WorkRippling · Time policies

Time AI copilot

Explain, create, and validate policies; answer “why” outcomes

Role: Lead Designer, Time
Timeline: Jun 2025 concept → Q4 2025 build
Product area: Time policies
Scope: Explain · Create/Edit · Check
Status: Platform rollout in Q1 2026

In-context answers that translate configuration into plain language

Behind the work

Role

Lead Designer, Time — identified high-ROI use cases, designed flows/prompts/UI, created 3 prototypes, aligned with Platform for shared Copilot patterns.

With

Platform Design/Eng • Time Eng (prototype spike) • Leadership reviewers.

Project snapshot

Timeline
Concept Jun‘25 → Build Q4 25
Scope
Explain/Create/Edit/Check
Alignment
Platform primitives

Checkout the Rippling AI launch announcement!Link coming soon

Overview

Time policies (Time tracking, Overtime, Breaks, Time off) are rich and jurisdictional—great power, little clarity. I defined an AI Framework for Time: a copilot that explains policies, creates/edits them via natural language with real-time compliance checks, and answers employees’ “why” outcomes (e.g., no OT applied). The work aligned with Platform AI primitives now shipping across Rippling features, with Time implementation starting Q4 2025 to reduce policy-related support tickets and increase trust.

The problem

Great power, little clarity

Time policies (Time tracking, Overtime, Breaks, Time off) are rich and jurisdictional. The configuration is powerful; the outcomes are hard to explain.

Complexity
Global workforce management is hard—customers struggle with country/state rules
Support burden
Routine “what/why/how” questions pull Eng, managers, and support off higher-value work.
Configuration fear
Admins hesitate to customize defaults; risk of misconfiguration.
Decentralized AI efforts
Needed Time-specific use cases to guide platform investment

Time Off configuration – Australia – Time policies

FIG. 02Layers of policy complexityFour policy types, each carrying country and state rules, each customizable per company.

Solution highlights

Explain. Create. Check.

A copilot that explains policies, creates/edits them via natural language with real-time compliance checks, and answers employees’ “why” outcomes (e.g., no OT applied).

FIG. 03The three capabilities, and how they hand off to each other

AI explanation and policy-editing video — Plain-language answers; compliant edits; proactive risk scanning.

Explainability (in-context)

“Summarize UK maternity,” “Explain AU annual leave payout at termination,” “Why no OT today?”

Creation & editing

NL setup: “Create UK maternity policy with 100% pay for 12 weeks”; real-time compliance checks with citations; what-if scenarios.

Proactive risk

Run compliance across policies; flag gaps; offer compliant presets and trade-off notes.

Reusable patterns

Standard entry points, chat panel behaviors, response cards, sources/retention/guardrails; shared across Rippling features.

Process · 6 chapters

Feasibility

Could plain language summarize a policy?

Engineering proof of concept for plain-language AI summaries

Time AI feasibility primary product UI concept — FIG. 05The first proof of concept and a later concept exploration, side by side.

Time AI feasibility product UI concept three — FIG. 05The first proof of concept and a later concept exploration, side by side.

Discovery

Where the confusion actually lived

Support analysis → “explainability/why/how” ticket clusters; biggest ROI areas

FIG. 06Support themes behind the paradoxRoutine what, why, and how questions clustered around explainability.

Research

The pain points

The current reality for our customers is a paradox: Rippling provides the power to configure highly customized policies, but this power often comes at the cost of clarity. Our policy creation forms, while comprehensive, are incredibly complex. This leads to:

Confusion and errors

Customers often find it challenging to grasp the implications of each setting, leading to mistakes that can result in non-compliance or unexpected financial burdens

Increased support burden

This confusion results in a significant number of customer support tickets, as administrators seek assistance to understand or rectify their policy configurations. This not only drains our resources but also frustrates our customers.

Hesitation to customize

While we offer ‘Rippling Default Policies’ that comply with statutory minimums, such as a UK Maternity Leave policy, customers frequently wish to enhance them—like raising maternity pay from 80% to 100%. However, the fear of non-compliance or misinterpreting the effects of their changes often holds them back, forcing them to choose between providing better benefits for their employees and sticking to a rigid default.

Initial strategy

Our CEO challenged the team to devise a strategy for reducing customer support tickets. As the company explored ways to integrate AI experiences into our product, I recognized that enhancing explainability through AI capabilities was a key part of the solution.

Common questions such as, “How is Overtime calculated?” “Am I eligible for a payout due to a missed break?” and “How much time off can I take at year-end?” frequently reach Managers, who often struggle to interpret policies set by others in the organization. AI could play a crucial role in addressing these inquiries.

Explainability

The initial focus was on enhancing product feature explainability to clarify Time policies. An engineering partner developed a proof-of-concept using AI to generate clear and digestible policy summaries. Ultimately, our vision is to enable users to input prompts for policy creation and to explore how AI can elevate Rippling’s policy engine, empowering our customers.

Use cases

Mapping the moments AI could help

Mapping admin journeys

FIG. 07Time AI admin journey and use-case carousel

Concepts

From explanation to action

3 prototypes around Create/Edit • Explain • Check

Intelligent policy creation

Proactive compliance & risk management

Play clip: Transparency & employee empowerment

FIG. 08Three prototypes, one per capability.

Platform alignment

Making the pattern reusable

Contribution to system patterns (entry points, response types, citations)

FIG. 09Shared Copilot anatomyThe patterns Time contributed, and where they now apply.

Panel pattern — FIG. 10Platform design-system patterns (Pebble)

Agent thinking pattern — FIG. 10Platform design-system patterns (Pebble)

Implementation

From framework to product

Implement in Time, Q4 2025; Rippling beta launch Mar 26; measure ticket reduction, setup time, resolution rate.

Time AI Copilot implementation and final view — Play clip: Time AI Copilot

What we measure

Ticket reduction · Setup time · Resolution rate

What I learned

Trust starts with explainability and the “why”
Tie AI to domain objects + citations; don’t ship a generic chat.
Platformizing early keeps AI consistent—and safer—across products.
The strength of AI lies in clarity and action—diagnosing and fixing, surfacing unknowns, and guiding smarter setup

“Adrian shared the Time Product org’s framework for AI, looking at how it will support the creation and understanding of often complex policies. [He] did an awesome job framing the opportunity and showing the concept through a realistic scenario. Great guidance for other teams shaping their AI initiative”

— Ryan Lucas, VP of Design, Jun 2025