AI Features
Overview of HypLab's AI capabilities -- template generation, people generation, persona creation, and more.
AI Features
HypLab uses AI agents to generate templates, create realistic synthetic people, and build personas automatically. This page gives you an overview of what the AI can do and how it works.
What the AI does
Template generation
When you complete workspace onboarding, the AI:
- Analyzes your onboarding answers (company context, industry, audience, use case).
- Generates 5 dynamic templates tailored to your business.
- Selects relevant people fields and persona fields for each template based on category (B2B or B2C).
The result is a set of ready-to-use templates that match your context. For example, a B2B SaaS company might get templates like "Enterprise CTO", "Startup Founder", and "IT Procurement Lead".
People generation
When you generate people from a template, the AI:
- Uses the template's people fields to determine which fields to populate.
- Generates realistic, coherent person records -- names match demographics, companies match industries, behavioral data is consistent.
- Marks the people as "synthetic" so you can distinguish them from manual or real data.
You can generate 1-10 people at a time.
Persona creation
When you generate people, the AI also:
- Checks if a persona already exists for that template.
- If not, creates a new persona with an AI-suggested name based on the template and generated people.
- If a persona exists, adds the new people to it.
- Populates persona fields (buyer profile, triggers, obstacles, etc.) using the template's persona field list.
Persona from text
The Persona Generator agent can create a complete persona from a text description -- useful when you already know who you want to represent.
Onboarding enrichment
The Enrichment agent analyzes your onboarding data to provide industry insights, competitive positioning, and recommendations. This feeds into template generation.
AI agents architecture
HypLab's AI capabilities are powered by separate Python services called agents:
| Agent | Port | Purpose |
|---|---|---|
| Enrichment | 4098 | Analyzes onboarding data for industry insights |
| Synthetic | 4099 | Generates realistic people and persona data |
| Templates | 4100 | Generates templates from onboarding context |
| Persona Generator | 4101 | Creates complete personas from text descriptions |
Agents communicate with the main API via WebSocket, which allows real-time progress streaming during generation.
Supported LLM providers
HypLab uses LiteLLM for multi-model support. You can use:
- Google Gemini
- OpenAI (GPT-4, GPT-3.5)
- Ollama (local models)
- Any OpenAI-compatible API
See LLM Setup for configuration details.
Tips for better AI results
- Be specific during onboarding. The more detail you provide about your company, industry, and audience, the better the generated templates will be.
- Review before using. Always check generated people and personas for quality. AI output is a starting point, not a final product.
- Iterate. Generate small batches (3-5 people), review the results, and adjust templates as needed.
- Create custom templates. After seeing what the AI generates, create your own templates with exactly the fields you need.
Next: Settings and Customization -- customize fields, preferences, and workspace settings.