# AlphaFold Brand Voice

> Academic, precise, and utilitarian, prioritizing scientific utility over marketing flair.

## Positioning
AlphaFold is an AI-driven protein structure database developed by Google DeepMind and EMBL-EBI. It serves the global scientific and academic community by providing high-accuracy structural predictions to facilitate biological research and functional discovery.

## Voice principles
*   **Academic:** Uses formal, peer-reviewed language and prioritizes proper citation and attribution.
*   **Transparent:** Openly discusses limitations, confidence levels, and the specific methodologies used to generate data.
*   **Functional:** Focuses on utility and "how-to" instructions, using direct verbs to guide researchers through the database.
*   **Collaborative:** Emphasizes the partnership between AI developers and biological institutes, framing the tool as a contribution to the scientific ecosystem.

## Tone by context
| Context | Tone |
|---|---|
| Technical Documentation | Precise and instructional. Uses specific terminology like "homodimeric" and "PAE". |
| News & Updates | Informative and milestone-oriented. Focuses on the scale and impact of new data releases. |
| Legal & Licensing | Formal and protective. Strict adherence to copyright, attribution, and citation requirements. |
| Troubleshooting (FAQs) | Helpful and systematic. Offers logical steps and alternative search methods. |

## Lexicon
*   **Use:** Facilitates, predictions, high-confidence, attribution, proteome-wide, functional and mechanistic discoveries, open source, academic and commercial use.
*   **Avoid:** Revolutionary (uses "top-ranked" or "accurate" instead), magic, instant, easy (uses "facilitates" or "reliable" instead).

## Messaging do's and don'ts
*   **Do:** Cite specific papers and researchers when discussing data origins.
*   **Do:** Use quantitative descriptors (e.g., "~2.2M homodimeric structures") to show scale.
*   **Do:** Provide clear instructions for search failures (e.g., "Try searching by protein or gene name").
*   **Don't:** Make absolute claims about accuracy without mentioning "limitations" or "confidence".
*   **Don't:** Use hyperbolic marketing adjectives; stick to the data and its potential for "biological research".
*   **Don't:** Refer to the AI as a sentient entity; it is an "AI system" or a "method".

## Evidence
*   "AlphaFold was the top-ranked protein structure prediction method by a large margin."
*   "This release facilitates functional and mechanistic discoveries across biology."
*   "If your use case isn't covered by the database, you can generate your own..."
*   "Sequence search is the most reliable method."
*   "EMBL-EBI expects attribution (e.g., in publications, services, or products)."