# Upside Lab

> Upside Lab is a reviewed AI agency in Kraków, Poland, listed in the AIAgencies.eu register and last assessed 2026-09-08.

Canonical page: https://www.aiagencies.eu/agency/upside-lab
Generated: 2026-09-28 · Curated by Gabor Kiss · No paid placements — inclusion, score, and rank are never sold.

## Facts
- Location: Kraków, Poland
- Website: https://upsidelab.io
- Hourly rate: €85-130/hr (directory-listed)
- Team size: 10-49
- Founded: 2017
- Services: AI Consulting, AI Development, AI Automation
- Industries: Healthcare, E-commerce, SaaS & B2B
- LinkedIn: https://www.linkedin.com/company/upside-lab

## About
Kraków consultancy building enterprise AI on top of governed data platforms — AI strategy and roadmapping, AI-ready data foundations, custom AI applications, enterprise integration and MLOps. Also publishes its own open-source agentic AI platform for e-commerce and an LLM-driven ETL tool.

## Documented clients
Documented engagements: University of Toronto (T-CAIREM), SimOptions, Winderen, Deltabadger, Artemest, and Samsung. All 6 carry a source URL on file.

## Curator assessment
Reviewed 2026-09-08 by Gabor Kiss against published evidence.

The Health Data Nexus is the one to read: a governed platform built with the University of Toronto's T-CAIREM so researchers can work on real clinical data, cleared through a formal Privacy Impact Assessment and Threat Risk Assessment under PHIPA and TCPS 2, with isolated cloud environments and no download path.

By May 2025 it hosted nine datasets—among them 14,000 patients across 22,000 visits, over 1,000 CT scans and a voice dataset—served researchers from five countries and carried four datathons; note where the AI sits, though: the compute story ends at Google Vertex AI access for researchers rather than at a model Upside trained.

The rest of the portfolio is e-commerce and platform engineering with commercial numbers attached—17% conversion improvement at SimOptions, 10x revenue growth after Winderen's MVP, $6.5M of automated purchases at Deltabadger—and the firm publishes Enthusiast, an open-source e-commerce assistant that runs against hosted APIs or self-hosted models.

The weaker fit is a buyer who needs a model built and evaluated; what is demonstrated here is the governed data foundation and the product around AI, which is usually what decides whether a model ever reaches users.

### Key strengths
- Regulated data engineering is documented rather than claimed: PHIPA and TCPS 2 compliance, a Privacy Impact Assessment and Threat Risk Assessment, isolated environments with no download capability
- Adoption is reported, not promised—nine datasets, researchers in five countries, four datathons, and named hospital datasets behind them
- Publishes Enthusiast, an open-source e-commerce AI tool that supports self-hosted models such as LLaMA alongside SaaS APIs

### Good to know
- In the flagship engagement the AI is what the platform enables; Upside's own work is governance and infrastructure—ask what modeling the team has done itself
- Outcomes elsewhere are commercial (conversion, revenue, transaction volume) rather than model metrics—ask for an evaluation you can inspect
- At 10-49 people, ask who operates the platform after launch; the €85-130/h shown here comes from third-party listings rather than from Upside

---
Source: AIAgencies.eu — the curated register of European AI agencies. Rate provenance: "agency-confirmed" = disclosed via submission or claimed profile; "directory-listed" = Clutch or two agreeing B2B directories (lone unverified sources are never written). Full method: https://www.aiagencies.eu/methodology · Rate index: https://www.aiagencies.eu/rates