Get in touch

PROJECT OVERVIEW

Chrome Extension for Academic Research

[ BROWSER EXTENSION MODIFICATION FOR ACADEMIC RESEARCH ]

Extending an open-source browser extension into a research-grade instrument for capturing how users interact with an AI shopping assistant.

Assembly Requirements

MODULE INVENTORY (5 COMPONENTS)

A university research team needed to study how real participants interact with Rufus, Amazon's AI shopping assistant, in the wild. Rather than build a measurement tool from scratch, they came to us with Webmunk — an existing open-source Chrome extension used in web research — and asked us to extend it to capture Rufus-specific interactions at research grade.

We shipped a customized Webmunk build that listens for Rufus events, normalizes them into a typed schema, and streams the data through Rudderstack into the team's analytics stack.

Primary focus

Extend a trusted research extension rather than reinvent one

Capture Rufus-specific interaction events end-to-end

Ship clean, structured data to the team's analytics pipeline

What we delivered

Customized Webmunk build with Rufus-aware content scripts

Research-ready event schema flowing through Rudderstack

Packaged extension ready for participantdistribution

STEP 02 — CHALLENGES IDENTIFIED

Assembly Requirements

MODULE INVENTORY (5 COMPONENTS)

REQUIRED CAPABILITIES

Tracking a Moving Target

Rufus is an actively evolving feature inside a production Amazon UI. The tracking layer had to resolve the right elements reliably across DOM changes, survive Amazon's in-flight updates, and degrade gracefully when selectors drifted — without requiring a re-deploy mid-study.

Research-Grade Event Fidelity

Casual web analytics tolerates lost events. Academic research does not. Every interaction needed to be captured once, in order, with accurate timestamps and consistent participant identifiers the research team could trust across the full study window.

Scoped, Respectful Data Capture

A university study carries strict constraints on what's collected, where it travels, and how long it lives. The extension had to stay scoped to the study by default — no drift, no over-collection, no surprises surfacing downstream in the data pipeline.

STEP 03 — ASSEMBLY PROCESS

EXTENSION CUSTOMIZATION

SOLUTION APPROACH

MODULE 3.1

● Core Instrumentation

We forked Webmunk and added a Rufus-aware content script that hooks into the shopping assistant's DOM lifecycle, watching for message exchanges, product recommendations, and user actions.

Each captured interaction is shaped into a typed event and streamed through Rudderstack to the research team's warehouse — giving them a clean, queryable dataset from day one rather than a pile of raw logs to parse later.

Workflow diagram showing data flow from Rufus UI through content script generating typed events, streamed to Rudderstack, and stored in a warehouse.

MODULE 3.2

● Delivery & Handoff

We packaged the extension for the research team to distribute to study participants, including a sanity-check mode that let the researchers validate tracking against their own interactions before going live.

Flowchart showing extension package leading to sanity check, then participant rollout to N participants.

STEP 04 — ASSEMBLY COMPLETE

Delivered Impact

MODULE INVENTORY (5 COMPONENTS)

Event-level
capture
fidelity

IMPLEMENTATION

Typed schema. Precise timestamps.
Drift-resistant selectors.

Clean data,
research-
ready

RESEARCH CAPABILITY

Rudderstack transport.
Queryable warehouse.

Packaged
for
rollout

DELIVERY

Self-serve install. Sanity-check mode.

PERFORMANCE OUTCOMES

Research team received a working data pipeline on day one — not a pile of raw logs to parse

Tracking layer survived Amazon's in-flight UI updates across the full study window

Study ran on the planned Feb–Sep 2025 timeline without instrumentation blockers

ASSEMBLY PHILOSOPHY

We don't just code.
We build systems.
Piece by piece,
designed to snap perfectly
into your vision.

Start building now

Ready to start your project?

Chrome Extension for Academic Research

Extending an open-source browser extension into a research-grade instrument for capturing how users interact with an AI shopping assistant.

Assembly Requirements

Assembly Requirements

EXTENSION CUSTOMIZATION

Delivered Impact