ChallengeUpdated on 23 December 2025
Automatic Generation of Live Micro-Segments Using Generative AI and Predictive Algorithms
Open Innovation Sr Analyst at REPSOL
Móstoles, Spain
About
-
Challenge Context: We aim to evolve from traditional segmentation models (based on static attributes: age, gender, location, purchase history) towards a dynamic, continuous, and self-learning system that automatically identifies emerging micro-segments within the customer base, each with its own needs, behaviours, and motivations.
The goal is to achieve a capability not yet mature in the market: detecting small groups (10–50 users) based on latent patterns invisible to human analysis and enabling activation of messages, creatives, and offers generated by generative AI for each group.
This challenge is part of the company’s strategy to incorporate advanced AI to personalise and optimise marketing actions in real time, increasing perceived relevance for the user and improving conversion -
General objective of the challenge: Identify live and dynamic micro-segments from first- and third-party data. (To be confirmed during the configuration).
Generate understandable descriptions of these micro-segments (insights, motivations, frictions).
Propose and/or automatically generate messages, creatives, and offers for each micro-segment.
Track the performance of each micro-segment and readjust the model as they evolve or disappear.
Enable marketing teams to activate micro-segments in digital channels easily. -
Functional Scope: Includes data ingestion and unification, automatic detection of micro-segments, automatic description, generation of messages and creatives, continuous evaluation, and channel activation.
-
**Technical Requirements:**Architecture and data model for large volumes (>500,000 users), AI models (transformers, clustering, generative LLMs), explainability, security and privacy (GDPR compliance).
-
Project Deliverables: Technical architecture document, operational model, dashboard, API connectors, message generation system, final validation report.
-
Proposal Evaluation Criteria: Technical capability, experience, creativity, security, scalability.
-
KPIs Required to Validate the Challenge: Discovery of at least 5 relevant micro-segments, 10–15% conversion increase, ability to update every 30 days, creative generation <30s, message coherence >90%.
-
Expected Proposal Format: Detailed technical document (max 20 pages), demo or video, methodology description, risk mitigation, detailed budget.
Topic
- Artificial Intelligence
- Audiovisual
- Multimedia
- Content creation
Type
- Proof of concept/pilot testing
- Investment (Venture capital and corporate venturing)
Organisation
Similar opportunities
Challenge
Active Listening and Campaign & Bid Optimisation
- Multimedia
- Audiovisual
- Content creation
- Artificial Intelligence
- Text analysis and tagging
- Image analysis and tagging
- Proof of concept/pilot testing
- Investment (Venture capital and corporate venturing)
LAURA GARCIA DIEGO
Open Innovation Sr Analyst at REPSOL
Móstoles, Spain
Challenge
Hyper-personalisation Strategy for Businesses
- Machine Learning
- Content creation
- Artificial Intelligence
- Natural Language Processing
- Design / User interface / UX
- Proof of concept/pilot testing
Jordi López Almazán
Head of Innovation at Banc Sabadell
Barcelona, Spain
Challenge
Innovation in Design & Product Development
- Artificial Intelligence
- Text analysis and tagging
- Image analysis and tagging
- Natural Language Processing
- Proof of concept/pilot testing
- Systems Integration & Advanced Manufacturing
Beatriz Lario
Technology Innovation Specialist at DESIGUAL
Barcelona, Spain