Streamlining Financial Services Content Production with AI Agents
We built two AI agents for a global asset manager: one to improve its SEO and one to keep its charts on brand.
Optimizing financial services content production with a global asset manager
Ext. Marketing built two AI agents for this client’s content team, which publishes a high volume of white papers, blog posts and market commentaries. The firm was already using Microsoft Copilot, and requested we leverage the same AI tool for this initiative.
The challenge: applying firm standards to every article and chart
The team used its AI tool for one-off prompts, so articles and charts still needed manual optimization and review before they could go live.
The firm’s brand, style and search engine optimization (SEO) standards lived in static PDFs that AI could not read. There was also no central IT support for team-level AI projects. The team needed AI that followed its own rules without adding a new tool, vendor or login.
Our solution: two AI agents built on the firm’s existing tools
In eight weeks, we built an agent for each job inside the firm’s Microsoft 365 environment, so the team could keep using the tools it already had. Every output is approved by the team before it goes live.
Our SEO and AEO agent reviews a finished article against the firm’s SEO standards and returns
on-page recommendations: page titles, meta descriptions, subheads, internal links and answer engine optimization (AEO) readiness. Our data visualization agent checks a chart against a machine-readable style guide, corrects labels, colour use and scale, and writes a short narrative
to go with the commentary.
Before
- Articles and charts reviewed by hand.
- Standards stored in static PDFs that AI could not read.
- AI used only for one-off prompts.
- No central IT support for team-level AI projects.
What We Built
- Firm standards rewritten so AI can read and apply them.
- SEO and AEO agent reviews finished articles and returns on-page recommendations.
- Data visualization agent brings charts in line with the style guide and writes a narrative.
- A team member approves outputs before they go live.
Results
- Two AI agents, each built for one repeatable content task.
- Zero new platforms, vendors or logins for IT to approve.
- 100% of client data stays inside the firm’s environment.
- Eight weeks from discovery to a coached, self-sufficient team.
