Testing and Improving AI Research Agents for Offshore Wind Data – AgentZero Energy Ltd

Company Information:

We use artificial intelligence, guided by experts, to collect and analyse energy market intelligence. We sell our data and analysis to the energy industry. They use it to make better strategic decisions.

Working Pattern

Remote working with weekly in-person catch-ups at Orbis Energy, Lowestoft, and regular online supervision.

Project Overview

How Good Is Your AI? Testing Research Agents on Offshore Wind

AgentZero uses AI agents to automatically research and build databases tracking major energy projects worldwide. Companies rely on our data for big investment decisions. We currently cover subsea cables and specialist ships; offshore wind is next. This project is valuable because the intern will configure our AI agents to gather wind farm data, monitor their accuracy, and correct errors. This directly builds a product we sell while testing our automation on a new domain – proving our AI platform scales beyond its current datasets.

Project Aim

To build and validate an offshore wind project dataset using AI research agents, while assessing their accuracy and identifying where human oversight is required to improve data quality.

Key Research Questions:

• How effectively can AI agents collect offshore wind project data from public sources?

• Which data sources provide the most reliable information on offshore wind projects?

• Where do AI agents perform well, and where do they make errors?

• What types of errors occur most frequently?

• How can AI agent performance and data accuracy be improved?

What You Will Do

• Research offshore wind projects using public databases, reports, and news articles

• Configure and run AI research agents to extract offshore wind project data

• Monitor AI output by comparing extracted data with original sources

• Identify, flag, and correct errors within the dataset

• Track AI agent performance and accuracy

• Build a verified offshore wind database covering at least 50 projects

• Rank and catalogue the best public data sources for offshore wind research

• Analyse results and develop recommendations to improve AI accuracy

• If appropriate, present findings to the AgentZero team

Science, Engineering and Technical Understanding

• Understanding how artificial intelligence systems collect and process data

• Learning about data pipelines, automation, and quality control

• Applying analytical thinking to assess accuracy and performance

• Exploring offshore wind as an energy technology and industry

• Understanding the balance between automation and human judgement

Planning and Organisation

• Plan and manage your work independently

• Keep clear records of data sources and verification checks

• Organise datasets and performance notes systematically

• Use evidence to justify conclusions and recommendations

• Take part in regular check-ins to review progress

Skills Developed

• Research and data analysis

• Problem solving and critical thinking

• Independent working and self-motivation

• Communication of technical findings

• Understanding AI systems and energy markets

Final Outputs

• A populated offshore wind database (50+ projects)

• A ranked source catalogue of public offshore wind data sources

• A performance report assessing AI accuracy and limitations

• Clear recommendations for improving AI research agents

• A presentation summarising findings and lessons learned

Reflection and Evaluation

Your report will include:

• What you learned about offshore wind and AI research systems

• Examples of where AI performed well and where it struggled

• How human oversight improved data quality

• How confident you are in the final dataset and why

• What you would improve if the project continued

https://www.crestawards.org/resources/crest-gold-student-guide