What if running payroll was as simple as asking for what you need?

The legacy experience required users to navigate multiple screens, understand system specific workflows and execute a sequence of operations before discovering whether something had gone wrong.

When issues appeared at the end of the process, users had to investigate the cause, correct the underlying problems, roll back and run payroll again.

The platform also became unavailable for other work while certain processing operations were running.

Meanwhile, accessing workforce information such as employees, timesheets and rosters required users to navigate different areas of the system and understand where information lived.

The underlying problem was not simply poor navigation. It was the way users accessed information, executed operations and handled exceptions.

Rather than improving every legacy screen individually, I saw an opportunity to rethink the interaction model itself.

What if users could describe what they needed in natural language, while the system handled the complexity behind the scenes?

This became the foundation of the product vision: a conversational layer that could connect users to workforce data, business rules and operational workflows without requiring them to understand the entire platform.

Payroll became the first use case. The broader ambition was to establish a reusable capability that could extend to other modules across the Workforce Management System.

The run stays visible outside the conversation

When the conversation isn't active, the dashboard keeps the full picture of each run: its scope, every step in the timeline, its status and what is blocking it, such as unapproved timesheets or overtime to review.

Users can check progress at any time without asking again, and recover with a rollback when they need to.

The difference is not just fewer clicks. It is a different operating model, designed to prevent avoidable delays and reduce the effort required to complete complex work.

1. 01

   Make complex capabilities accessible

   Use natural language to reduce the need for users to understand system specific navigation. The goal was not to eliminate payroll expertise where business decisions require it, but to reduce dependency on technical knowledge of the platform.

2. 02

   Move validation earlier

   Identify prerequisites and potential blockers before critical execution stages. Surface relevant information, generate alerts and notify the people who can resolve issues.

3. 03

   Separate interaction from execution

   Allow users to interact flexibly through conversation while retaining explicit business rules, validation logic and controlled state transitions for sensitive operations.

4. 04

   Remove unnecessary waiting

   Enable processing to continue in the background, with status updates and notifications allowing users to remain productive while operations progress.

5. 05

   Establish a reusable product capability

   Design the initial implementation as a foundation for conversational access to other workforce management capabilities, rather than as a standalone payroll chatbot.

The experience combines an LLM, a set of tools connected to APIs and deterministic workflows. Each component has a distinct responsibility.

The conversational layer makes interaction more flexible. The deterministic workflow preserves structured execution where the business process requires it.

The same conversational model makes it possible to access operational information across the platform.

Instead of navigating multiple modules to locate employee information, review timesheets or inspect rosters, users can ask for the information they need and interact with the relevant capabilities through connected tools.

The key change is that users no longer need to know which section of the platform contains each piece of information. They can focus on the operational task rather than learning the system.

One of the most important changes was the processing model behind payroll.

In the legacy approach, timesheet processing was concentrated around payroll execution. Clockings, punch records, blocking logic and related processing accumulated within the same critical window.

In the redesigned approach, timesheets are processed daily. Relevant clocking and punch processing, including applicable blocking logic, is handled in advance.

When payroll is initiated, the remaining work can focus on the final calculations, including workload, penalties and allowances, together with any remaining applicable validations.

Combined with background processing and event based notifications, this reduces the amount of work concentrated in the payroll window and avoids making users wait for every operation to finish.

Two distinct improvements

• A shorter critical processing window

  Daily preparation reduces the work that remains when payroll execution begins.

• Less interruption to the user

  Background operations and event driven updates allow users to continue working while the system processes tasks.

These are related but different outcomes. One concerns the processing time of the business operation. The other concerns the time users spend waiting and the availability of the platform during that operation.

User validation and process design helped establish four core product outcomes.

• Reduced dependency on technical system knowledge

  The conversational experience provides a more accessible path to executing workflows without requiring users to learn the full platform navigation.

• A non blocking operating experience

  Background processing and event driven updates allow users to continue working instead of waiting for an operation to finish before using the platform again.

• Less work concentrated in the payroll window

  Daily timesheet processing prepares relevant information in advance, leaving fewer processing steps to complete when payroll is initiated.

• Earlier identification and resolution of blockers

  Pre execution validations, contextual guidance and notifications help surface issues before they disrupt critical processing.

How we would measure success

The next step is to quantify the impact through production and usability data.

• User independence

  Measure successful task completion by new users, onboarding time and the proportion of users who complete workflows without specialised assistance.

• Processing efficiency

  Measure total payroll processing time, the proportion of timesheets prepared before payroll and the time spent resolving blockers.

• Platform availability

  Measure active waiting time, time during which users cannot continue other work and the proportion of operations that run without blocking the wider platform.

• Execution reliability

  Measure payroll completion without rollback or reexecution, along with the number of correction cycles and the time needed to resolve issues.

These measures separate the outcomes observed through validation from the quantitative results that still need to be established.

I led the product vision and use case definition, translating operational pain points into a prioritised strategy for conversational payroll execution.

Working closely with engineering, I prioritised the tools and workflows, helped define the business rules and shaped the conversational experience required to connect user intent with product capabilities.

In collaboration with the technology team and Chief Architect, I contributed to the design of the solution architecture, aligning the product direction with the technical foundations required for future expansion.

My focus was to ensure that the first implementation solved a meaningful operational problem while establishing a reusable approach to connecting data, tools and workflows across the Workforce Management System.

Product leadership responsibilities

1. 01

   Defined the product vision and prioritised initial use cases.

2. 02

   Prioritised tools and workflows with engineering.

3. 03

   Shaped business rules, validation behaviour and conversational interactions.

4. 04

   Collaborated with engineering and the Chief Architect on the solution architecture.

5. 05

   Established a foundation intended to extend beyond the initial payroll use case.

Zero Touch Payroll was the first step towards a broader product vision: a Workforce Management System that users can interact with through intent rather than relying exclusively on traditional navigation.

The long term opportunity is to connect more workforce data, tools and workflows through a consistent conversational experience, making operational capabilities easier to access while preserving the controls required by critical business processes.

The ambition is not simply to make software easier to navigate. It is to make complex work easier to understand, initiate and manage.