Use AI to help Sellers Sell
Use AI to help Sellers Sell
How I Led the End-to-End Product Strategy for Sugar AI — and Built an Architecture of Trust.
How I Led the End-to-End Product Strategy for Sugar AI — and Built an Architecture of Trust.
Embedded Files
AI PROTOTYPE
Before I dive into the case study, let me show you the video of AI prototype and the full result.
Before I dive into the case study, let me show you the video of AI prototype and the full result.
Full story of how I build the AI Prototype and challenges I encounter during the process is here.
The Case Study
The Case Study
THE PROBLEM
SugarCRM was at an inflection point. The company had decided to become AI-first — but "AI-first" is a strategy, not a product. Without a clear definition of what that meant for users, the risk was real and expensive: we would ship a generic chatbot that enterprise sales teams would reject on day one.
SELLERS ARE NOT LOOKING FOR MORE TOOLS...
Sellers are not looking for AI. They are looking for results. If the AI couldn't prove its value in the first interaction — in a high-stakes, high-skepticism enterprise sales environment — we would lose a multi-million dollar opportunity to lead the market in AI-driven CRM.
The pressure was not just on the product. It was on design to define what this product actually was.
MY ROLE
This was not a feature project. It was a 0-to-1 platform transformation, and I owned the design strategy end-to-end. During the process I coached 3 designers on LLM mental models, human-in-the-loop patterns, and technical constraints
This was not a feature project. It was a 0-to-1 platform transformation, and I owned the design strategy end-to-end. During the process I coached 3 designers on LLM mental models, human-in-the-loop patterns, and technical constraints
20+
User Interviews 20+
User
Research
User
Research
Led discovery alongside PMs and the VP of Strategy
100%
Responsibility100%
Service
Blueprinting
Service
Blueprinting
The blueprint became engineering's canonical roadmap
6+
Executives Involvement & Collaboration6+
C-level
Communication
C-level
Communication
Presented and aligned at C-suite level throughout
4+
Teams & Cross Collaboration4+
Team
Collaboration
Team
Collaboration
UX, PM, Engineering, and Customer Success.
~100h
Workshop Facilitations~100h
Workshop
Facilitation
Workshop
Facilitation
From discovery through MVP definition and delivery.
100%
Responsibility for AI Design System100%
AI Design
System
AI Design
System
Established interaction patterns, component standards, and accessibility principles across 4 enterprise platforms
The
Discovery
The
Discovery
STEP-BY-STEP SOLUTION
A Framework for
A Framework for
Strategic AI Adoption
Strategic AI Adoption
From research to implementation - SugarCRM became AI Platform
for Precision Selling.
From research to implementation - SugarCRM became AI Platform
for Precision Selling.
1.
Service
1.
Service
Blueprinting
Blueprinting
I visually synthesized complex research to align VP-level stakeholders on a unified, evidence-based roadmap.
2.
The Architecture
of Trust
2.
The Architecture
of Trust
I designed a side-by-side verification framework, allowing sellers to validate AI insights against CRM "Sources of Truth".
3.
Systemic
Scaling
3.
Systemic
Scaling
I established standardized "Response & Action" rules that were integrated into our core Design System to support all future AI features.
Define
MVP Target
Define
MVP Target
Research → Workshops → MVP Definition
Research → Workshops → MVP Definition
RESEARCH RESULTS
Sellers
do not need AI -
they need Strategic Partner
Sellers
do not need AI -
they need Strategic Partner
Research showed that Sellers are focused on selling and in essence, they do not need AI or any other tool. Instead they need a Strategic Partner.
Research showed that Sellers are focused on selling and in essence, they do not need AI or any other tool. Instead they need a Strategic Partner.
WORKSHOPS RESULTS
Use Data to
Use Data to
Define Priority
Define Priority
Using the Service Blueprinting process we identified that Pre-Meeting Prep was the highest friction point (MVP) with the lowest barrier to AI integration, allowing for a faster 0-to-1 launch.
Using the Service Blueprinting process we identified that Pre-Meeting Prep was the highest friction point (MVP) with the lowest barrier to AI integration, allowing for a faster 0-to-1 launch.
MVP DEFINITION
"Jobs to be Done"
"Jobs to be Done"
for the Conversational AI
for the Conversational AI
Based on the findings and cross-collaboration, we redefined our MVP goals from "Building a Chatbot" to creating the "Sales Assistant".
Based on the findings and cross-collaboration, we redefined our MVP goals from "Building a Chatbot" to creating the "Sales Assistant".
1.
Meeting
1.
Meeting
Prep
Prep
Pulling data from silos to prepare the seller before the call.
2.
Smart
Prompts
2.
Smart
Prompts
Removing the manual "remind me to call" task.
3.
Risk
Visualization
3.
Risk
Visualization
Using AI to highlight which deals were likely to stall.
DESIGN PHASE
Design AI
Experience for Trust.
Design AI
Experience for Trust.
I established the visual logic that allows the AI to act as a proactive partner—delivering 'Precision Selling' insights while ensuring the seller maintains absolute control over every customer touchpoint, so that users can gain trust in AI features.
I established the visual logic that allows the AI to act as a proactive partner—delivering 'Precision Selling' insights while ensuring the seller maintains absolute control over every customer touchpoint, so that users can gain trust in AI features.
1.
1.
Beyond the
Beyond the
Reactive Chatbot
Reactive Chatbot
Seamless interaction where the AI anticipates seller needs rather than waiting for manual inputs.
Seamless interaction where the AI anticipates seller needs rather than waiting for manual inputs.
2.
Give Users
Control
2.
Give Users
Control
Contextual validation and data provenance, allow sellers to compare AI suggestions side-by-side with the 'Source of Truth' in the CRM.
Contextual validation and data provenance, allow sellers to compare AI suggestions side-by-side with the 'Source of Truth' in the CRM.
3.
Keep Consistency
Across the Platform
3.
Keep Consistency
Across the Platform
Visual rules—ranging from KPI visualizations to conversational prompts — ensure extreme consistency across the platform.
Visual rules—ranging from KPI visualizations to conversational prompts — ensure extreme consistency across the platform.
4.
Transparency over
'Black Box' Failures
4.
Transparency over
'Black Box' Failures
This proactive feedback loop is what keeps users engaged even when the system faces limitations.
This proactive feedback loop is what keeps users engaged even when the system faces limitations.
THE FINAL UI LAYERS
'Human-in-the-loop' principles.
'Human-in-the-loop' principles.
While the AI provides the heavy lifting of data synthesis, the seller remains the strategic lead, empowered by precise insights.
While the AI provides the heavy lifting of data synthesis, the seller remains the strategic lead, empowered by precise insights.
IMPACT & STRATEGIC VALUE
22%
22%
Sales Velocity Gap
Driving Business Transformation
Driving Business Transformation
22% sales velocity gap identified directly influenced MVP scope decision. Service blueprint adopted as engineering's canonical roadmap — eliminating 3+ months of cross-functional realignment. AI Design System delivered on schedule across 4 platforms, unblocking 2 engineering teams. 4 PMs requested collaboration by name within a single quarter of design lead engagement. Prototype validated in 3 days using Claude Code, saving ~4 weeks of engineering speculation.
100%
100%
Business Alignment
From Confusion
to Clarity
From Confusion
to Clarity
The Service Blueprint, I initiated, became the Foundational Roadmap for Engineering and Product, eliminating cross-functional friction and ensuring 100% business alignment throughout the product lifecycle.
The Vision
The Vision
Raise the bar for all
future AI initiatives
Raise the bar for all
future AI initiatives
The AI Visual Rules and Interaction Logic I established, raise the bar for all future AI initiatives at SugarCRM, ensuring that as the platform scales, every AI-driven feature remains consistent, trustworthy,
and human-centric.
LESSON LEARNED
Sellers don't want AI to do their job
Sellers don't want AI to do their job
They want
AI to give them their time back.
They want
AI to give them their time back.
By distilling complex data into actionable momentum, we enabled sellers to move away from 'data-entry clerk' roles and back to what they do best: strategic closing.
UP NEXT
Page updated
Google Sites
Report abuse