Step 1
Frame It
One-liner
Cluely is a real-time AI copilot that surfaces contextual answers during live conversations — invisible to the other side.
Before I look at segments or features, I want to anchor on the structural insight that makes Cluely interesting as a product:
Before
LeetCode, Pramp, Notion
Prep decays under pressure. Memory ≠ performance.
During
Nothing — until Cluely
The gap. High-stakes moments have zero real-time tool support.
After
Otter.ai, Fireflies, Gong
Post-mortems don't help you when you're blanking mid-answer.
Every competitor either helps you prepare or helps you review. Cluely is the only product that helps you perform — in the moment. That is the whitespace they claimed.
Step 2
The User
Cluely launched targeting one segment sharply and expanded. The three segments are distinct — different JTBD, different economics, different churn profiles.
Job to be Done
Pass technical interviews despite DSA gaps or performance anxiety
Core Pain
LeetCode prep doesn't hold under pressure. Every question feels like a blank.
Willingness to Pay
High — already spending on LeetCode Premium, prep courses
Churn Risk
High — they stop using once hired
Job to be Done
Surface objection responses, competitor comparisons, and pricing in real-time
Core Pain
Can't memorize every SKU, case study, and competitor differentiator. Gets caught flat-footed.
Willingness to Pay
High — if it closes one extra deal, the tool pays for itself in minutes
Churn Risk
Low — becomes part of daily workflow
Job to be Done
Appear prepared, surface context, capture key decisions in real-time
Core Pain
Meetings move faster than preparation. Key decisions happen when you haven't read the brief.
Willingness to Pay
Medium — value is less acute than interview or sales use cases
Churn Risk
Low — recurring weekly use case
The Tension
The launch wedge (job seekers) has the highest conversion and the worst LTV. The sales rep segment has lower conversion but 10x better retention. This is why the expansion strategy isn't optional — the interview use case builds brand and distribution, but it cannot sustain the business alone.
Step 3
Product Anatomy
Breaking down the product surface — what each piece does and, more importantly, the PM reasoning behind each decision.
A transparent overlay that doesn't show up in screen share tools like Zoom, Google Meet, or OBS. This is the core technical differentiator.
PM Read
Why invisible? Because visible makes it useless. The entire value proposition collapses if the interviewer can see it. This wasn't a feature decision — it was a product survival constraint.
Captures both sides of the conversation live. Transcribes and routes to the AI model with minimal latency, surfacing responses before the user has to stall.
PM Read
Why audio AND screen? Screen gives code context. Audio gives question context. Together they enable accurate, contextual responses — not just generic completions. Removing either degrades the output significantly.
Upload your resume, job description, and talking points before a session. The AI grounds all responses in your specific context rather than producing generic answers.
PM Read
This is the feature that separates Cluely from 'just open ChatGPT on your phone.' Context injection makes the output 10x more relevant and significantly harder to replicate with off-the-shelf AI.
Interview Mode (coding + behavioral), Sales Mode, Meeting Mode. Same AI engine underneath — different context templates and prompt tuning per mode.
PM Read
This is the expansion playbook in the product itself. Interview was the wedge. Sales and Meetings are the monetization path. Building modes signals where the company is going — and what segments they're betting on for retention.
Step 4
Business Model
Model
B2C Subscription
Individual users — monthly and annual plans
Wedge
Job Seekers
High intent, low CAC via organic controversy
Expansion
B2B Sales Teams
Higher ACV, better retention, procurement path
Revenue Driver
Seat-based pricing
Per-user as they move to team plans
The Unit Economics Problem
If the average job seeker finds a role in 3 months and pays ~$30/month, that is $90 LTV. Even with low CAC from organic virality, this is not a scalable business on its own. The job seeker segment builds the user base and funds early growth — it is not the end state.
A sales rep using Cluely to close deals will use it indefinitely. A $30/month subscription from an AE who attributes one closed deal per month to the tool is renewal-certain. That is the real business. The interview use case is the acquisition channel.
Step 5
Growth Loop
Cluely didn't grow despite the controversy — it grew because of it. The launch was a growth hack disguised as a brand statement.
Provocative Launch
"I cheated my way through 30 interviews" — a single headline that made every engineer feel something strong. Anger, envy, or recognition.
Media Amplification
Haters wrote think-pieces. Fans shared clips. Ethics boards issued statements. Every article drove organic reach to people who had blanked in an interview last month.
Organic Search Capture
Searches for "how to cheat coding interviews" and "AI interview tool" spiked. Cluely owned the top of that high-intent funnel with near-zero ad spend.
High-Intent Signups
People who find the product through controversy already have a real, acute pain. Conversion is high because the problem is visceral — not aspirational.
Viral Success Stories
Users share "I passed FAANG with Cluely" stories. Each success story is a new loop trigger — more media, more searches, more signups.
The Key Insight
Most founders try to avoid controversy. Roy Lee weaponized it. Every think-piece calling Cluely unethical was free advertising to someone who had blanked on a coding round under pressure. The positioning was implicitly: “if you're angry about this, you've never bombed a technical interview.” That reframe turns critics into amplifiers.
Step 6
Competitive Map
Mapped on the two axes that matter most for this category: when in the conversation workflow does the tool operate, and who is the primary buyer — individual or team. Click a dot to read more.
Real-time AI overlay during conversations — invisible to screen share detection.
Feature Comparison
| Capability | Cluely | Otter.ai | Gong | LeetCode |
|---|---|---|---|---|
| Real-time AI suggestions | ✓ | — | — | — |
| Invisible to screen share | ✓ | — | — | — |
| Live transcription | ✓ | ✓ | ✓ | — |
| Post-meeting analysis | — | ✓ | ✓ | — |
| Interview / prep context | ✓ | — | — | ✓ |
| Sales intelligence | ✓ | — | ✓ | — |
| Team-level analytics | — | Partial | ✓ | — |
Step 7
Moat & Risks
What protects them
First-mover brand
In a new category, the pioneer brand is the moat. 'Invisible AI copilot' maps to Cluely in most engineers' minds right now.
Context model depth
After millions of real interview and sales conversations, their prompt tuning and context extraction is battle-tested in a way competitors can't shortcut.
Latency optimization
Real-time requires near-instant response. Getting this right across different devices, network conditions, and conversation types takes serious engineering time.
Controversy flywheel
Organic viral CAC driven by strong brand reactions. Hard to replicate because it requires both a real product and a willingness to make people uncomfortable.
Risks — click to expand
Step 8
My Take
What they got right
They found genuine whitespace
Before and After are crowded. During was empty. That's not luck — identifying a structural gap in the workflow and betting on it before anyone else is sharp product thinking.
Controversy as a CAC strategy
Organic controversy is the highest-leverage growth mechanism if your product actually solves a real problem. They earned the right to be controversial because the pain they address is real and widely felt.
The pivot timing
Moving from 'cheat tool' to 'AI meeting assistant' language before the interview use case became their ceiling shows self-awareness. The window to do this cleanly is short — they seem to be moving at the right time.
What I'd do differently
Kill the 'cheat' framing faster
The viral moment was built on cheating. But enterprise doesn't buy from brands associated with misconduct. I'd have a hard brand cutover to 'ambient AI assistant' within 6 months of launch and let the founding story become lore, not positioning.
Build the retention loop in the product
Job seekers churn once hired. I'd build a 'career mode' that transitions users from 'use during interviews' to 'use during onboarding, performance reviews, and 1:1s' — extending the relationship past the job search.
Invest in the context layer as the moat
The invisibility tech is copyable. The context model — fine-tuned on millions of real interview and sales conversations — is not. I'd orient roadmap investment toward deepening context quality over adding new feature modes.
The Big Bet
Ambient AI becomes cognitively normalized — like GPS.
We don't say we're “cheating at navigation” when we use Google Maps. If ambient AI during conversations follows the same normalization curve, Cluely's entire risk profile changes. The brand stops being a liability and becomes the pioneer story. That bet depends more on culture than on product — and that makes it either the biggest risk or the biggest tailwind, depending on which way the decade goes.
0 gap
Workflow timing gap
Only real-time in-conversation AI at launch
0
User segments targeted
Interview → Sales → Meetings
0
Growth loop steps
From provocation to viral success stories