25.8.2024

Building interactive personas with Volkswagen Group

25.8.2024

Building interactive personas with Volkswagen Group

25.8.2024

Building interactive personas with Volkswagen Group

VW brands used various independent AI tools for customer insights, leading to inconsistent data and higher costs.

Lakmoos AI personas embedded into VW’s Sedcards, providing instant, representative responses for key segments.

Early testing showed the centralized tool could streamline brand insights and reduce redundancies, while maintaining high data accuracy and compliance.

Deep dive

Overview

Company

Volkswagen Group (VW)

Time

Q3 2024

Timeline

September: Pilot setup and configuration

October: Persona development and integration into VW Sedcards

November: Testing, showcase, followed by pilot assessment and potential broader rollout

Main goal

Explore integration of Lakmoos AI personas as a scalable, central insight tool for consistent, high-quality data across VW brands.

People involved

Skoda X: project management

Volkwagen: integration and frontend

Lakmoos AI: backend and API

100% cooperation happened online.

Who participated

Heroes of the project

We save money, time and effort

Pilot Solution

We save money, time and effort

Pilot Solution

We save money, time and effort

Pilot Solution

"Whoever say corporates are slow to adopt innovation is simply wrong. This team managed to organise and implement a possible group-wide solution in two months from the first conversation. This shows exceptional ability to absorb innovation, which is in the current climate of the race of implementing AI for better productivity absolutely key."

Kamila, Lakmoos

1

Persona Interaction via Chat

Users could query personas, representing target customer segments in e-mobility within the Czech market.

2

Instant Feedback on Group Fit

Each answer included a “group fit” metric to ensure responses accurately represented the intended segment.

3

Data Compliance

With no additional datasets shared, Lakmoos operated fully within VW’s compliance guidelines.

The pilot utilized Lakmoos AI personas configured to represent key VW segments defined earlier in 2024. Embedded in VW Sedcards, these personas were designed to generate consistent, accurate, and timely customer insights.

“The AI chat is working, and as always, working with you has been a lot of fun. And we were fast, very fast. Light speed, or better Chuck Norris speed.”

Klaus, Global Brand Strategy, Volkswagen AG

We save money, time and effort

Expected results

We save money, time and effort

Expected results

We save money, time and effort

Expected results

1

Cost Savings

Centralized data through personas should reduce the need for multiple independent AI tools, with initial findings indicating reduced costs for brand-specific insights.

2

Time Efficiency

Pilot users should gain immediate responses, bypassing traditional data gathering and analysis delays.

3

Stakeholder Buy-In

A test group of VW stakeholders reported improved access to customer data and insights, demonstrating the tool’s strategic potential.

Pilot Findings - What Worked

Key Benefits Observed

Pilot Findings - What Worked

Key Benefits Observed

Pilot Findings - What Worked

Key Benefits Observed

Centralized Insights

Provided a unified, scalable platform, facilitating cross-brand data aggregation.

We validated answers against real-world data.

Simultaneously, we validated our baseline dataset with AI department.

Also, users often asked the model a question to which they already knew answers from datasets that were not shared with us.

We validated answers against real-world data.

Simultaneously, we validated our baseline dataset with AI department.

Also, users often asked the model a question to which they already knew answers from datasets that were not shared with us.

We validated answers against real-world data.

Simultaneously, we validated our baseline dataset with AI department.

Also, users often asked the model a question to which they already knew answers from datasets that were not shared with us.

Accurate and Fast

AI personas generated quantifiable responses that could be used in both qualitative and quantitative research contexts.

Yes, you can.

We only ran into difficulty with sharing 20 login details via email. Flagged as spam and reformatted, we will be smarter next time.

Yes, you can.

We only ran into difficulty with sharing 20 login details via email. Flagged as spam and reformatted, we will be smarter next time.

Yes, you can.

We only ran into difficulty with sharing 20 login details via email. Flagged as spam and reformatted, we will be smarter next time.

Cost-Effective

Lower resource demands compared to traditional customer insight collection methods.

Lakmoos distributed 21 licenses, 17 of them (81%) were activated and 14 of them (82%) were actively using the app.

Within 31 days, users submitted 243 sensible questions.

Users tested not only the model but also our patience and submitted 48 nonsensical questions, such as “How do I ask a girl out?” or “What was first: chicken or egg?”. Surprise surprise, a model trained for banking products did not know the answers. We are not ChatGPT.

Lakmoos distributed 21 licenses, 17 of them (81%) were activated and 14 of them (82%) were actively using the app.

Within 31 days, users submitted 243 sensible questions.

Users tested not only the model but also our patience and submitted 48 nonsensical questions, such as “How do I ask a girl out?” or “What was first: chicken or egg?”. Surprise surprise, a model trained for banking products did not know the answers. We are not ChatGPT.

Lakmoos distributed 21 licenses, 17 of them (81%) were activated and 14 of them (82%) were actively using the app.

Within 31 days, users submitted 243 sensible questions.

Users tested not only the model but also our patience and submitted 48 nonsensical questions, such as “How do I ask a girl out?” or “What was first: chicken or egg?”. Surprise surprise, a model trained for banking products did not know the answers. We are not ChatGPT.

Pilot Findings - What’s Next

Areas for Development

Pilot Findings - What’s Next

Areas for Development

Pilot Findings - What’s Next

Areas for Development

-

Expanded Persona Library

To support broader market insights, the persona library will need to grow to encompass more regions and topics.

-

Data Integration

Ensuring ongoing compatibility with VW’s ecosystem while maintaining strict data protection.

Future potential

Future potential

Future potential

Based on the pilot’s success, Volkswagen Group plans to explore a phased expansion across more segments and regions. This includes adding personas for other countries and strategic topics like autonomous driving. A broader rollout would allow the Volkswagen Group to further unify its customer insights, optimizing decision-making processes across all brands.

Get in touch

Want to learn more about our Pilot programs?

Write us any time.

Get in touch

Want to learn more about our Pilot programs?

Write us any time.

Get in touch

Want to learn more about our Pilot programs?

Write us any time.

Get in touch

Want to learn more about our Pilot programs?

Write us any time.

Honest answers on a 20 % budget. Lakmoos answers surveys with data models instead of real people. We aim to replace 20 % of traditional surveys with real-time insights by 2030, saving $30 Bn in research costs and 35 Bn hours of fieldwork globally each year.

Quick contact

Příkop 843/4

Brno 60200

VAT CZ19395108

Lakmoos AI s.r.o. 

Copyright © 2024 Lakmoos. All rights reserved.

Honest answers on a 20 % budget. Lakmoos answers surveys with data models instead of real people. We aim to replace 20 % of traditional surveys with real-time insights by 2030, saving $30 Bn in research costs and 35 Bn hours of fieldwork globally each year.

Quick contact

Příkop 843/4

Brno 60200

VAT CZ19395108

Lakmoos AI s.r.o. 

Copyright © 2024 Lakmoos. All rights reserved.

Honest answers on a 20 % budget. Lakmoos answers surveys with data models instead of real people. We aim to replace 20 % of traditional surveys with real-time insights by 2030, saving $30 Bn in research costs and 35 Bn hours of fieldwork globally each year.

Quick contact

Příkop 843/4

Brno 60200

VAT CZ19395108

Lakmoos AI s.r.o. 

Copyright © 2024 Lakmoos. All rights reserved.

Honest answers on a 20 % budget. Lakmoos answers surveys with data models instead of real people. We aim to replace 20 % of traditional surveys with real-time insights by 2030, saving $30 Bn in research costs and 35 Bn hours of fieldwork globally each year.

Quick contact

Příkop 843/4

Brno 60200

VAT CZ19395108

Lakmoos AI s.r.o. 

Copyright © 2024 Lakmoos. All rights reserved.