
The brief: Dynamiq lets enterprises build, deploy and monitor AI workflows and agents inside their own infrastructure. The team came to us at pitch stage and needed a complex, technical product to feel approachable without hiding its power.
What I did:
- Helped kick off the project and set up how we worked - one dedicated designer and regular reviews.
- Reviewed flows and UI for the workflow builder, deployments, observability and guardrails.
- Joined client calls to align product vision and priorities.
- Guided the design system choice - GitHub Primer, customized to Dynamiq's brand.
- Directed brand, website and marketing work alongside the product.
Result: Dynamiq grew from an investor pitch to a full platform with one design language across product, site and marketing - on a design system the developers already knew.




Design process
1.Research & exploration
- Stakeholder interviews
- Competitor analysis
- User journeys & flows
- Low-fidelity wireframes
2.Branding
- Visual direction
- Client testing
- Brand guidelines
3.Design & prototyping
- High-fidelity screens
- Design system
- Work with developers
4.Website & marketing
- Website pages
- Ads, social and blog visuals
Research & exploration
I helped the team map user stories role by role - from an admin who approves deployments and invites users to a first-time user who signs up, creates a team and learns how workflows work.

Wireframing & iterations
Workflow creation and deployment were simplified in wireframes first, then refined over several rounds so moving between workflows, metrics and agents stays easy.

Design system
We built on GitHub Primer - familiar to the developers - and customized it to Dynamiq’s brand, which kept the product consistent as it grew.



Features
Workflows
A drag-and-drop builder for AI modules, states and tools like OpenAI and databases - visual enough for non-technical users.


Deployments & metrics
A guided path from development to production, with performance and efficiency metrics in real time.


Observability
Each deployment gets its own quality metrics, like relevance and correctness. Traces show every step of a run with its time and tokens, and the run tree opens the full input and output.



Collaboration
A chat for deployed assistants with shared and personal folders, so teams keep conversations organized and share them in one click.

Branding
A clean, technical identity - the logo, the type and illustrations of the integrations and model providers Dynamiq connects to.




Website
A website that explains the platform to enterprise buyers - on-premise deployment, compliance and the full toolset - with pages for industries like healthcare.



Marketing
Ads, social posts and blog covers in the same visual language - from a CTO’s guide to GenAI to articles on LLMs in banking.



