noelog37@gmail.comLinkedIn ↗

Adoption research · Kroger / Cooklist

Meal Assistant

Two studies, one adoption story — Kroger / Cooklist

Year
2026
My role
Design Research Director
Method
Moderated Wizard-of-Oz + unmoderated study
Partner
Cooklist, Inc.
moderated Wizard-of-Oz interviews, 1:1 Zoom
N=5
studies triangulated: diagnostic + coverage
2

Kroger/Cooklist Meal Assistant — Two studies, one adoption story

How trust, realism, and plan-to-cart reliability shape weekly meal planning behaviors.

The proof of concept

A short demo of the Kroger/Cooklist Meal Assistant, the concept explored in the two studies below.

Why two studies

One study goes deep on why behaviors happen (diagnostic), the other goes wide on where they break across the end-to-end flow (coverage). Together they triangulate adoption barriers and product requirements.

Moderated Wizard-of-Oz interviews

Wizard-of-Oz is a prototyping method where users interact with a ‘live’ assistant, but a researcher simulates the system behind the scenes to validate the experience before engineering is complete.

Method: 60-min 1:1 Zoom (Wizard-of-Oz simulation)

  1. Sample: N=5
  2. Goal: compare baseline vs AI-assisted flow for speed, confidence, and trust

Moderated Wizard-of-Oz interviews (60 min, Zoom). I ran 1:1 sessions using a structured discussion guide to keep each interview consistent while still leaving room for open-ended probing. Participants worked through a realistic weekly-meal-planning scenario, first describing how they’d normally decide what to cook, then interacting with a “live” Meal Assistant. Behind the scenes, I simulated assistant behavior in real time (Wizard-of-Oz) using prepared response patterns to test the experience before full automation. This approach let us evaluate trust, clarity, and perceived effort—especially what people needed to feel confident the plan would work for their household.

Themes from the Wizard

Next case studyCALI: Competitive Agent Longitudinal Intelligence