Research

Google Gemini

5 desktop + 4 mobile screenshots

Desktop (5)

Mobile (4)

Funnel Overview

Google Gemini — Funnel Overview

Funnel Summary

  • Total steps: 2-4 (Google account already exists for most users)
  • Funnel type: Ecosystem-embedded PLG with contextual onboarding
  • Time to complete: Under 1 minute (for existing Google account holders)
  • Data collected: Leverages existing Google account data (name, email, preferences). Additional: Search history, Gmail content, Docs activity, Photos (with permission for personalization)
  • Payment timing: Free tier available. Gemini Advanced at $19.99/month (included with Google One AI Premium). No credit card required for basic use.
  • Personalization level: Heavy — passive personalization through connected Google apps (Search, Gmail, Docs, Photos, YouTube). Active personalization through Gems (custom personas) and preference settings.

Funnel Flow

Step 1: Discovery / Entry Point

Multiple entry points — this is Gemini's key differentiator:

  • Direct: gemini.google.com
  • Gmail: "Help me respond" side panel
  • Google Docs: "Help me brainstorm" sidebar
  • Google Sheets: Formula generation assistant
  • Chrome: Browser integration
  • Android: Power button hold or "Hey Google" voice command
  • Google Search: AI-enhanced search results

Most users discover Gemini through contextual prompts in tools they already use, not through a standalone signup flow.

Step 2: Terms Acceptance

For first-time Gemini users: brief terms of service acceptance. No new account creation needed — Google account serves as authentication.

Step 3: Initial Interaction

User lands on chat interface with suggested prompts based on context (if entering from Gmail, suggestions relate to email; from Docs, suggestions relate to writing). "Google It" verification button available for fact-checking responses.

Step 4: Personalization (Ongoing)

Personalization deepens over time:

  • Connect Search history for tailored responses
  • Gemini Memory stores preferences across sessions (Advanced tier)
  • Create custom Gems (personas) for specific tasks
  • Adjust response style and depth preferences

What Works Well

1. Zero-Friction Entry Through Existing Accounts

Most users already have Google accounts, eliminating signup entirely. This is the most frictionless "funnel" in the AI space — there's literally no funnel. Users go from discovery to product use in one click.

2. Contextual Discovery in Existing Workflows

Side panels in Gmail, Docs, and Sheets introduce Gemini where users already work. The AI appears as a helpful assistant within familiar interfaces, not as a separate product requiring adoption. This "ambient AI" approach has the lowest cognitive barrier of any onboarding.

3. Passive Personalization Through Connected Data

Rather than asking users 20 questions about preferences, Gemini learns from existing Google data (with permission). Search history, email patterns, and document activity create a rich user profile without any explicit onboarding effort.

4. "Google It" Verification Button

Unique trust signal — users can instantly verify AI responses against web search results. This addresses the #1 concern with AI tools (accuracy) through a familiar, trusted mechanism.

5. Suggested Follow-Up Questions

Instead of tooltips or tutorials, Gemini guides users by suggesting related questions after each response. This teaches capabilities through natural conversation flow rather than formal instruction.

What Could Be Better

1. Fragmented Experience Across Entry Points

Users may be confused about Gemini vs. Google AI vs. Gemini Advanced vs. Copilot Chat vs. individual Google app AI features. The proliferation of entry points can create brand confusion.

2. No Structured Onboarding for First-Time AI Users

Users new to AI assistants get no structured guidance. The "just start typing" approach works for tech-savvy users but may leave AI newcomers unsure how to begin or what's possible.

3. Personalization Requires Trust

Connecting Search history, Gmail, and other Google data requires significant trust. Privacy-conscious users may resist personalization, reducing the product's effectiveness.

4. Advanced Features Hidden

Gems, Deep Research mode, and other advanced features are not prominently introduced. Users may never discover capabilities that could drive upgrade conversion to Advanced tier.

5. No Clear Upgrade Path Visibility

The transition from free Gemini to Gemini Advanced isn't prominently marketed within the product experience. Users may not know what they're missing.

Key Psychological Principles Used

Familiarity / Mere Exposure Effect

By appearing within tools users already use daily (Gmail, Docs, Chrome), Gemini benefits from the comfort and trust associated with familiar interfaces. Users are more likely to engage with AI that feels like a natural extension of their existing tools.

Default Effect / Path of Least Resistance

Google can make Gemini the default assistant across its ecosystem. Users encounter AI capabilities without actively seeking them — the path of least resistance leads to Gemini use.

Anchoring Through Free Access

The free tier creates an anchor — users experience significant value at no cost, making the $19.99/month Advanced tier feel like a reasonable upgrade rather than a new expense.

Progressive Disclosure via Conversations

Rather than front-loading features, Gemini reveals capabilities through suggested follow-up questions. Users naturally discover advanced features through conversation, creating organic "aha" moments.

Relevance to Twofold

Direct Transfers

  1. Contextual Entry Points: Twofold should explore EHR integration — appearing as a sidebar or assistant within the tools clinicians already use, rather than requiring a separate app launch.

  2. Passive Personalization: Rather than asking clinicians extensive setup questions, Twofold should learn from early sessions — note format preferences, terminology patterns, documentation style — and progressively customize.

  3. Suggested Actions: After generating a note, Twofold could suggest related actions: "Would you like to also generate a treatment plan?" / "Want to create a patient summary?" This teaches capabilities through natural workflow.

  4. Verification Mechanisms: Like Gemini's "Google It" button, Twofold should make clinical accuracy verification easy — perhaps a "Review clinical details" button that highlights key findings for clinician confirmation.

Adaptations Needed

  1. Standalone Product Required: Unlike Gemini, Twofold can't rely on existing EHR accounts for authentication. However, EHR integration partnerships could reduce the "new product" friction.

  2. Explicit Onboarding Needed: Clinicians need structured guidance on recording best practices, template selection, and compliance features. A "just start typing" approach would miss critical clinical workflow education.