Customer Success Managers (CSMs) at enterprise companies juggle complex relationships, shifting priorities, and endless context switching. Yet most Customer Success Platforms force them into rigid, object-driven workflows: logging activities in predetermined fields, navigating between multiple screens and modals to piece together account stories, and spending precious relationship-building time on administrative tasks.
Traditional CSP structure was creating friction where CSMs needed flow.
I set out to design an AI-enhanced notepad that would let CSMs capture thoughts organically while intelligently connecting those insights back to their structured work.
I conducted discovery interviews with our internal CSMs over Zoom, focusing on how they currently capture and act on customer insights. Key research questions included:
Walk me through how you prepare for a Monday morning. What information do you need to gather, and where do you currently find it?
If you could have a conversation with Totango, what would you ask it to help you with?
Show me how you currently take notes during customer meetings. What happens to those notes?
Through these conversations, I identified two core friction points:
Context Fragmentation
Critical insights were scattered across emails, tasks, and touchpoints (Totango's structured version of logging engagements). Forming a cohesive story about the state of a customer required looking across these records and holding context in your head, instead of in the platform.
Administrative Overhead:
To jot down a quick thought or action item, a CSM would have to create a task or touchpoint. The creation flows for these records open in modals, contain required fields to help with indexing, and then disappear into a far away data table on save.
We built and tested progressively complex prototypes:
Static mockups to validate the weekly organization concept
Interactive prototypes to test the /AI interaction patterns
Functional demos with sample data to validate the conversion workflows
Each iteration revealed new insights about how CSMs wanted to interact with AI assistance.
This wasn’t a solo design effort. I worked closely with our AI/ML team to understand what contextual information the system could realistically access and process. Together with product management, we identified which AI use cases would deliver the highest impact with our current technical capabilities.
The engineering team helped me understand the constraints around real-time data access, which shaped decisions about how quickly the AI could surface relevant account information.
The notepad is organized by weeks, each serving as its own entry point. CSMs can flip between past weeks to review previous thoughts, or jump to future weeks to jot down plans and reminders. This temporal organization mirrors how customer success managers naturally think about their work cycles.
The notepad lives as a slide-out drawer accessible from anywhere in the app. No context switching, no losing your train of thought. CSMs can capture insights while reviewing accounts, preparing for meetings, or analyzing customer health scores.
When they’re ready to take action, any line of text can be highlighted and converted to structured CRM objects — tasks, touchpoints, or objectives — with full context preserved.
The real magic happens when CSMs need AI assistance. By typing /AI or highlighting text, they can access intelligent prompts that understand their specific context:
Meeting Follow-ups: After a customer call, they can prompt the AI to draft a follow-up email based on the meeting transcript, then refine and send it directly from the notepad
Team Management: CS Leaders can ask the AI to analyze their direct reports’ performance and suggest specific accounts that need attention or tasks to assign
Task Creation: A simple to-do list becomes actionable with one click: each line converts to a structured task
Since this feature is still in active development, I focused on creating a measurement framework that would capture both user behavior and business impact from day one. Rather than waiting for post-launch data, I worked with our product and data teams to identify leading indicators that would tell us quickly whether we were solving the right problems in the right way.
The challenge with designing AI features is that success isn't just about adoption — it's about whether the AI assistance actually improves how people work. This required thinking beyond traditional feature metrics to understand the quality of AI interactions and their downstream effects on CSM productivity.
I identified two leading indicators that would predict success:
How well does this addition fit with their existing platform usage?
Frequency of /AI prompt usage per session
Time spent in notepad relative to other platform areas
Conversion rate from notepad to structured CRM objects, particularly emails
How integrated does our notepad feel with their day-to-day work streams?
Percentage of users accessing notepad from different app sections
Rate of cross-week navigation (indicating they’re building a useful knowledge base)
AI suggestion acceptance rates
I prioritized these specific metrics because they indicate the feature is solving our core hypothesis: that CSMs want to think naturally but still need structured outcomes. High AI prompt usage suggests they find the assistance valuable. Strong conversion rates from notes to objects means we’re successfully bridging unstructured thinking with actionable workflows.
The cross-week navigation metric is particularly telling. It indicates CSMs are building a persistent knowledge base rather than treating this as just another note-taking tool.
Designing AI experiences requires a different mindset than traditional feature design. The AI isn’t just a tool, it’s a collaborator that should understand context, intent, and workflow. The most successful AI interactions feel like natural extensions of human thinking rather than separate “AI features.”
The weekly organization structure taught me that temporal scaffolding can be just as important as hierarchical organization. CSMs think in cycles, and our tools should match that mental model.
Most importantly, this project reinforced that the best B2B design doesn’t just solve functional problems, it reduces cognitive load and lets professionals focus on what they do best: building relationships and driving outcomes.
