavioral assumptions, forces real trade-offs (build vs partner), and identifies cohort-specific economic challenges (irregular schedules = lower LTV).
Friction Genie
Pitch: It's 2035. Maya is the last human employee at a fully automated logistics company - not because she's essential, but because regulatory law requires at least one human on payroll. Her job is to sit in an office, approve AI decisions she doesn't understand, and take legal responsibility when things go wrong. When a catastrophic shipping error threatens to bankrupt the company, she's the only one who can be prosecuted. The story follows her attempt to understand what the AI actually does, why it made the mistake, and whether she's complicit in a system she never controlled. Part corporate thriller, part meditation on accountability in an automated world. Told from Maya's first-person POV with flashbacks to how she ended up in this job.
We pitched the same story premise to all three tools:
ChatGPT asked basic story structure questions:
What specific event or "shipping catastrophe" triggers the plot, and why does the AI make that particular mistake?
Asking for the storyWhy does Maya stay in a job where she carries legal liability for systems she doesn't control?
Surface character questionWho or what opposes Maya as she investigates the AI, the company, the legal system, the AI itself, or her own uncertainty?
Generic plot questionWhat does Maya ultimately discover about responsibility: is someone actually accountable, or is the system designed so that no one truly is?
Asking for the endingThese are writing workshop 101 questions. "Tell me about your character and plot." No engagement with whether the premise actually works.
ChatGPT 5.3 on free tier
Gemini showed more genre awareness:
If Maya doesn't understand the AI's logic, what specific "inciting incident" or technical clue allows her to begin investigating a system that is intentionally opaque?
Better craft questionWhat is the emotional or financial "tether" that keeps Maya in such a high-risk, soul-crushing position instead of resigning before the inevitable catastrophe occurs?
Character motivationHow do you raise the stakes of the corporate thriller without making the AI a sentient "villain," keeping the conflict focused on systemic accountability and legal loopholes?
Genre awarenessIn a world of total automation, who are the secondary characters that provide Maya with friction or information, given that her role is defined by isolation?
WorldbuildingShows understanding of craft and genre, but feels like a writing instructor's feedback rather than a reader's real concerns.
Gemini 3 Flash on free tier
Friction Genie generated discussion that questioned the premise itself:
What specific regulatory framework in 2035 would require a single employee to be criminally liable for AI decisions rather than allowing companies to distribute liability across shell companies or purchase comprehensive insurance?
Worldbuilding plot holeIf Maya is approving thousands of decisions daily without understanding them, what exactly is she reviewing on her interface, raw data, confidence scores, summaries, and how does meaningful human oversight work at that volume?
Concrete storytelling detailBeyond discovering how the AI system works, what concrete actions can Maya actually take with that knowledge that would meaningfully change her legal situation or the story's trajectory?
Does the plot work?How does the story avoid becoming either 'pure misery porn' where Maya is completely powerless, or implausible wish fulfillment where individual investigation somehow defeats systemic problems?
Tonal tightropeWhat makes Maya's investigation different from simply uncovering that the system is working as designed, if the AI optimized for unstated metrics, who bears actual responsibility and why would that shift liability from her?
Core premise questionThese questions challenge the fundamental premise: why would 2035 law work this way? What can Maya actually DO? If the AI just did its job, there's no story. Forces you to solve worldbuilding logic or realize the premise doesn't work.
Friction Genie
Pitch: Instead of hiring influencers to promote our new skincare line, we're paying them to tell their followers what NOT to buy from us. The concept: "We make 12 products. You probably only need 3 of them. Here's how to figure out which ones." Influencers get paid per video where they genuinely tell people to skip products that won't help their specific skin type. The tagline is "Buy Less, Choose Better." We're betting that the honesty builds trust and cuts through the noise of every brand telling you to buy everything. Target demo is Gen Z women who are exhausted by 10-step routines and overconsumption. Launch on TikTok and Instagram with a quiz tool that narrows down which 3 products you actually need.
We submitted the same marketing campaign to all three tools. Here's how they responded:
ChatGPT asked basic campaign questions:
How will you ensure influencers' "don't buy this" recommendations feel genuinely honest rather than like a scripted marketing tactic?
Vague concernIf the campaign encourages people to buy fewer products, how do the economics work compared to traditional skincare marketing that pushes larger baskets?
Surface-levelWhat prevents competitors from copying the same "buy less" positioning once it proves effective?
GenericHow will the quiz determine which three products someone actually needs, and how will you build trust that the recommendations are unbiased?
Fair but basicGeneric "have you thought about authenticity and economics?" questions. First-pass thinking that doesn't dig into real behavioral problems.
ChatGPT 5.3 on free tier
Gemini identified specific strategic risks:
How do you prevent influencers from "reverse-selling", deliberately bashing a cheap product to nudge followers toward a more expensive 3-item bundle, which would destroy the authentic "honesty" brand?
Smart specific concernWith a "Buy Less" business model, what is your strategy for customer lifetime value (LTV) and repeat purchases once a user has already bought the only 3 products they supposedly need?
Good LTV questionHow do you reconcile the high cost of influencer marketing with a strategy that intentionally suppresses sales volume and lowers the average order value (AOV)?
Economic tensionIf the quiz tool is the ultimate authority on what a customer needs, what is the specific added value of the influencer beyond simply driving traffic to that tool?
Role clarity"Reverse-selling" and LTV concerns show understanding of the model, but still feels strategic/abstract rather than user-behavior focused.
Gemini 3 Flash on free tier
Friction Genie surfaced real behavioral contradictions:
How do you maintain influencer honesty when you're paying them, and what prevents them from either curating their advice or defaulting to recommending fewer products as the easier content strategy?
Lazy influencer arbitrageIf the core value is recommending 3 products instead of 12, how do you grow revenue per customer or scale the business without eventually pressuring customers to buy more or launching additional products that undermine the original promise?
Business model trapWhen customer skin needs change seasonally or with hormones, how does a quiz-based 3-product recommendation system adapt without either breaking the 'buy less' promise or creating analysis paralysis when recommendations become outdated?
Real user behavior breaks promiseHow would you resolve contradictory recommendations if different influencers genuinely disagree about which products to skip for the same skin type, and how would that affect the 'honest brand' positioning?
Brand consistency problemWhat's your measurable success metric—conversion rate, customer lifetime value, brand sentiment, or something else, and how do you distinguish this from standard marketing when influencers must legally disclose paid partnerships?
Does honesty work when disclosed?Identifies the business model trap (can't grow without breaking promise), real user behavior that breaks the concept (seasonal skin changes), and asks if disclosed paid partnerships undermine the honesty angle.
Friction Genie
Pitch: A calendar feature that analyzes your schedule and tells you when you're lying to yourself. It learns how long tasks actually take you (vs. how long you schedule them), flags when you've booked back-to-back meetings with no buffer, and warns you when you've scheduled "focus time" during hours you historically just scroll Twitter. At the end of each week, it shows you a comparison: "You planned 6 hours of deep work. You did 2. Here's when you actually had focus time." The idea is to surface the gap between aspirational scheduling and reality, so you can plan more honestly. Optional feature: "Accountability Mode" shares your reality check with a friend or colleague.
We submitted the same product feature to all three tools. Here's how they responded:
ChatGPT asked basic product questions:
How will the system accurately detect what someone is actually doing (deep work vs. scrolling) without feeling invasive or requiring heavy tracking?
Surface concernWhat makes this meaningfully different from existing calendar analytics or time-tracking tools that already show how time is spent?
Lazy comp questionHow do you present the "you're lying to yourself" insight in a way that motivates behavior change rather than making users feel judged and abandon the tool?
Fair UX concernWhy would users choose to enable "Accountability Mode," and what incentives make sharing their productivity gaps with someone else appealing rather than uncomfortable?
Basic questionBasic product thinking. "How will you detect it? What makes it different? Will people use it?" First-pass questions without deep engagement.
ChatGPT 5.3 on free tier
Gemini showed better product thinking:
How does the tool differentiate between a task taking longer due to a user's "lying" versus external factors like a meeting overrunning or an urgent, unscheduled fire drill?
Edge case thinkingWhat is the specific "retention hook" that prevents a user from uninstalling the app once the novelty wears off and the constant reminders of their productivity failures become demoralizing?
Good retention questionHow do you handle the significant privacy and technical hurdles of tracking "Twitter scrolling" or off-calendar activity across multiple devices and browsers to prove a user wasn't working?
Technical realityIn "Accountability Mode," how do you prevent the data from being weaponized by toxic managers or causing unnecessary friction in professional and personal relationships?
Misuse concernRetention hooks and weaponization concerns show product awareness, but still feels like PM review rather than user reality.
Gemini 3 Flash on free tier
Friction Genie challenged the fundamental premise:
How will you get consistent, accurate data about actual task completion when calendar APIs don't capture whether someone was actually in 'flow state' versus just occupying time?
Technical reality checkWhat prevents this from becoming another abandoned productivity tool that users find addictive for week one but ignore by week twelve, like RescueTime and Toggl before it?
Specific failed precedentsIf the core value is helping people estimate better, why not start with just a simple weekly email asking 'you planned X hours of deep work, how many did you actually get' instead of building complex tracking and AI pattern recognition?
Occam's Razor kills the productWho actually benefits from the social accountability features, and won't sharing productivity failures create awkward workplace dynamics or just invite manager surveillance rather than peer support?
Real organizational dynamicsIs the real problem actually lack of self-awareness about time patterns, or is it structural constraints (back-to-back meetings, interruption-heavy jobs) that no individual app can solve?
Might invalidate entire premiseTechnical reality (calendar APIs can't measure flow state), specific failed competitors (RescueTime/Toggl), Occam's Razor (why not just a weekly email?), and the killer question: is this solving the wrong problem?
Friction Genie
One is surface level. The other is a depe dive.
Multiple personas debating your idea. See who loves it, who's skeptical, and why.
Iterate from the feedback and see the conversation evolve.
Reception, skepticism, clarity, excitement, quantified so you know what needs work.
Overall and per comment so you get full visibility on the discussion.
Not generic advice. Actual questions extracted from the discussion to answer.
View the forum and see why the questions were raised.
Answer the questions, refine your idea, run it again. See your scores improve.
Compare results and see what worked and what doesn't.
Ask ChatGPT to critique your idea. Then try Friction Genie.
The difference will be obvious in 10 minutes*.
Seriously, it's this simple
A few sentences is fine. Tell us what you're making and who it's for.
We check your idea and if needed, expand on your idea.
In 10 minutes*, we generate a realistic discussion about your idea.
Hand crafted personas with an interest in your idea, not generic voices.
Read the discussion, see your scores, get specific questions to answer.
Answer the questions, refine your idea, run it again. Compare scores.
We wanted to make it free and we worked hard to try and get some free tier working but as a solo founder, I haven't got the capital to support 1000 free tries. hopefully in the future I can.
2 minutes to submit your idea. 10 minutes* to get results. That's it.
Yes. Everything is private unless you choose to share it. We do use an LLM API (Anthropic) so your idea is submitted to Anthropic anonymously.
Anything people might have opinions about. Startups, books, campaigns, products, content, services.
It's AI simulating realistic feedback. Not a replacement for real users, but way faster and cheaper for early testing.
Try it! You'll get generic questions. We generate full discussions with specific, actionable feedback. See the difference
Sometimes it will be. Use it to spot patterns and blind spots, not as absolute truth.
Nope. If you can write an email, you can use this.
The feedback isn't personal, use the feedback to refine your idea.
Friction genie will highlight some nuances buts it's not a replacement for professional advice.
Not on purpose, some people may find the feedback a little blunt, if you are of a sensitive disposition, maybe don't use Friction Genie.
thats an opinion and that's fine, if you don't want to use it, then don't.
It takes 10 minutes*, and might save you months of wasted effort.
Test Your Idea TodayResults in 10 minutes* • Save time and money