AI Efficiency Meets Profound Human Insight
Challenge
Deepen insight gathering working with AI, or wait for this human expertise to become obsolete?
As Large Language Models (LLMs) transform the data analysis landscape, The Strategy Group, a notable strategy and business consultancy with diverse clients in infrastructure, healthcare and government, faced a pivotal challenge. Could they leverage AI to deepen customer insights and retain a competitive edge, or risk falling behind? The question was whether they could harness the speed and power of LLMs to even surpass the quality and depth of traditional methods, while preserving the expertise that defines their brand. This was the pivotal question driving a journey towards their innovation in data synthesis.

Approach
Harnessing AI to Transform Customer Interview Analysis
To address this challenge, I spearheaded the UX development of InsightWise, a bespoke AI tool built on a Large Language Model (LLM).
My contributions involved a multi-faceted approach:
- Collaborating with an AI Engineer, business analysts, and researchers, addressing feasibility parameters and key insights needed from transcripts (current-state analysis and supporting evidence).
- Leading the design process from early ideation to creating high-fidelity interactive prototypes in Figma, while mentoring analysts and junior designers on design tools and methodss.
- Crafting prompts grounded in Natural Language Processing (NLP) best practices to ensure clarity, specificity, and factual grounding.
- Conducting iterative testing with Figma prototypes ranging from low to high fidelity, gathering feedback to enhance the design's effectiveness and prompt accuracy.
- Continually refining prompt engineering approaches to eliminate ambiguity, reduce speculative interpretations, and align with the desired insights.


Outcome
Shows potential to boost customer insight analysis, reveal hidden patterns and even surpass human accuracy, offering a competitive edge
Early Proof-of-Concept demonstrates:
- Reductions in analysis time, outperforming traditional manual methods, unearthing significant operational efficiencies
- Tangible improvements in the factual accuracy of AI-generated insights, narrowing the gap with traditional analysis, with iterative prompt engineering
- Uncovering previously hidden patterns and connections within customer transcripts, revealing nuanced perspectives and potential insights beyond the reach of traditional methods
- The benefit of combining human expertise with AI's speed and precision, with the potential to surpass traditional methods to unlock new frontiers in customer research, creating a competitive advantage for The Strategy Group
AI Beyond the Strategy Group
My passion for AI extends beyond just this project:
- Generative AI Prompt Engineering: Since September 2022, extensive engagement with generative AI, refining prompt engineering skills for complex tasks.
- Practical AI Applications:
- Experiments with a ReactJS app on AWS for streamlined data analysis using Langchain and Embedchain.
- Created Streamlit apps integrating OpenAI APIs and Langchain for enhanced document queries.
- EEG Data Analysis:
- Innovated prompts for AI to assess cognitive load in data over time, contributing to new UX research methodology.
- Transcript Analysis Efficiency:
- Leveraged prompt engineering expertise to speed up transcript analysis, aiding in market research and academic studies.
- AI-Driven Innovations:
- Developed a tailored student marketing system using AI for rubric-based assessments, improving teacher efficiency.
- Implemented AI in ideation sessions, accelerating brainstorming and fostering creativity.
- Utilised AI in Human-Centered Design workshops for instant participant feedback analysis, deepening user understanding.
- Cross-Domain Impact: Demonstrated versatility and proficiency in applying generative AI and NLP, effectively unlocking insights and streamlining workflows across various sectors.