The tech world is full of brilliant prototypes that never see commercial success. Often, it’s not because the tech doesn’t work — it’s because the founders can’t bridge the gap between scientific innovation and market adoption.
Katerina Axelsson’s journey with TASTRY is a masterclass in how to close that gap. Starting as a quality control chemist in the wine industry, she built AI that could “teach computers how to taste.” But her real achievement wasn’t just in the tech — it was in translating that complex innovation into a platform that solves measurable, high-value problems for multiple stakeholders in the supply chain.
Lesson 1: Start with a Precise, Painful Problem
TASTRY’s origin story begins with a surprisingly simple insight: two identical wines, sold under different labels, can perform drastically differently in the market.
Katerina recognized this wasn’t just a wine issue — it was a universal problem in consumer product development:
-
High product failure rates (≈85%).
-
No reliable way to predict consumer preference before launch.
-
R&D driven by intuition, not data.
She focused on building tech that could connect the chemistry of a product to the palates of individual consumers. The early target? Wine — a market with complex sensory variables, massive waste, and high margins for winners.
Lesson 2: Build Proprietary Data Before You Build the Business Model
Unlike many SaaS founders who start with a software interface, TASTRY began with a data asset — a proprietary “flavor matrix” mapped at the molecular level.
The problem? They couldn’t just scrape or buy this data — it had to be generated in-house. The solution was to launch an API-driven personalized wine recommender, deployed in grocery retail as a whimsical 10-question quiz.
The benefits were twofold:
-
Validated the technology in-market.
-
Collected high-quality consumer palate data to fuel their AI engine.
This dataset became their defensible moat.
Lesson 3: Sequence Your Markets for Maximum Leverage
TASTRY’s go-to-market path is a blueprint SaaS founders can borrow:
-
Initial wedge → Consumer-facing wine recommendations in retail (data gathering + early revenue).
-
Next layer → Sell insights to retailers to optimize shelf assortments and reduce waste.
-
Full integration → Work directly with manufacturers on product formulation, market entry targeting, and launch risk mitigation.
By vertically integrating after proving the tech in multiple contexts, TASTRY created a flywheel:
-
Retail deployments feed better manufacturer insights.
-
Manufacturer partnerships increase retailer value.
-
Both sides reinforce each other’s adoption.
Lesson 4: Position the Tech as an Outcome, Not a Process
TASTRY’s AI and machine learning models are deeply technical — but the pitch isn’t “AI flavor profiling.”
It’s:
-
“Increase your wine sales and margins.”
-
“Know exactly where your product will sell before you launch.”
-
“Reduce product waste and failed launches.”
For SaaS & tech companies, this is critical: buyers don’t want your algorithm — they want the business impact your algorithm creates.
Lesson 5: Use Constraints as Acceleration Points
COVID could have slowed TASTRY’s growth — retail expansion stalled for nine months. Instead, they pulled their 2021 product roadmap forward and went straight to wineries with market prediction and product formulation tools.
This “accidental acceleration” gave them traction on both sides of the supply chain a year earlier than planned, strengthening their vertical integration strategy.
Lesson 6: Hire for Belief, Not Just Skill
Katerina’s hiring philosophy is blunt:
-
Convince candidates not to join.
-
See who leans in anyway.
One of her best hires turned down offers from top tech companies, accepted 40% of his lowest salary offer, and asked to trade more salary for equity. That level of alignment can’t be bribed with perks — it comes from belief in the mission.
Consulting Takeaway for SaaS & Tech Leaders
TASTRY’s growth offers a repeatable framework for deep tech commercialization:
-
Anchor your innovation to a specific, high-value problem.
-
Build your proprietary dataset before scaling the platform.
-
Sequence market entry to build leverage layer by layer.
-
Sell outcomes, not processes.
-
Turn external shocks into acceleration sprints.
-
Hire missionaries, not mercenaries.
When you can combine cutting-edge tech with a disciplined, buyer-centric go-to-market sequence, you don’t just build a product — you build a platform that changes an industry.