A symbiotic designer’s practice
Symbiotic Designer OS is the operating system for my work with AI. I build it alongside custom skills for research, design and delivery planning. It gives those skills project context, and I check their output before making a decision.
My system for working with AI
I built a workspace that brings project knowledge, AI skills and decision rules together. Each new session starts with current sources, the reasoning behind earlier choices and a clear next step.
Building an AI workspace
I build a workspace for research and design. It needs a shared knowledge base, a way to select skills and a record of decisions across sessions.
Goal · AI boundaries · Criteria
I start with the brief, sources and earlier decisions. I specify what should come back for review.
Brief · Sources · Expected output
A research skill organises sources, a design skill develops options and a planning skill turns the chosen direction into next steps.
Research · Options · Delivery plan
I check what supports a recommendation. If the evidence is thin, I know what to verify before choosing a direction.
Evidence · Risk · Decision
I record the choice, its reasoning and the next action. The next session starts from those decisions.
Decisions · Rationale · Next step
Michał Aleksander / experience
I bring experience in usability research, leading a UX team and designing products to Symbiotic Designer OS. My work spans online shops and booking services through to advertising platforms.
See my career history on LinkedInUsability research and shopping journey design.
Leading a UX team, researching and designing educational products.
Booking journeys for an international travel service.
Designing buyer journeys for a classifieds platform across several markets.
Research and SaaS app design for crews working in the field.
Complex B2B platforms, design systems and rules for working with AI.
I started working in e-commerce in 2005. The dates mark the start of selected design roles.
A skill is a reusable instruction for AI. I choose one for the task, give it project material and define what should come back for review. For prototypes, I use the Design Cortex module.
Research Workbench · Design Conference Research
Design Direction · UX Architecture
Prototyping module
Three design decisions from Rekon projects.
Information architecture
On this interactive 3D map of Luboń, select a project to see its reported stage, who is responsible and which documents support the information.
See the Miasto z bliska case study
Visual exploration with AI
We compared five opening concepts. The chosen scene shows the operator at work alongside an example of anomaly detection.
i-Rays / five opening concepts
Prototype comparison
I compared seven watch-face concepts, including after an activity and with missing readings. I checked the simplified always-on display in the simulator.
ORBIT / from multisport to themed collectionsI start with what the audience needs to understand or do. I write down how we will know the result is ready to use.
I draw on conversations, sources, brand material and product rules. These give the chosen skill the context it needs.
I describe the intent and use AI to prepare options. Depending on the project, I check a print proof, a user journey or a working prototype.
I choose a version, record why and deliver the agreed result. If a finding can help future projects, I update the rule.
A case from practice / Design Cortex
A check found colours outside the design system. They came from external material, so I defined an area with its own rule and ran the prototype check again.
Read the full story on SubstackStudio sources
After Future Product Days I wrote about how a design system can guide an agent on when to use a pattern. That question came up again in Cortex when I set a colour rule for external material.
