Now Assist Context menu
Designed a platform-level setup wizard that empowers enterprise AI admins to easily configure and customize AI widgets.
Challenge
With the rapid adoption of AI capabilities across ServiceNow products, customer demand for greater control has significantly increased. A common request is the ability to customize AI widgets that surface contextual outputs—whether within forms, writing editors, or side panels—to better align with their users' workflows.
Key goals
-
Tailor AI output placement to match workflow needs.
-
Control what content the AI surface
-
​Ensure AI UI components match our brand and language
-
Experiment and iterate with minimal technical effort
-
Monitor and improve AI engagement
Business impact
Faster roll out
Enables faster rollout of GenAI capabilities across the organization, reducing time-to-value and supporting broader adoption at scale.
Scalability
Provides flexibility to evolve with business needs, accelerating innovation by empowering teams to build and deploy tailored GenAI solutions efficiently.
Improved decisioN
Supports smarter business decisions and responsible AI use by surfacing context, dependencies, and expected outcomes of each skill.
Project timeline and team
Timeline: 2 weeks
Collaboration: 2 product managers, PMs from the business units, 1 content writer, 8 engineers, 1 designer
Design approach
Internal research and audit: Conducted a thorough review of the existing implementation and configuration process. Discovered that all customization was handled by engineers, who were overwhelmed by a high volume of requests from various business units. Compiled and prioritized these requests based on frequency and impact to inform design decisions.
Design thinking workshop: Led a cross-functional design thinking workshop with key stakeholders—including PMs, engineers, and solution architects from the Now Assist AI Admin team, along with representatives from all connected product areas—to align on user needs, pain points, and integration opportunities across touchpoints.
Rapid iterative test: Since we only had a week left for getting the designs ready, I took the approach of rapid iterative testing and prototyped a first veriosn of it to take feedback from multiple stakeholders and internal admins as quickly as possible. Synthesized and incorporated the findings to further refine it.
