Client work, and my own products under KateThought Studio. Client cases are anonymized: no product or client names, no screenshots.
Self-Service re-platform and slot booking: enterprise IT-migration SaaS
The product. An enterprise platform that large organisations and financial institutions use to migrate and modernise their IT estates: devices, software, users. Each client runs its own instance, and decisions about thousands of devices are made on its data.
The ask. Port Self-Service, the platform’s data-collection pages, to a new architecture and UI. The legacy interface and the new front end ran over the same operational database, and user answers wrote straight back into it.
What I did. I joined as a Feature Product Owner and was promoted to Area Product Owner within six months, responsible for two development teams. Self-Service was one of our focus areas. Every “port it as is” had to become an explicit keep / change / drop decision. Requirements came through in-house consultants, with no direct access to end users. The hardest part was slot booking: letting employees book device-migration slots from capacity-managed calendars.
What I delivered. User stories with full acceptance criteria, flow definitions and rule sets, joint design reviews with the designer and product manager, and release demos on realistic test data.
Outcome. The feature shipped and worked reliably. Booking edge cases were caught in specification and testing, not in production incidents. Slot-based self-scheduling is still one of the product’s publicly promoted capabilities.
Customer-journey & usability review: B2B messaging SaaS (EU)
The product. An early-stage B2B SaaS for customer messaging in small, owner-operated businesses: cafés, salons, repair shops. AI drafts every reply, and a person must approve each message before it is sent.
The ask. Go through the product as a first-time visitor, from the landing page to the decision to sign up. Say what is clear, what is confusing, and where the value lands.
What I did. I built three small-business personas, each testing one thing: the full desktop journey, the same journey on a phone, and approving a reply in a language you don’t fully read. I combined a 5-second test, usability heuristics, journey mapping, a funnel review and a cookie-consent check.
What I delivered. In 3 business days: a written report with 11 severity-rated findings (2 critical, 3 important, 6 minor), a prioritised fix list, and a 15-minute think-aloud Loom recorded in persona.
Outcome. The client shipped one fix before the report was delivered and closed the contract with a 5.0 review: “We changed so many things after her review and detailed report.”
Onboarding & activation audit: vertical B2B CRM (US)
The product. A vertical CRM sold to small US teams. In most accounts one person does the whole job: they capture data in the field on an iPad, and months later they work with it at a desk on the web.
The ask. The founder ran onboarding personally and wanted to keep that customer contact. The brief was to find where the product itself could carry more of onboarding.
What I did. I went through first login to first value on both devices, importing my own deliberately messy files rather than only reviewing screens. I mapped real customer questions to the places in the product where users got stuck. There was no product analytics, so evidence came from the product itself, walkthrough videos and the founder’s call.
What I delivered. In one week: a written report with 23 severity-rated findings, prioritised recommendations, three clickable screen prototypes, a proposed first sprint, and a measurement plan with one activation definition.
Key finding. Of eleven real customer questions, not one was a bug report. Every one was about finding something that already existed. So the fix was guidance at the task, not another tour at first login.
Outcome. 5.0 review on Upwork. My work ended at handoff, so I have no post-implementation results to report.
GymBuddy: Guided strength
My own product: an Android app for strength training that I designed and built solo with AI coding tools. Structured weekly planning, set-by-set workout logging, history, and conservative adaptation of future weeks. Released on Google Play in October 2026 after a closed test with 12 testers.
See the mobile fitness concept and interactive onboarding prototype →
GymBuddy: coach’s sight
in development
A companion Android app for personal trainers, built on the GymBuddy training engine.
A coach invites a client with a one-time link. The client fills in a short web questionnaire in English or Russian, with no account and no app to install. The coach reviews the answers and accepts the client, and GymBuddy’s rules build the program. From then on the coach runs every workout from their own phone: logging sets, rescheduling or cancelling, for up to three clients at once.
Clients are known to the coach by a pseudonym and can delete their data at any time with a personal code.
The screens are designed, and the app, the server side and the client questionnaire are in development. A small pilot with personal trainers comes next.