Context
Genetec’s unified platform includes a Reports module with different types of data searches (Access control, Alarms, Bookmarks, Camera events, Incidents). A new category was introduced: Forensic search, designed to help operators locate people and vehicles captured across multiple cameras. This feature leveraged existing ML models to recognize attributes such as clothing, vehicle type, and colors.
The goal was to provide an experience that could combine traditional filters with the convenience of natural language queries (NLQ), for example:
“Find a red car and a person with a black shirt near the north entrance yesterday.”
Challenge
Design work was not about building the technology from scratch, but rather about integrating an already existing capability into the product in a way that was usable, clear, and scalable.
Key constraints:
- Limited user research: end-users (security operators) were very difficult to access directly. Needs were relayed through sales and account managers who had closer ties with customers.
- Language coverage: the interface had to support 6 languages from day one.
- Technical constraints: the global search bar could not be reused for NLQ due to architectural limitations, because they are using micro front-ends (MFE).
- Time-to-market pressure: development teams already had parsing and detection models working, the priority was to design a simple, intuitive entry point.
Approach
The design process was pragmatic, balancing what was technically possible with what would be clear to users:
- Mapped user needs via conversations with sales/account teams: operators wanted less manual filtering, faster workflows, and a way to describe what they were looking for in plain terms.
- Collaborated with developers to understand what was already available and what would be costly to change.
- Explored three integration concepts, focusing on different balances between discoverability, clarity, and effort.
- Considered multilingual aspects early: placeholder texts, examples, and query feedback had to be designed in a way that worked across all supported languages.
Design Concepts
Concept 1:
Slash Command in Search Bar
- Typing “/” activates NLQ mode in the global search bar.
- Very lightweight to implement.
- Suitable for advanced users but lacks visibility for the broader audience.
Concept 2: Global Assistant Button
- Dedicated button next to the global search.
- Opens a modal with input field, speech-to-text option, recent queries, and keyword highlighting.
- Strong discoverability, positions the feature as a flagship capability.
- Redirects users to Reports with filters pre-applied.
- More effort required from a development perspective.
Concept 3: Reports-only Assistant
- A new button placed within Reports, next to the Generate action.
- Opens a modal similar to Concept 2.
- Clearer scope: the Assistant is framed as an alternative to filters.
- Less globally visible but simpler to deliver.
Evaluation
| Criteria | Concept 1 | Concept 2 | Concept 3 |
|---|
| Discoverability | Low | High | Medium |
| Development effort | Low | High | Medium |
| UX clarity | Risk of confusion | Strong | Clear but scoped |
| Market positioning | Low | High | Medium |
- Short-term recommendation: Concept 3, as a pragmatic compromise balancing feasibility and clarity.
- Long-term vision: evolve towards Concept 2, which elevates the feature as a recognizable capability across the product.
Outcome
- Delivered wireframes, interactive mockups, and flows illustrating each concept.
- Defined a roadmap: start with Concept 3 to enable rapid adoption, then invest in Concept 2 for broader visibility.
- Designed multilingual interface elements: labels, hints, and parsing feedback adapted for 6 languages.
- Helped the development team integrate the NLQ feature quickly without architectural rework.
Key Learnings
Proxy research: when direct access to end-users is limited, sales and account managers can act as valuable proxies for uncovering needs and expectations.
Designing for feasibility: in contexts where the technology is ahead of design, the role of UX is to frame the capability in a usable and discoverable way.
Progressive disclosure: offering different entry points (shortcuts vs. buttons) balances the needs of power users and new users.