Sasha Kucharczyk · Case studies

Selected work.

Four one-page cases from Preteckt, the applied-AI company I co-founded. Each shows an operating problem, the analysis I led, the decision that followed, and the result.

01 / Market entry

Building a transit vertical

As COVID put existing trucking, OEM, and defence work under pressure, I identified public transit as an adjacent market and reframed our offer around reliable bus service.

I secured an anchor pilot, tested the positioning with other agencies, helped win ten more paid pilots, and shaped transit pricing and product priorities. Once the motion worked, I handed repeatable sales and account work to the team.

~$5M Signed annual transit contract value at full deployment, built over three years. Read the one-page case (PDF) ↗

02 / Unit economics

Reducing unit costs without weakening the product

Growth had made cost optimization a lower priority. When revenue fell short, I revisited unit economics without sacrificing the Repair Plans customers relied on.

I decomposed unit COGS and worked with expert reviewers and engineering to identify excess storage, collection, and telematics costs. I recommended the changes; engineering implemented them, while we continued monitoring product results.

~40% lower Total COGS per monitored vehicle-month, based on vendor cost structures. Read the one-page case (PDF) ↗

03 / Commercialization

Turning pilots into procurement

After strong early pilots, only about one in ten later sponsors pursued procurement, even as they said the product worked.

I interviewed customers and analyzed Repair Plan frequency, fleet size, adoption, and pilot length. I recommended a 50-vehicle, six-month format to create steadier value signals and decision-ready evidence.

~10% → ~50% Early pilot-to-procurement progression after the redesign; revenue per pilot also rose roughly 3×. Read the one-page case (PDF) ↗

04 / AI adoption

Making AI outputs work for frontline teams

A frontline maintenance use case lost value when Repair Plans began reaching technicians without the supervisor judgment that had previously focused the work.

I confirmed the productivity gap in the data, joined customer interviews and leadership discussions, and helped define shorter plans centered on the likeliest issue. Engineering implemented the output redesign.

~4% → ~56% Measured gain in maintenance work completed per labour hour at steady state. Read the one-page case (PDF) ↗

Let’s talk

Want to discuss the decisions behind the results?

I’m happy to share more context on the trade-offs, measurement, and what the cases taught us.

Email Sasha LinkedIn