2 months | August 23 - October 23
Leveraging Gen-Ai to improve productivity for users
User Research - Discovery - Prototyping - Proof of Concept - Concept Sprint
Project subject to a Non-Disclosure Agreement (NDA) - all information has been sanitized.Helping a global information services company generate quicker and higher-quality insights about consumer behaviour from a large structured dataset.
Today the process is highly manual and time-consuming, with analyses not incorporating the full breadth of data available.
The objective of the 6-week POC was to build a GenAI-enabled chat interface, allowing users to ask common and time-consuming queries, and receive text & visual responses
CONTEXT
IMPACT
〰️
IMPACT 〰️
~30-50%
Productivity uplift from deploying in internal workflows to answer customer queries
~10-20%
Revenue uplift from rolling out as a new product to customers to directly usequeries
Challenges
Building trust in a GenAI tool
Developing a new tool, especially with Generative AI, requires establishing robust guardrails to prevent inaccuracies and ensure that the generated responses are relevant and accurate
Establishing a new way of working among employees
A new tool has the potential to enhance the productivity of existing employees. However, it's essential to ensure that we're developing a tool to support them rather than replace them in the future. Involving end-users from the outset is crucial for shaping the Proof of Concept (POC) strategy together with them.
Approach
Over a 6-week period, we developed a POC where we
1. Identified users’ 20 most common and time-consuming queries
and core GenAI capabilities required to answer these (e.g., intent identification, data visualisation)
2. Mapped the workflow
and pain points with client end-users, which informed the POC features, design vision and potential impact
3. Developed and prototyped
ideal concepts to stress test with users. Prepared ‘best-practice’ sample answers and glossaries working closely with client SMEs
4. Refine concepts and build
in parallel a POC journey, developed by the technical team. Rapidly deployed a re-usable solution on internal platform for quick iterative testing & experimentation
5. Pivot based on new needs identified
and developed a new ‘storytelling mode’ to compile best-practice frameworks and core-analyses to answer ‘strategic’ questions
Results
Launch a Proof of Concept (POC) within 8 weeks by collaborating with a diverse team comprising data scientists, data engineers, business analysts, and designers:
Empower users through a seamless end-to-end experience
Single source of truth leveraging an aggregated database
Quick and accurate generation of charts and insights
Ready-to-access library of outputs with the ability of easily collaborate
Empower clients to pitch the POC to main stakeholders, igniting interest and securing funding
Your copilot
contact us
Want to hear more? Let’s talk!