CHAPTERS
Veriff’s core product: verifying a person against a government ID in one flow
Kaarel Kotkas explains Veriff’s purpose and the basic user experience: capturing an ID and a selfie (often as video) to confirm the person is real and present. He also hints at a broader ambition: making digital identity portable and trusted across borders and platforms.
- •Matches passports/driver’s licenses/ID cards to a live selfie within a single flow
- •Uses device camera with on-screen instructions to confirm real-time presence
- •Adapts by country: sometimes leverages government registries instead of requiring an ID
- •Positions Veriff as enabling trust for online services, not just a checkbox feature
Origin story: from hacking PayPal as a teen to rebuilding Wise’s verification
Kaarel traces the founding insight to his early experience bypassing age restrictions online and later stress-testing real KYC flows with better tools (Photoshop/Illustrator). A consulting-style request from TransferWise (Wise) in 2015 turned into the conviction that identity verification needed a ground-up rebuild.
- •Early lesson: systems were easy to game (changing DOB, reusing the same ID)
- •2015 trigger: Wise asked him to improve their identity verification flow
- •He tested how easy it was to impersonate someone online and found the problem unsolved
- •Conclusion: verification was inaccurate, didn’t scale, hurt conversion, and fraud was cheap
Identity as infrastructure, not compliance SaaS
The conversation reframes online identity verification as foundational infrastructure akin to Estonia’s digital identity system. Kaarel argues that compliance-driven approaches are insufficient for preventing fraud, especially as real-time payments remove the window for after-the-fact transaction monitoring.
- •Most providers built for compliance requirements, not for stopping fraud
- •Regulatory checklists can coexist with high fraud; accuracy needs deeper signals
- •Estonia example: digital identity underpins services far beyond financial services
- •Shift to real-time payments makes identity accuracy more critical upstream
Why video and “1,000+ data points” beat photo-based KYC
Kaarel explains why early KYC patterns (e.g., “three pictures”) were fundamentally weak and easy to spoof. Veriff’s approach emphasizes richer capture (video end-to-end) and many objective signals to enable scalable, consistent decisions.
- •Static photos provide weak assurance and are easy to manipulate
- •Video capture increases information and reduces spoofing opportunities
- •Large feature sets (“1,000+ data points”) support objective decisions at scale
- •Goal: higher accuracy without destroying user conversion
First traction: traditional banks, then Uber’s driver onboarding
Despite fintechs initially viewing Veriff as overbuilt, traditional banks became early adopters because they needed online verification to match or exceed in-person checks. Uber in Estonia followed when local teams needed strong remote onboarding without office visits.
- •Early buyers were Nordic/traditional banks needing bank-grade assurance
- •Fintech feedback: Veriff seemed like “nuclear-powered submarine” for simple needs
- •Uber Estonia used Veriff to enable remote driver onboarding and training
- •Local autonomy in Uber city teams helped land the customer
Bootstrapped constraints and early profitability pressure
Kaarel describes building Veriff as his first and only job, fueled by a small initial investment that forced rapid profitability. The company grew closely alongside customer needs, reinforcing the product’s infrastructure-level ambition.
- •Started full-time right after secondary school; Veriff became his life’s work
- •Initial ~€60k funding created a 6–7 month path-to-profitability constraint
- •Customer-driven iteration shaped the product and conviction early
- •Operated with a long-term infrastructure mindset from the start
YC rejection, cap table crisis, and the two-week turnaround
YC initially passed due to two issues: Kaarel was a solo founder and had an extremely unfavorable cap table after raising early money. He recounts learning what a cap table was, then executing a fast rescue—adding a co-founder and restructuring ownership, including taking a personal loan to buy back equity.
- •YC concerns: solo founder + “messed up” cap table
- •He had sold ~95% for €60k (driven by early bank funding needs)
- •Recruited an early employee (Janer) as a later-stage co-founder
- •Took a ~€1M personal loan to reset ownership; restructured toward ~75/25 founder-team vs. others
YC batch dynamics: customers wanted to buy Veriff, not just use it
During YC, many large companies expressed strong interest but pushed acquisition or in-house builds, sometimes withholding contracts that would make fundraising easier. Kaarel explains why remaining independent mattered: cross-industry collaboration against fraud works better when the infrastructure provider isn’t owned by a single platform.
- •Bigger companies tested the product but claimed Veriff wasn’t “big enough” yet
- •Some pressured acquisition or in-house alternatives instead of signing contracts
- •Independence seen as a strategic moat enabling industry-wide cooperation
- •Kaarel turned down life-changing offers after buying the company back
Post-YC scaling: funding, infrastructure, and global media performance
After YC, Veriff raised seed and then a sizeable Series A, returning to Estonia to scale execution. Kaarel highlights the technical focus on making video-based verification work globally, including in low-connectivity conditions, supported by strong engineering talent in Tallinn.
- •Raised ~$1.23M after YC, then closed a ~$7.7M Series A soon after
- •Focus shifted from proving value to scaling reliably and globally
- •Built infrastructure for media/video analysis that works with latency and low bandwidth
- •Leveraged Tallinn’s deep engineering talent (e.g., Skype-era expertise)
Crypto boom stress test: massive volumes and fraud at global scale
Crypto’s growth created both regulatory KYC needs and intense fraud pressure, making Veriff’s accuracy advantage valuable. A large airdrop campaign drove unprecedented global volumes, forcing rapid hiring and operational scaling while managing customer concentration risk.
- •“Fraud follows the money” — crypto amplified both volume and attack intensity
- •Airdrops/promotions triggered heavy signup fraud outside US/EU markets
- •Veriff scaled quickly, hiring ~200 people in ~6 months (mix of ops + engineering)
- •Described as “building a plane while flying it,” with early customer concentration challenges
COVID acceleration: remote life made identity verification essential
The early pandemic was uncertain and financially awkward (customers delayed payments), but demand surged as services moved online. Veriff became key to remote work, remote examinations, and notarized contracts—reinforcing identity as a baseline requirement for digital society.
- •Initial shock: uncertainty and delayed invoices, even as demand rose
- •Volume “skyrocketed” as more services moved online and remote-first
- •Use cases expanded: remote employment verification, exams, contracts/notarization
- •Validated the need for trust infrastructure when people can’t meet in person
From hypergrowth to organizational design: becoming the coach
Kaarel discusses the shift in founder role from builder to coach as the company scales. He emphasizes keeping the org effective and flat, avoiding unnecessary management layers, and continuously iterating on structure to match the company’s growth rate.
- •Founder transition: individual execution to coaching a team for leverage
- •Need for people’s growth to outpace company growth to sustain performance
- •Complexity from rapid scaling and financing required organizational redesign
- •Focus on flatter structures and iterative trial-and-error improvements
AI and deepfakes: why single-signal verification is breaking
Deepfakes are now cheap and widely accessible, turning identity fraud into an arms race. Kaarel argues that approaches relying on a single credential (voice, image, or basic biometrics) create false confidence; robust verification requires layered signals and device-level integrity checks.
- •Deepfakes aren’t new, but costs dropped dramatically and availability exploded
- •Single-factor methods (e.g., voice auth, photos-only KYC) are increasingly vulnerable
- •Layered signals: behavior in flow, accelerometer/motion, video dynamics, network/device checks
- •Detection includes verifying real camera capture vs. injected/synthetic media
The digital passport / trust infrastructure vision: identity + trust + ongoing authentication
Kaarel outlines a broader system where companies need answers to three questions: identity (who), trustworthiness (are they trusted), and continuity (is the same trusted person still behind the account). Veriff’s ambition is portable trust—so good actors carry reputation across platforms while honest users face less friction.
- •Three core questions: identity, trust, and re-authentication/continuity
- •Trust is broader than government ID; it includes behavior and history across services
- •Vision: users carry trust globally to access services with less repeated friction
- •Goal: industry-wide collaboration to reduce fraud and “make honest people trusted”
