Project Details:
Client: E-commerce | Automotive industry | Auto parts | B2B2C | Marketplace
Cooperation: Since 2022 | 3rd joint tech product development process within the capital group.
Time: 4 months of cooperation / 4 weekly sprints / 22 days in total.
Goal: A “Last-chance” strategy for a complex tech product with a 10-year history that had never reached the break-even point.
Context
This involves a tech product in a highly competitive market (only 7% of the total TAM, shared among many aggressive international competitors). The product needed a strategic refresh due to current and upcoming market changes (e.g., new international agreements, the so-called “Trump tariffs,” new logistics chains, etc.) after 10 years of market presence without reaching profitability.
The Product:
The application is a professional Workshop Management System (WMS), oriented towards optimizing work during peak traffic periods (e.g., seasonal tire changes). It serves as an interactive command dashboard that connects human and hardware resource planning with advanced client communication.
The main goal of the application is maximizing workshop efficiency by eliminating “dead spots” in the schedule, automating the online reservation acceptance process, and building customer loyalty through precise targeting of service offers.
The application acts as a central operational hub, integrating incoming customer inquiries (online) with the physical capabilities of the workshop (number of lifts/bays). It allows for dynamic management of mechanics’ time and precise visit planning, taking into account the specifics of different services (from a 30-minute wheel change to multi-day engine repairs).
AI-Powered Product Development Process
This was the 3rd product development process conducted with the same company. The teams differed each time, but after a couple of joint projects, we knew how to operate together to make the process efficient and productive in terms of time (especially since the executive management team: CTO, Managing Director, were involved).
As usual, we were contracted based on two factors: scope & time (4 sprints), with the caveat that the process could change due to project specifics and progress.
Sprint Scope Breakdown
Crucial assumptions:
- The list of data to gather would be shared with the client before the process, and delivered before work commenced.
- Target group research would be conducted by the client team based on the process designed by me.
- After the second sprint, the client would receive 3 potential product development & new business model scenarios to consider before further development.
The actual process
As the client needed to develop product changes quickly due to the market situation, we decided to use an AI toolbox to speed up the research and analysis process. Here you can read more aout it and watch the video about it >>
Even though the research process was finally developed by me, it did prolong the entire strategic process for only 2 days, thanks to AI-powered product development practices.
Crucial Results
Product Strategy including New Pricing
The changes in the core of the business model were so crucial that the entire business model needed redefinition. As a result of the process, we:
- Defined 14 personas in the C (Customers) part of the business model, and 4 in the B (Business) part.
- Defined the Unique Value Proposition (UVP) per segment (general UVP + UVP targeted per each customer segment).
- Defined the offer for each persona among segments.
- Defined new automatic marketing & sales funnels.
- Defined a new dashboard of metrics to measure and put into the Balanced Scorecard.
- Redefined the income structure.
- Created a new partnership structure.
- Defined new pricing.
- Created the product development roadmap for 2 years ahead.
- Developed the marketing & sales roadmap for 2 years ahead.
Note: The target group and UVP per persona were defined more precisely, which clearly influenced the pricing. The new pricing was built based on the new functionality set that reflects user needs, as well as a “Good-Better-Best” scheme, including a freemium model and add-ons. This pricing reflected the value delivered to customers while allowing the brand to reach profitability based on specific plans.
(Re)Branding
Due to major changes in the UVP and target groups, and taking into account 10 years of market presence, there was a huge need to reposition the brand, including its visual language. One of the key results was the brand redesign.
The redesign process included:
- New tone of voice.
- Defined brand archetypes.
- Brand values.
- New value proposition translated into the claim and marketing messaging.
- New visuals including sign, typography, signet, and colors.
In the proposal for the board, I prepared 2 complete new brand books that were consistent with the portfolio branding. At the same time, they allowed for the creation of new logotypes for sub-brands easily and without additional costs using an AI model (I utilized the Nano Banana model by Google Gemini). The brand name remained the same.
Automation of the Marketing Funnel & Content Production
As the marketing team dedicated to the product is small, automating the content generation and publishing process was crucial. I built the funnel in a way that allowed easy automation based on established AI & media creation tools. The flow was designed using a modern AI video stack—starting with a single file video input and ending with many marketing outputs.
Recommended Tech Stack:
- OpusClip AI
- Google NotebookLM
- Grok Imagine
- Google AI Studio & Google Advanced (Gemini, Veo, Imagen models)
- WordPress
- MailerLite
Created outputs were also integrated into the pricing model as an addition to the UVP for product customers in terms of recognition and brand promotion within the B2B sector.
Refreshed Product Prototype with AI
One of the key results of the entire AI-powered product development process was the visualization of the modified product. As the number of product changes was huge (following new pricing, new product upselling strategy, new data visualization, new booking process, and a new set of add-ons, e.g., premium SMS packages), I decided not to use design tools like Figma or Sketch to visualize the new product flow. Instead, I decided to “vibe code” it from the ground up (front-end & back-end, just without integrations).
I used the Google Gemini 2.5 Pro model through Google AI Studio for vibe coding purposes. The model was fine-tuned with the new pricing plans, new target groups and personas, the list of desired functionalities, and the design of the app flow.
The app I redesigned had been coded by the previous software team for almost 2 years. However, due to vibe coding possibilities, I was able to redesign and prototype the app—including all crucial functionalities in front-end and back-end, as well as crucial corrections and add-ons—in less than 2 days.
Tech Stack of the Booking App: React, TypeScript, Tailwind, date-fns.
Architecture and Solutions
- ES Modules & Import Maps: The project uses native browser modules. Instead of a heavy bundler (like Webpack), dependencies are mapped and downloaded directly by the browser thanks to import maps.
- Responsive Design: The interface is fully responsive (RWD), automatically switching between daily and weekly views depending on screen resolution.
- Print Styles: Dedicated @media print styles in index.html, optimized for printing the daily schedule on A4 sheets (black and white view, hiding unnecessary UI elements).
Crucial Functionality
- Multi-Bay Calendar (Bay Management):
- Managing multiple stations simultaneously (lifts, diagnostic stations).
- Intuitive Drag & Drop system for moving reservations between hours and stations.
- Visual distinction of service categories (Tires, Fast-Fit, Heavy Mechanics).
- Online Booking Integration:
- Notification system for new visit requests sent by customers via the website.
- Quick acceptance and appointment assignment with “one click”.
- Advanced CRM and SMS Marketing Module:
- Client Segmentation: Distinguishing groups (fleet, clients with tire deposits).
- Loyalty Campaigns: Automatic generation of reminders for oil changes (e.g., for people who haven’t been to the service center in 6 months).
- Time-limited promotions for specific, less busy services.
- Analytics and Workload Planning:
- Weekly view presenting the percentage utilization of the workshop’s processing capacity.
- Real-time detection of appointment collisions.
- Operational Optimization:
- Dedicated Printing Mode: Generating black-and-white task lists for mechanics for a given day (according to the A4 standard).
- Management Reports: End-of-day summaries sent via email, containing visit status and new inquiries.
- Scalability System (Plans):
- Built-in subscription system limiting or expanding functions (number of stations, access to SMS) depending on the workshop size.
Business Functions (Mock-ups)
Subscription Logic: A subscription tier system (Start, Standard, Pro, Premium) controlling access to functions such as adding workstations or SMS campaigns.
CRM & Segmentation: Logic inside SmsModal.tsx simulating customer segmentation (fleet, deposits, oil change reminders).
Market analysis
The process was developed in the mid 2025, when major changes in terms of logistics chains, European Union regulations, US -EU tariffs changes due to the Trump policies happened. It in the obvious way interferred the markets and influenced the company, product, offer, and strategy itself.
The need for the analysis was broad and rapid, new data and changes appeared every day. As the scope of the analysis was really huge, I decided to use the Google Advanced (Gemini 2.5) uptrained with the internal data of the company, as well as insights coming from executives.
Market Analysis
The process was developed in mid-2025, when major changes in terms of logistics chains, European Union regulations, and US-EU tariff changes due to Trump policies occurred. This obviously interfered with the markets and influenced the company, product, offer, and the strategy itself.
The need for analysis was broad and rapid; new data and changes appeared every day. As the scope of the analysis was huge, I decided to use Google Advanced (Gemini 2.5) uptrained with the internal data of the company, as well as insights coming from executives.