Self-Service Resignation

Situation

Leaving a job is one of the most stressful professional experiences someone can face. When employees have direct control over how and when they resign, they feel respected, not processed - and this completely changes the temperature of the relationship between employee and employer. Self-service resignation gives people clarity on their timeline, visibility into next steps, and the dignity of handling something personal on their own terms.

For Amazon, that meant turning an anxiety-inducing moment into a straightforward, human experience. Prior to this initiative, there was no unified, self-service way for employees to submit a resignation, understand what happened next, or access resources that might address the very reasons they were leaving. Every resignation that could have been prevented by timely access to a growth resource, a different team, or a flex shift option represented recruiting and onboarding spend that had to be made all over again.

The People Experience Foundation team recognized that the exit flow was an underutilized intervention point, and that designing it well could reduce attrition while treating departing employees with the dignity they deserved.

"The exit flow wasn't just a process to close out. It was the last moment the organization had to listen - and sometimes, to change someone's mind."

The challenge

As the lead designer on this project, I owned the end-to-end experience across a cross-functional team spanning Engineering, Product Management, Legal, Compliance, and HR/People Experience leadership. That meant meeting directly with senior leaders to align on vision, running workshops with both stakeholders and employees to surface what the experience needed to be, and serving as the consistent advocate for making a necessarily complex process feel human.

The product had to serve two meaningfully different user populations - full-time Amazon employees and contingent workers operating under different access policies and legal frameworks. The design challenge was layered. At the surface, it was a multi-step form with conditional logic. Underneath, it was a set of deeply human tensions:

The hardest single challenge was making the business case for the project itself. Unlike a revenue-generating feature, this was an investment in people, and proving its value required translating human outcomes - retention, satisfaction - into financial ones: avoided recruiting spend, reduced attrition cost.

The flow

The resignation experience was designed as a 6-step guided flow, surfaced within the A to Z app on mobile and web. Each step was carefully sequenced to gather information, surface relevant resources, and confirm intent - while giving employees multiple natural off-ramps.

The flow opens with a personal acknowledgment - "Hi Tony, we're sorry to hear you're thinking of leaving Amazon" - before asking the employee to select a primary reason for leaving. The phrasing matters: it signals care without being saccharine. What follows is a conditional drill-down question set specific to the category selected, surfacing more nuanced signals about what's actually driving the decision. This information then informs all following screens, where alternative options are offered - ranging from training courses relevant to their career growth to alternate shift times if scheduling was the issue.

How I led it

Starting with alignment, not wireframes

Before any design work began, I met with People Experience leadership to establish a shared vision for what this experience needed to accomplish - including objective retention metrics, but also a clear understanding of how this affected the actual people using it. From there, I facilitated a series of workshops with stakeholders across Legal, Compliance, HR, Product, and Engineering.

With so many groups holding distinct and sometimes competing priorities, getting alignment early wasn't a courtesy - it was a design decision. Those sessions surfaced assumptions that would have otherwise become friction during reviews, and created the shared framework we returned to every time a tradeoff needed to be made.

Discovery: listening to employees who had already left

The research phase included user interviews and contextual inquiry sessions with both current employees who had considered resigning and former employees who had followed through. The dual perspective was critical. From it, we built service blueprints that tracked the experience of FTE and contingent workers in parallel, making visible where their needs diverged and where a single design could serve both.

One consistent finding stood out: employees leaving for "growth opportunities" often weren't aware of internal resources that directly addressed their concern. They had simply never encountered them at the right moment.

The pivot: cross-team training access

Midway through the project, a defining insight emerged from ongoing research and stakeholder conversations. The most meaningful signal wasn't just that employees wanted more training - it was that they wanted access to training from other teams across Amazon. They weren't just trying to do their current job better. They were trying to see a path forward inside the company.

This reframed the product's core value proposition. What began as a resignation form evolved into a retention intervention - a system that could meet employees where they were and offer them a reason to stay.

The Explore hub became a central design artifact: a surface that gave employees exiting for any reason a set of concrete alternatives - career growth resources, team openings and leader profiles, flex shift options, and self-service resignation - within a single, coherent experience. Each pathway was a potential reason to stay.

Designing the A to Z Assistant integration

One of the more technically and experientially complex aspects of the project was integrating the resignation flow with the A to Z Assistant, Amazon's AI-powered support tool embedded in the app. The Assistant needed to recognize resignation intent expressed in natural language, respond with empathy and useful resources, and guide employees through a structured decision path without feeling robotic.

Designing the conversational components required close collaboration with Engineering on the NLP layer and with Legal on the precise language used for consent and confirmation. Rapid prototyping and iterative stakeholder critique drove the majority of this work, with each round of feedback integrated into the next prototype before broader review.

Confirmation, compliance, and the right to change your mind

The final steps of the flow were shaped heavily by Legal and Compliance requirements. The final step presents a full summary of resignation details with inline edit affordances, giving employees clarity and control before the irreversible action of submitting. The design explicitly reinforces that employees can change their mind up to 24 hours before their last day.

Post-submission, employees land on a Resignation Information page that confirms their last day, surfaces a personal checklist of next steps, and provides a clear path to the alumni experience through the A to Z app. The cancel modal was designed with deliberate friction - confirming intent without being obstructive - reflecting a principle that ran through the entire design: respect the employee's autonomy while giving the organization every opportunity to make a compelling counter-offer.

Outcomes

The Self-Service Resignation flow shipped within the 3-6 month timeline and is now live in the A to Z app, serving tens of thousands of Amazon employees globally, across both full-time and contingent worker populations.

Making the business case for a people investment

Perhaps the most significant design outcome wasn't on the screen at all. By modeling the cost of attrition - the recruiting spend, onboarding time, and productivity loss associated with replacing a departing employee - and comparing it against the retention lift produced by the flow, I created a financial argument that translated human experience into business value. This reframing changed how People Experience leadership evaluated and prioritized employee-facing investments going forward.

The project didn't just reduce attrition. It changed the organization's mental model of what employee experience design is worth.

Org-level impact

The dual-journey framework developed for this project - tracking FTE and contingent worker experiences in parallel - has become a methodology reference for subsequent PX Foundation work. The integration pattern between a structured form flow and the A to Z Assistant is now a foundation for other self-service HR features. And the principle of designing exit flows as retention opportunities has been adopted more broadly across the People Experience organization.

Reflection

This project reinforced something I believe about design at enterprise scale: the hardest problems are rarely about interface. They're about alignment, trust, and the ability to speak the language of business while advocating for the language of people.

Navigating Legal, Compliance, HR, Product, and Engineering - each with legitimate and sometimes competing priorities - required the kind of facilitation and strategic communication that doesn't show up in a component library. I had to hold the human experience as a north star while translating it into language that resonated with lawyers, engineers, and business leaders. Designing for two user populations with different access rights and legal protections required systems thinking that went well beyond a single flow. And making the business case for a people investment, in an environment oriented around measurable output, required connecting human outcomes to financial ones with enough rigor to be taken seriously.

The moment I return to most is the pivot toward cross-team training access. It didn't come from a brief or a business requirement. It came from staying close to employees long enough to hear what they weren't quite saying: that they didn't just want to learn, they wanted to belong to an organization that believed they could grow. That's what we tried to build.