Silicon Valley Girl$18B AI CEO: How to Build a Million-Dollar Business in the Age of AI
CHAPTERS
From agency pain point to a browser whiteboard (Miro’s true origin story)
Andrey explains that Miro wasn’t born from a grand ambition to build a unicorn, but from a practical collaboration problem he faced running a creative agency. The early goal was simple: reach break-even quickly while solving a real pain for remote work with clients.
- •Started from a personal workflow problem: collaborating with remote customers
- •Initial product concept: “bring a whiteboard into a browser”
- •Early focus was survival and break-even, not massive scale
- •Rebuilt early versions after realizing initial signals showed “it’s not working”
- •Market thinking still mattered: choose a problem with a sufficiently large addressable market
When it became obvious it could be big: key growth inflection points (2015 → pandemic)
The conversation maps Miro’s step-changes in momentum: technical shifts, enterprise readiness, and the pandemic-driven remote work wave. Each phase revealed a new ceiling for how large the business could become.
- •2015 shift from Flash to HTML clarified a path to $1–5M revenue
- •2018–2019: discovering enterprise sales unlocked a path to ~$100M revenue
- •Pandemic era validated broad scaling for remote collaboration tools
- •User base jumped from ~5M to ~50M in ~18 months
- •Post-pandemic growth flattened but continued to ~100M users over the next ~36 months
What Miro evolved into: the “AI innovation workspace” and end-to-end delivery
Andrey reframes Miro’s positioning beyond visual collaboration into an end-to-end workflow where teams move from idea to delivery. AI becomes a collaborator that helps teams progress through stages faster, making the canvas a powerful modality for AI workflows.
- •Expansion from brainstorming tool to end-to-end innovation/delivery workspace
- •“Canvas as modality” for AI: collaborate with AI inside team workflows
- •AI shifts work from manual steps to AI-assisted progression between steps
- •Focus on compressing time from ideation to outcomes
- •Product evolution is framed as a new horizon, not a small feature add-on
The original growth flywheel: virality, delight, and search before sales layers in
Andrey breaks down how Miro acquired users early: obsessing over user experience and incentivizing collaboration invitations. Only after organic channels were strong did the company layer on more intentional marketing and enterprise sales.
- •Primary optimization: product experience and frictionless UX
- •Designed virality: users invited others to collaborate inside boards
- •Word of mouth activated by “wow, this exists” delight
- •SEO became a second major growth channel alongside virality
- •Intentional marketing and sales came later, after organic flywheel was proven
Building in 2025: PMF still wins, but brand and trust matter more than ever
With AI lowering the cost of building, Andrey argues fundamentals still rule: product-market fit and real problem-solving. However, in a crowded market of “nice-looking” tools, brand, trust, and emotional attachment become critical differentiators.
- •AI makes building faster/cheaper, but quality still requires heavy investment
- •If the product doesn’t solve a real problem, it won’t grow (fundamentals unchanged)
- •Brand becomes a stronger moat in an oversupplied product landscape
- •Trust becomes a core purchasing and usage driver, especially for AI tools
- •He emphasizes “lovemark” energy—excitement and emotional resonance with a brand
How to find PMF now: customer discovery + prototypes (especially for AI-first UX)
Andrey outlines a practical PMF process: validate the problem and market, then use open-ended customer conversations to test hypotheses. For AI-first productivity tools, prototyping is essential because users often can’t articulate what they need until they see it.
- •Start with: is the problem real, big, and tied to a large enough market?
- •Use open-ended interviews to prove/disprove hypotheses
- •Prototype early—users may not be able to “request” novel AI experiences
- •Put prototypes in front of customers and observe behavior/feedback
- •Signal can emerge from ~7–20 deep interviews, depending on depth and clarity
Why you need to fail: managing a portfolio of bets and iterating the solution
Miro embraces a deliberate failure rate to ensure the company is pushing boundaries. Andrey explains how to distinguish between a wrong problem and an imperfect solution, and why consistent iteration is the real engine of durable PMF.
- •Target success rate ~50–70%, leaving ~30% for failures
- •Applies across product, growth, and even acquisitions (portfolio mindset)
- •Balance “safe bets” with moonshots—some should fail by design
- •Key diagnostic: wrong problem vs. wrong/unfinished solution
- •Relentless iteration is normal (e.g., onboarding redesigned many times over years)
Trust as the new currency in AI products (and why users hesitate)
Marina emphasizes that users now feed AI apps highly sensitive data, making privacy and trust central to adoption. This section underscores that trust isn’t abstract—it directly shapes consumer behavior and willingness to try new tools.
- •Users share photos, docs, voice, and screenshots into AI apps—raising stakes
- •Consumers often don’t understand how their data is handled
- •Privacy concern affects product adoption and retention
- •Trust becomes a differentiator when switching costs are low and tools are abundant
- •Sets context for compliance/consent management as part of trustworthy product building
Name vs brand vs lovemark: rebranding RealtimeBoard to “Miro” to stand out
Andrey explains his framework for company naming and why Miro’s rebrand aimed beyond recognition toward emotional resonance. The “lovemark” concept ties the product to inspiration and creativity rather than generic “software.”
- •Three tiers: descriptive name → brand → lovemark
- •RealtimeBoard felt literal and undifferentiated; Miro chosen to stand out
- •Lovemark = an emotional feeling when you hear the name
- •Inspired by artist Joan Miró and the idea of an inspiring canvas
- •Positioning goal: not “another tool,” but an inspirational part of daily work
Strategy reset under AI: from 3-year “paint the picture” to 6–12 month planning cycles
Miro previously used a 3-year narrative planning practice (“paint the picture”), but AI’s pace makes long-range prediction unreliable. Andrey argues companies must anchor on mission while planning in shorter increments and constantly reassessing where they can win.
- •Adopted Atlassian-style “paint the picture” (multi-page future narrative)
- •Their 2022 plan executed well—but AI wasn’t predicted and changed assumptions
- •He believes predicting beyond ~12 months is increasingly unrealistic
- •Now commits to ~6-month plans; holds a looser 12–18 month (closer to 12) view
- •Key filter: focus on mission + choose battles where you have “permission to win”
Why building is more exciting now: LLMs expand the solution space and interfaces
Despite increased competition and ambiguity, Andrey is more optimistic because AI enables many new ways to solve problems. He describes a broad reinvention of interfaces and “surfaces,” making this era feel like a builder’s playground.
- •LLMs redefine the solution space: more ways to solve the same job-to-be-done
- •Expect reinvention of interfaces, workflows, and software “surfaces”
- •CEO lens: build sustainably; builder lens: explore new building blocks
- •Opportunity is vast but requires navigating a rapidly changing ecosystem
- •He frames the moment as a “candy shop” for builders
AI Canvas demo: multimodal workflows, image generation, and conversational editing in Miro
Andrey walks through Miro’s AI Canvas, demonstrating how teams can wire references into a workflow that generates outputs like images and structured artifacts. The demo highlights model selection, iterative prompting, and a “sidekick” for conversational edits—hinting at a multimodal future.
- •Switches into AI Canvas as a distinct mode with a new toolbar/workflow
- •Connects multiple references (style/location) into an image placeholder
- •Adds a prompt and selects a model (e.g., Stable Diffusion) to generate results
- •Iterates via follow-up steps (variants, refinements) and conversational “sidekick”
- •AI Canvas outputs can feed into other modalities (tables, Kanban boards, etc.)
Most important founder qualities now: curiosity, critical thinking, and resilience in chaos
Andrey argues that constant change will only accelerate, making mindset a competitive advantage. Founders must stay curious, think critically about where they can win, and build resilience to handle ongoing ambiguity and intensity.
- •Continuous curiosity is essential as tools, models, and markets shift
- •Critical thinking prevents “building blindly” without understanding the landscape
- •Resilience is required as daily volatility increases for founders
- •The environment will get more overwhelming as AI expands into physical/robotic experiences
- •Founders who enjoy ambiguity and complexity are best positioned to thrive
Where opportunities are: multiplayer AI, vertical AI, and fast consolidation
Andrey outlines where he sees the next wave: vertical AI companies reinventing entire end-to-end workflows. He cites legal, coding, and marketing as standout areas and predicts rapid consolidation—rewarding founders who move quickly and escape the early-adopter niche.
- •Founder-market fit: choose domains aligned with your strengths and passion
- •Shift from “single-player AI” to “multiplayer AI” (team productivity acceleration)
- •Goal: collapse months-long project cycles into hours during collaborative sessions
- •Vertical AI poised to scale quickly (legal, coding/context engineering, marketing/content/ROI)
- •Market will consolidate fast (~18–24 months); to build big, move very fast
Closing recommendations: books, favorite AI apps, and the daily founder question
The conversation ends with Andrey’s resources and personal operating advice. He recommends management/scaling books, shares the AI tools he relies on, and offers a simple daily question to sustain founder energy.
- •Books: *High Growth Handbook* and *High Output Management*
- •Favorite AI apps: Granola (AI note-taking), Perplexity, ChatGPT, Anthropic tools, plus Miro
- •He switches tools depending on the job-to-be-done
- •Founder habit: ask “Do I love what I’m doing?” each morning
- •If yes, that intrinsic energy sustains the grind; if no, reconsider the path
