Overview
─────
Reducing Setup Time by 96% for Developer-First Product
01
Challenge
The client’s platform enables companies to deploy vision AI solutions, e.g., a logistics firm using AI cameras to automate loading operations. However, the developer site was difficult to navigate, making it hard for developers to discover critical information, such as how to correctly set up the AI cameras, test the inference, and deploy AI models.
02
Discovery
I conducted a comprehensive audit, including information architecture analysis, user interviews, stakeholder interviews, and quantitative research across key touchpoints: documentation landing page, navigation, entry templates, and search experience.
During discovery, I also created a product ecosystem map, a visual representation showing how different products, services, platforms, and stakeholders interact and create value within the client’s defined market.
I also hands-on tested the SaaS platform as a new user, setting up Raspberry Pi AI cameras and exploring pre-trained templates.
03
Approach
- Identified key user needs
- Created navigation sitemap
- Designed low-fidelity content wireframes for key website touchpoints: documentation landing page, documentation navigation, documentation entry template, homepage
- Created a strategy doc for the documentation site setup
- Created a roadmap for the implementation phase for PMs and technical writers
Example of a tactical IA improvement recommended: During setup, I found the AI camera only works with 2.4 GHz WiFi. This crucial detail was missing from the documentation. Users with 5 GHz WiFi couldn’t enroll devices and were left unaware.
Clearly displaying this info and potential actions would prevent user frustration and reduce drop-offs, especially since many routers default to 5 GHz.
04
Impact
| Setup time reduction | From several days to ~1 hour (96% improvement) |
| Template findability | Significantly increased |
| User onboarding | Streamlined orientation for new developers |
05
Client feedback
Senior PM at a F100 company
“This was a messy project, and we appreciated Delfina taking on all aspects of its messiness.
What was so valuable with the service was how Delfina was able to put into words how the problems impact our users. We would definitely purchase these services again!”