Frontiers Media | Developing a communications AI suite 

Researcher Personas GPT

About:
I developed the Researcher Persona GPT — an interactive AI trained on behavioral and communication data from my previous research project. Designed to simulate authentic academic responses across career stages (early-career, mid-stage, senior), it enabled our marketing and CX teams to test, refine, and adapt their tone for greater relevance and empathy.

Needed:
I wanted to create a scalable, conversation-based way to test and validate communications with distinct researcher types. The goal was to understand what matters most to each audience segment, ensure empathy-driven outreach, and align with brand messaging principles.

Work included:

Built prompt architecture that allows for tone analysis and rewording by persona.

Led user testing with around 30 internal specialists across four pilot journals.

Collected and analyzed both qualitative and quantitative feedback on clarity, tone, and motivation.

Iterated based on measurable success metrics — operational efficiency, response rate, and user confidence.

Outcomes:

Delivered actionable insights into how different researcher segments respond to messaging.

Enabled dynamic A/B testing of tone and calls-to-action.

Significantly increased confidence in audience targeting and personalized communication across campaigns.

---

Brand Guardian GPT

About:
I created the Brand Guardian GPT — an AI-powered content review assistant trained on Frontiers’ tone, glossary, and metrics index. It functions as a live editorial companion, helping teams keep communications consistent, compliant, and unmistakably on-brand.

Needed:
I identified the need for a fast, reliable way for teams to check copy for tone, compliance, and voice before publication — without slowing down production. The system needed to centralize brand reinforcement, integrate with evolving frameworks, and be optimized for large language model compatibility.

Work included:
Rewrote, reformatted, and restructured our brand guidelines into a machine-readable framework.

Embedded tone-of-voice detection and suggestion systems to guide clarity and confidence.

Piloted “email checker” workflows using real marketing communications.

Built feedback loops to continuously refine tone detection and expand the glossary and metrics index.

Outcomes:
Reduced manual brand compliance reviews and improved publishing speed.

Strengthened message consistency across global teams.

Established a scalable model for AI-assisted brand governance.

---

UX Writing Assistant GPT

About:
I developed the UX Review Assistant to help product teams craft and refine communications — from email campaigns to journal promotions. Acting as both editor and collaborator, it supports the UX design team in maintaining precision, empathy, and accessibility across all copy, while remaining in the Frontiers brand.

Needed:
The objective was to build an intuitive AI tool that enhances writing quality and tone alignment without diminishing human creativity. It also needed to support non-writers producing on-brand, high-quality content at scale, fully compatible with our UX and CX workflows.

Work included:
Designed GPT logic capable of repurposing content based on objectives, format, and audience.

Integrated the assistant with testing frameworks to measure clarity, engagement, and response.

Partnered with researchers, developers, and QA teams for iterative tuning.

Analyzed pilot feedback to refine tone sensitivity, phrasing, and contextual adaptability.

Outcomes:
Streamlined content workflows and shortened review cycles.

Boosted team confidence in writing while reducing dependency on editorial oversight.