# Paradigm Brand Voice

> Authoritative, precise, and action-oriented, focusing on scale, accuracy, and human control.

## Positioning
Paradigm is an AI platform that automates complex research workflows for decision makers and analysts. It replaces scattered, manual research with scalable, auditable "agent swarms" that verify data across multiple sources while leaving the final decisions to the user.

## Voice principles
* **Authoritative:** Speaks with absolute confidence about capabilities, using strong, definitive statements (e.g., "Accuracy as the baseline").
* **Action-oriented:** Leads with strong verbs to emphasize productivity and momentum (Scale, Run, Save, Encode, Pull, Define).
* **Precise:** Focuses on exactness, verification, and auditability rather than vague promises.
* **Direct:** Clearly contrasts the flaws of the old way (liability, breaks at scale) with the power of the new way.

## Tone by context
| Context | Tone |
|---|---|
| Marketing hero | Bold and visionary, focusing on superhuman capabilities and scale. |
| Feature descriptions | Instructional and clear, explaining exactly how the agents operate. |
| Trust and Security | Reassuring and definitive, focusing on safeguards, cross-referencing, and auditability. |
| UI/Status states | Extremely brief and functional (e.g., "Processed in Paradigm", "Report generated"). |

## Lexicon
- **Use:** Agentic research, workflows, scale, verify, audit, agent swarm, cross-reference, auditable, hallucination safeguards, decision makers.
- **Avoid:** Chatbot, magic, guess, maybe, autopilot (the brand emphasizes "you make the decisions").

## Messaging do's and don'ts
**Do:**
* Emphasize human control and oversight (e.g., "Agents do the work, you make the decisions").
* Highlight accuracy, verification, and the ability to audit sources.
* Use concrete numbers to illustrate scale (e.g., "5 rows or 50,000", "10,000+ decision makers").
* Frame manual research as a "competitive liability".

**Don't:**
* Overpromise AI autonomy without mentioning human oversight.
* Use fluffy or overly casual language.
* Obscure how the AI gets its answers (always highlight that it is "auditable by design").

## Evidence
* "Scale agentic research with superhuman a[ccuracy]"
* "Manual research is a competitive liabili[ty]"
* "Agents do the work, you make the decisio[ns]"
* "Run at any scale, 5 rows or 50,000"
* "Encode your best thinking once and run i[t forever]"
* "Define your workflow and run agent swarm[s]"
* "Accuracy as the baseline"
* "Safeguards against hallucination"
* "If the answer isn't findable, Paradigm l[eaves it blank]"