AI NativePersonas in AI Native are not job titles or demographic segments. They are defined by need-state: what the buyer needs the product to do for them. This distinction matters because AI assistant questions are shaped by what people want to achieve, not by what they are called.
What a persona contains
Each persona has four required fields:
Label. A short name used throughout the UI, such as "Value seeker" or "First-time buyer."
Need-state. A plain description of what the persona wants the product to do. For example: "wants the best price and terms" or "new to the category, needs guidance."
Business weight. A number between 0 and 1 that represents how commercially important this persona is relative to others. A weight of 1.0 means highest priority. Business weight feeds into the opportunity score calculation, so gaps in high-weight personas produce higher-ranked opportunities.
Profile. A structured description of the persona's goals, challenges, mindset, and buyer journey. This is generated during onboarding (by the LLM for live onboarding, deterministically for demo) and is editable. It drives the persona-stage prompt tree and provides context for the AI generation features.
The 6-stage journey
Each persona's profile includes a buyer journey. The journey has stages matching the five funnel stages in the platform (awareness, consideration, comparison, decision, retention) plus any additional context stages. Each journey stage records the buyer's mindset, representative unbranded questions, and representative branded questions.
You can expand the journey for any persona by clicking the route icon on the persona card. The stages display in order with the question examples from the profile.
The journey is not directly editable from the personas page but it is used by the prompt generation feature. When you generate prompts with steering, the LLM uses the journey context to write questions that match that stage's real intent.
How personas and the heatmap connect
The persona-stage matrix (the heatmap on the Matrix tab) is built by crossing each persona against each funnel stage. Every cell in the matrix corresponds to a set of prompts assigned to that persona at that stage. The recommendation rate in each cell is the measured outcome for that combination.
High-weight personas in weak cells produce the highest opportunity scores. The Focus tag on a persona card marks the persona with the highest-value gap across the product.
Persona performance metrics
Each persona card on the personas page shows three metrics from the latest scan:
- Recommendation rate for that persona's unbranded prompts
- Mention rate for that persona's unbranded prompts
- Average sentiment score for that persona's answers
Below the metrics, the weakest stage for that persona is identified with its recommendation rate and a link to filter the answers view for that specific persona and stage.
Adding a persona manually
Click "Add" in the page header to expand the add form. Enter the label, need-state, and business weight. The persona is created with no journey profile and no prompts. Add prompts to it from the prompt library, or use AI generation with a steering instruction that references the new persona.
Generating personas with AI
Click "Generate with AI" to expand the generation form. You can optionally provide a steering instruction and choose how many personas to generate (2 to 5). The LLM uses your product's positioning, existing competitor context, and the current persona set to draft new personas. Newly generated personas include a full journey profile.
Editing a persona
Hover over a persona card and click the pencil icon to edit the label, need-state, and business weight inline. Save with the Save button. Editing a persona does not change the prompts already assigned to it.
Deleting a persona
The delete icon on a persona card removes the persona. Prompts assigned to that persona are not deleted but become unassigned. They are still visible in the prompt tree.
Related docs
Questions
Why are personas defined by need-state instead of job title?
Buyers search for what they need a product to do, not for a description of themselves. A question like "which loan has no prepayment penalty?" comes from a need, not a job title. Need-state personas produce prompts that match real search and AI-assistant question patterns more accurately than demographic segments.
What is business weight and how should I set it?
Business weight represents the commercial value of a persona to your product, on a 0 to 1 scale. Set higher weights for personas who represent your largest revenue segments or highest lifetime value. The weight multiplies into the opportunity score, so gaps in high-weight personas get prioritised in the action queue.
Can I have more than four personas?
Yes. The platform does not enforce a hard limit. However, each persona adds to the prompt tree, which increases scan cost and time. Most products work well with four to six personas that cover distinct need-states without overlapping.
What is the focus persona?
The focus label marks the persona with the highest-value opportunity gap in the current scan. It is determined automatically from the matrix data. You cannot set it manually.
Does deleting a persona delete its prompts?
No. Prompts assigned to the deleted persona remain in the database but become unassigned. They are still counted in the prompt tree and can be reassigned from the import flow or by editing the persona assignment.
How does the journey profile affect scanning?
The journey profile does not directly control what questions are asked in a scan. That is determined by the prompt tree. The journey is used as context when generating new prompts with AI, to make the drafted questions match the realistic intent and mindset of the persona at each stage.
Can I import personas from a spreadsheet?
There is no direct persona import. Use the manual add form for each persona, or use AI generation with a steering instruction that describes the persona type you want. Persona generation produces fully profiled personas with journey stages.