Plan Mode

Plan Mode is the conversation you have with atypica.AI before a study runs. You describe what you want in plain language, the system figures out what kind of study fits, and shows you the full plan. You confirm or modify, then execution starts.

It's the layer between "I want to know X" and "running 8 interviews with Tier 2 personas for 2 days."

What the plan actually shows you

A plan has four parts, and you can read each one before confirming:

Study type. Which of the four modes your study runs in:
- User Research — qualitative research (interviews, discussions, observation). The default for most research questions.
- Fast Insight — turn a topic into a podcast and brief in a few hours. No interactive personas.
- Product R&D — market trend and product opportunity research.
- Panel Only — build a reusable persona panel without running a study.
Methodology framework. The lens your study will use — JTBD, KANO, STP, user journey, etc. The framework shapes what questions get asked and how the output is structured.
Method mix. Which research methods to combine: Interview Chat (1-on-1 deep dive), Discussion Chat (3–8 people, ideas collide), Scout Agent (social-media observation), Fast Insight. Often more than one.
Persona pool. Which personas to use, how many, and any filters (age range, region, gender skew, etc.).

If you're not familiar with a framework or method name in the plan, ask the model what it means before confirming.

How long it takes

There's no fixed duration. A clear brief ("competitive teardown of Notion AI") might produce a plan in one or two exchanges. A vague goal ("understand our users") might involve several rounds of clarifying questions — age range? region? what kind of understanding are you after?

The plan is generated and persisted either way. You can always come back to it.

Skipping plan confirmation

For routine studies or when you've specified everything explicitly, you can skip the confirmation step and have the plan accepted automatically. This is for users who already know what they want and don't want the round-trip of approving every plan.

The plan is still generated and saved — you're just skipping the click.

When Plan Mode is the right tool

Use it when:

Your research goal is fuzzy or you've never done this kind of study before.
Multiple methodologies could fit and you want to see the tradeoff.
You're working with a team — the plan is a checkpoint for alignment.
The study is expensive (high-tier personas, long interviews) and you want a sanity check before committing.

Skip it when:

You've specified the exact studyType, framework, and method already.
It's a routine study you've done before.
You're an experienced user and the dialogue would just be overhead.

Modifying the plan

Once shown, you can ask the model to change:

The study type ("actually I want Fast Insight, not full research")
The persona filters ("narrow to 28–32 year old women in tier-1 cities")
The method mix ("drop the interviews, just run a discussion")
The participant count
The framework

You can't ask for methods atypica.AI doesn't have. If you want a survey-based quant study, the system will redirect you — it's not the right fit.

What happens after you confirm

Once accepted, the plan is locked in. The session moves into execution and the study runs the chosen methods against the chosen personas. You can't change the plan mid-flight — modifying scope means starting a new session.

If you realize partway through that the framework or persona filter was wrong, finish the current study, then start a new session with the corrected plan.

Limits

All four study types are qualitative-leaning. You can't get a 1,000-respondent quant survey through Plan Mode.
The plan is the model's current best guess. Read it before confirming.
Once execution starts, the plan is fixed.

A worked example

You say: "I'm launching a sparkling water product aimed at young office workers. I want to know if the positioning should be 'health' or 'social'."

After a few clarifying questions (age range? region? how many personas?), the plan comes back:

Study type: User Research
Framework: STP
Methods: Scout → Discussion → Interview
Steps:

Scout watches Xiaohongshu and Douyin for sparkling water discussions

Discussion with 8 personas on 3 positioning directions

Interview 5 high-intent personas to validate the chosen positioning

Generate positioning report
Personas: 8 personas, age 25–32, female-skewed, tier-1 cities

You might adjust: "I want to focus on 28–32 year olds, not 25–32." The plan updates, you confirm, execution starts.