Deal registration left partners waiting and uncertain
Before submitting, partners couldn't tell whether a deal would qualify. They had to work out on their own which deal type to choose and which account to use. After submitting, they couldn't see where a deal stood, why it was denied, or how to fix it without starting over.
Every role built its own spreadsheet
The system didn't give anyone a complete view, so each group tracked deals on the side.
BEFORE: Statuses used internal terms like Pipeline and Demand Acceleration, and the Status Details column was empty. Nothing told a partner which deals needed them.
The ask was a better form. The real problem was trust in the decision.
When I joined, the team had already started improving the registration form and post-submission experience as separate initiatives. But partners experienced them as one journey. I connected the two efforts around a shared goal: help partners understand eligibility before submitting and know what to do after a decision.
Mapping the journey and the process behind it
I synthesized UX Research findings into a journey map, user workflow, and service blueprint, connecting partner pain points to the policies, systems, and reviews behind them.
Two breakdowns emerged:
Before submitting, eligibility checks happened too late, so fixable issues surfaced only at denial.
After submitting, partners couldn't understand decisions or fix a deal without starting over.
Mapping the full journey revealed where earlier validation and clearer decision guidance could prevent dead ends.
Aligning the team around one problem
I facilitated a workshop with 10 cross-functional stakeholders from Partner Engagement, Product, Business, AI Engineering, and Engineering.
Using the journey map, I asked:
What problem are we really solving?
We aligned on one problem definition and three experience principles:
This changed the plan.
Phase 1 focused on visibility and correction after submission, where pain was highest.
Phase 2 eligibility guidance moved into the form redesign, where it could be built on clear rules.
Transparency · Manage Deal Registration
From rejection to resolution
Partners received denials without clear explanations, and correcting a registration often meant starting over. I introduced Needs Attention, a new state that turns preventable denials into actionable tasks.
Prioritize what needs action: Surface registrations requiring attention at the top of the overview.
Use clear statuses: Replace internal terminology with Submitted, Under Review, Decided, and Closed.
Make deadlines visible: Show required actions and response deadlines.
Correct without restarting: Update information or accept proposed deal-type changes in place.
Keep everyone connected: Centralize comments and communication with the partner team.
SHIPPED: Needs Attention brings actionable registrations to the top, with clear statuses and deadlines.
SHIPPED: Partners can accept a change from Resale to Influence without restarting their registration.
Clarity · The registration form
Fix the workflow first
The original six-step form buried approval-critical information and relied on internal program terminology which created confusion. I redesigned the workflow around how partners actually sell.
Organize by selling motion: Group deal types into Sell Through and Sell With.
Speak the partner's language: Describe each deal type by the partner's role rather than internal program terminology.
Prioritize critical information: Bring products, customer details, and approval requirements forward.
Prevent avoidable errors: Disable ineligible deal types based on the partner’s territory eligibility and the customer’s business type, preventing invalid selections before submission.
Reduce complexity: Simplify the process from six steps to three.
BEFORE: Five deal types used internal program language, with an error shown before the partner entered any information.
VALIDATED DESIGN: Partners select products first, then their role in the deal, with eligibility guidance before submission.
First, make the form make sense. Then make it smart.
A clearer experience, with actionable next steps
Partners can understand where their registrations stand, what needs attention, and how to resolve issues without starting over. During user acceptance testing, the Partner Engagement team found the redesigned layout more intuitive, with less clutter and clearer information.
"The new layout is more intuitive and no longer overloaded with content that's difficult to understand."
— Paraphrased feedback, Partner Engagement team, user acceptance testing
Designing AI that guides, not decides
With the core workflows simplified, I explored how AI could help partners anticipate approval risks and move opportunities toward closing, with explainable guidance, protected information, and human oversight.
Approval Confidence Score
Know what's holding a registration back
I designed the Approval Confidence Score to evaluate registration readiness against business rules and give partners actionable guidance before submission.
Partner eligibility: Program requirements and financial compliance.
Territory authorization: Whether the partner can register the customer in the selected country.
Product and deal type: Whether the selected selling motion meets eligibility requirements.
Potential duplicate conflicts: Whether another registration or opportunity may affect approval.
Registration completeness: Whether the required information supports review.
Partners receive a High or Low confidence assessment, with explanations and next steps. The score supports approval readiness, not an automatic approval decision.
AI CONCEPT: Approval readiness with explainable risks and privacy-safe guidance.
Deal Coach · Agent-Assisted Sales Progression
Know the next move. Keep every deal moving.
Approval is only the beginning. I explored an AI-powered Deal Coach that guides partners across five sales stages, from Discovery through Transition.
Guide the next step: Highlight outstanding tasks based on the deal's current stage.
Make progress visible: Show completed and pending tasks, milestones, and deadlines.
Coordinate with CRM agents: Reflect activities logged by the CRM agent alongside updates entered by the partner.
Preserve accountability: Identify where information came from and distinguish agent-managed records from partner-editable notes.
Keep partners in control: Require confirmation and supporting notes where appropriate before marking tasks complete.
FUTURE-STATE CONCEPT: Contextual sales guidance connecting next actions, stage progress, and CRM-agent activity.
Agent Behavior · Trust by Design
Intelligence with boundaries
I defined expected agent behaviors and evaluation scenarios across three principles:
Privacy: Use restricted information for authorized checks without exposing other partners' business activity.
Explainability: Show why a risk or recommendation matters and what to do next.
Human oversight: Escalate exceptions and preserve human control over consequential decisions.
For Deal Coach, this also meant making agent-generated activity distinguishable from partner-confirmed work.
A technically correct AI response can still be wrong if it reveals confidential information or acts without authorization.
Prototyping · Aligning Business and Engineering
Turning ambiguous requirements into testable interactions
I built working prototypes with Claude in a day and reviewed them with eight stakeholders. The sessions uncovered missing territory rules and helped Product, Business, and AI Engineering align on expected behavior, partner guidance, and human handoffs.
AI evaluates and guides. Humans retain control. Private information stays protected.
