AI powered publishing for financial reports and data visualisation

AXA Investment Managers produces regular research reports that synthesise complex financial data and expert commentary for investors around the world.

These reports are critical communication tools, shaping how market insights are shared internally and externally. However, the existing process relied on a manual PowerPoint workflow that required significant editorial effort to update charts, adjust layouts, and maintain consistency across each edition.

Pattrn was asked to explore how modern digital publishing, structured content systems, and AI could dramatically improve the production workflow and enable a more scalable reporting platform.

The existing GMO reporting workflow created several operational and experience challenges.

Each new edition of the report required manually copying the previous presentation and replacing data, text, and visuals across every slide. Charts were recreated in external tools before being imported back into PowerPoint, making the process both slow and error-prone.

This resulted in several key issues:

• Time-consuming, manual updates for every edition
• Chart recreation across disconnected external tools
• Inconsistent visual and narrative standards
• Barriers to scaling reports across languages and channels

AXA IM needed a more modern approach to producing and distributing research content.

Pattrn structured the project as a technology and workflow exploration, designed to identify the most viable long-term solution for GMO report production.

The work was delivered in four phases:

1. Research

We evaluated a broad landscape of publishing platforms, data visualisation tools, and AI technologies to understand their capabilities within financial research workflows.

2. Assessment

Tools were assessed against the specific needs of the GMO reports, including structured content, chart generation, workflow automation, and compliance requirements.

3. Prototype development

Two prototype solutions were created to test different strategic approaches:

• A custom AI-powered publishing platform
• A hybrid solution using enterprise publishing tools

4. Proof of concept

The preferred custom architecture was developed into a working proof of concept to validate key features such as AI-assisted chart generation, structured content management, automated exports, and multilingual publishing.

Pattrn designed a modular architecture combining AI automation, structured content management and modern data visualisation.

The prototype platform integrated:

• AI automation using Claude for chart recommendations, translation, and content generation
• Headless CMS for structured content authoring and reusable report components
• Modern visualisation libraries for production-ready charts
• Automated export workflows for web, PDF, and PowerPoint outputs

This approach created a dedicated report-production platform tailored to the GMO workflow.

Key capabilities included:

• Structured report authoring
• Automated chart creation from raw data
• AI-generated commentary suggestions
• Multi-language publishing
• Automated application of design system rules
• Dynamic digital report outputs

To compare alternatives, Pattrn also developed a hybrid workflow combining enterprise publishing and data visualisation platforms.

This option allowed teams to create interactive digital reports using reusable templates while embedding visualisations from a specialist charting platform.

The hybrid approach delivered faster time-to-market and strong collaboration workflows but offered less flexibility and long-term extensibility than the custom architecture.