Designed a gamified experience that encouraged AI engagement and personalization.

This is a trust-first AI learning module designed to solve the cold-start problem in AI-powered estimating for field service businesses. Instead of attempting full automation upfront, the system focuses on visibly learning a business’s unwritten rules—pricing instincts, labor heuristics, material preferences, and operational patterns. It also serves as a "Knowledge Capture" tool to institutionalize tribal knowledge.

UX DESIGN
PRODUCT DESIGN
GAMIFICATION

Designed a gamified experience that encouraged AI engagement and personalization.

This is a trust-first AI learning module designed to solve the cold-start problem in AI-powered estimating for field service businesses. Instead of attempting full automation upfront, the system focuses on visibly learning a business’s unwritten rules pricing instincts, labor heuristics, material preferences, and operational patterns. It also serves as a "Knowledge Capture" tool to institutionalize tribal knowledge.


Visuals by Radfan Naushad

UX DESIGN
PRODUCT DESIGN
GAMIFICATION

Intent

Field service businesses operate largely on tribal knowledge:

  • Pricing logic lives in the owner’s or senior technician’s head

  • Cost and margin adjustments happen informally after jobs

  • Estimating rules are rarely documented or standardized

As a result, most AI estimators fail not due to lack of raw data, but due to lack of business-specific learning signals. This leads to:

  • Low trust in AI-generated estimates

  • High rates of manual edits

  • Poor adoption caused by opaque, “black-box” behavior

Compounding this is the AI cold-start problem: without sufficient historical data, early recommendations are weak, which further erodes confidence and discourages engagement.

Field service businesses operate largely on tribal knowledge:

  • Pricing logic lives in the owner’s or senior technician’s head

  • Cost and margin adjustments happen informally after jobs

  • Estimating rules are rarely documented or standardized

As a result, most AI estimators fail not due to lack of raw data, but due to lack of business-specific learning signals. This leads to:

  • Low trust in AI-generated estimates

  • High rates of manual edits

  • Poor adoption caused by opaque, “black-box” behavior

Compounding this is the AI cold-start problem: without sufficient historical data, early recommendations are weak, which further erodes confidence and discourages engagement.

/The Company

About Swivl

Swivl Tech is an AI-powered field service management platform designed for home service businesses such as HVAC, plumbing, electrical, roofing, landscaping, and cleaning companies. It helps businesses streamline their daily operations by combining customer management, scheduling, dispatching, job tracking, estimating, invoicing, and payments into a single platform. AI integrated across its workflows including tools for automated customer support, estimate generation, and operational assistance Swivl aims to simplify complex field service operations, improve team productivity, and help service businesses deliver faster and more efficient customer experiences.

Field service businesses operate largely on tribal knowledge:

  • Pricing logic lives in the owner’s or senior technician’s head

  • Cost and margin adjustments happen informally after jobs

  • Estimating rules are rarely documented or standardized

As a result, most AI estimators fail not due to lack of raw data, but due to lack of business-specific learning signals. This leads to:

  • Low trust in AI-generated estimates

  • High rates of manual edits

  • Poor adoption caused by opaque, “black-box” behavior

Compounding this is the AI cold-start problem: without sufficient historical data, early recommendations are weak, which further erodes confidence and discourages engagement.

/Overview of the Product

/Overview of the Product

What does Swivl do?

/About the Feature

/About the Feature

Max Training Center

Max Training Center

Most B2B products are designed to support complex workflows, so the design language is usually direct, information-dense, and functional. Visuals and gamification are often associated with B2C products, where they're used to improve onboarding, build habits, and keep users engaged.

What made designing the MAX Training Center interesting for me was bringing those engagement principles into a B2B environment. The challenge wasn't to gamify for the sake of it, but to help businesses understand the value of building their AI knowledge base. As users trained the system, they could see how their data was turning into a business asset capturing years of operational knowledge, identifying patterns, and helping the AI make better recommendations. The Training Center became more than an onboarding experience; it acted as a companion throughout the product, guiding users, rewarding progress, and helping them understand how different pieces of knowledge connect to create a smarter system.

/Objective & Scope

/Objective & Scope

Build trust by making AI learning transparent and easy to understand.

  • Encourage users to contribute structured, high-quality business knowledge.

  • Create a continuous feedback loop that improves AI recommendations.

  • Reduce repetitive manual work by capturing organizational knowledge.

  • Establish the foundation for intelligent automation and predictive workflows.

A guided training center for onboarding and knowledge creation.

  • AI training using data from customer estimates, jobs, customer invoices, and supplier invoices.

  • A gamified progression system with achievement levels (Helper → Apprentice → Journeyman → Master).

  • Contextual insights, learning recommendations, and automation opportunities.

  • Explainable rule creation, allowing users to understand how AI reaches decisions.

  • AI-generated draft templates to accelerate setup and reduce manual effort.

Gamification & Progression System

The progression system represents learning maturity, not feature usage or vanity metrics. Each level reflects the confidence and breadth of learned business logic, and directly governs what the AI is allowed to suggest or automate. Progression is cumulative, transparent, and irreversible.

/Levels, XPs and Capabilities

Helper

0 (Default)

Observation only; no AI output

Apprentice

5,000

Insights and contextual nudges

Journeyman

25,000

Insights, advanced suggestions, contextual nudges

Master

75,000

Insights, advanced suggestions, contextual nudges, automated drafts

/Rewards and Nudges

These Rewards are on the basis of activities, levels and XPs.

/What value does the center create for User

Level Progression Reward: Insights & Contextual Nudges

Insights - Proven patterns observed in ≥80% of relevant cases within the same category. The Insights can be from 5 to 50 in number for every sync. The broader classification of Rules and Non-Rules.

  • Guardrails (Rules)
    These should be:

    • Actionable and enforceable by the system

    • Binary or threshold-driven

    • Preventive or corrective

    • Clearly defined logic (if-this-then-that)

  • Insights (Advisory)
    These should be:

    • Non-enforceable

    • Contextual and exploratory

    • Helpful for decision-making

    • May include trends, suggestions, or observations

Designed a gamified experience that encouraged AI engagement and personalization.

This is a trust-first AI learning module designed to solve the cold-start problem in AI-powered estimating for field service businesses. Instead of attempting full automation upfront, the system focuses on visibly learning a business’s unwritten rules—pricing instincts, labor heuristics, material preferences, and operational patterns. It also serves as a "Knowledge Capture" tool to institutionalize tribal knowledge.

UX DESIGN
PRODUCT DESIGN
GAMIFICATION

Thank you, Keep in touch :]

IN

16:07

2026
Purva's's iPod
All Songs
Thank you, Keep in touch :]
IN

16:07

2026
Purva's's iPod
All Songs
Thank you, Keep in touch :]
IN

16:07

2026