Designing UX for a Dynamic Pricing Tool with AI Automation

Project Overview

This project investigates how AI-driven dynamic pricing tools can support small business owners in making informed pricing decisions. Given the challenges of fluctuating supplier costs, intuition-driven practices, and limited technical expertise, this study focuses on designing interfaces that balance automation, explainability, and user control.

A within-subjects A/B testing design was adopted to evaluate two interactive prototypes of a pricing dashboard, measuring both usability (SUS) and trust in automation (Madsen & Gregor, 2000), alongside qualitative feedback from semi-structured interviews.

Prototype Development

Prototype 1 (Control Group) – Streamlined Automation

Focused on usability and competitor comparison tables.
Provided automated price suggestions with minimal explanation.

Prototype 2 (Treatment Group) – Explainable AI & Scenario Simulation

Incorporated Explainable AI (XAI) elements for price reasoning.
Introduced scenario simulation (e.g., demand fluctuations, seasonal offers).
Integrated visual graphs to reduce cognitive load (Sweller, 1988).
Applied Dual-Coding Theory (Paivio, 1990) by combining textual and visual cues.

Project Results Analysis

Inferential Testing

Paired-sample t-test showed statistically significant improvement in usability and trust between Prototype 1 and 2.

Thematic Analysis

Shopkeepers valued automation, but demanded transparency in decision-making.
Trust increased when explanations matched their intuitive pricing logic.
Overloaded or technical explanations reduced adoption likelihood.
Users preferred visual clarity and competitor context over purely data-driven results.

Project Conclusion

For User Experience Design

Transparency in AI-driven tools is crucial for adoption.
Explainable AI (XAI) features enhance trust but must remain digestible and context-matched.
Scenario simulations empower users by giving them a sense of control and foresight.

For User Experience Research

Balancing automation vs. control is a core challenge in AI design.
Future research can explore long-term adoption and trust evolution.

Key Findings

Impact on UX Design

Visual explainability drives adoption – Shopkeepers placed the highest trust in pricing suggestions when supported by clear, visual explanations (competitor comparisons, sales trend graphs).
Trust improves with aligned logic – Participants trusted AI suggestions more when the reasoning mirrored their own intuitive pricing strategies, highlighting the importance of contextualised explanations.
Cognitive overload reduces usability – Overly technical or excessive information reduced trust and hindered usability, reinforcing the need for digestible, well-structured content.
Scenario simulations empower control – Features such as demand fluctuation or seasonal pricing scenarios gave participants a sense of agency and foresight, strengthening confidence in the tool.
Automation vs. control tension – Shopkeepers welcomed automation but wanted the option to override AI decisions, showing that hybrid human–AI decision-making fosters trust.

Transparency as a design principle – Dynamic pricing tools must embed Explainable AI (XAI) in ways that are accessible to non-technical users.
Balancing automation with user agency – Providing manual overrides and scenario simulations increases trust and adoption.
Reducing cognitive load in AI systems – Applying Cognitive Load Theory ensures users process complex data without overwhelm, supporting usability.
Visual-first communication – Graphs, trend lines, and competitor comparisons were more effective than text-heavy explanations, highlighting the power of visual storytelling in UX design.

Ankit Kumar Sinha