Complete Google AI Services Guide - Gemini, NotebookLM, Jules, Veo, Deep Search, and Stitch

Understanding the Complete Google AI Ecosystem

As of 2026, Google AI has evolved from a simple "chatbot" into an integrated AI platform.

Both Google AI Pro (¥2,900/month) and Google AI Ultra ($249.99/month) include nine services:

• Gemini 3.1 Pro
• NotebookLM
• Jules
• Veo 3 (Pro) / Veo 3.1 (Ultra)
• Deep Search
• Deep Think (Ultra only)
• Project Mariner (Ultra only)
• Stitch
• Flow

Each serves different purposes, and combining them delivers remarkable productivity gains.

1. Gemini 3.1 Pro — Conversational AI Core

Basic Specifications

Real-World Use Cases

Scenario 1: Complex Questions

Question: "Why has the generative AI topic been declining in Japan's 2026 IT trends?"

Gemini's response example:
There are three possible reasons:

1. Transition to maturity phase
   - 2023-2024: Attracted attention as "new technology"
   - 2026: Treated as "practical-stage technology"

2. Facing implementation challenges
   - During enterprise adoption, security, compliance, and ROI measurement issues have surfaced
   - No longer just "technology discussion"

3. Diversification and specialization
   - Interest has shifted from "AI in general" to "AI in specific fields"
     (medical AI, manufacturing AI, etc.)

Scenario 2: Text Analysis and Summarization

Paste large amounts of documents, emails, or news articles and ask for key points.

Scenario 3: Code Questions

# Paste the code below
def calculate_total(items):
    total = 0
    for item in items:
        total = total + item['price'] * item['quantity']
    return total
 
# Question: "Does this code have performance improvement potential?"
 
Gemini's response:
- Use sum() and map() instead of for loop
- Convert to list comprehension
- Add type checking functionality

Getting Started

1. Access Google AI Studio (web): ai.google.dev
2. Sign in with your Google account
3. Start a new chat

The Pro/Ultra difference isn't in cost or context window size, but effectively in request limits.

2. NotebookLM — Unified Knowledge Management and Analysis

What It Can Do

NotebookLM is a tool where you "upload multiple documents and have AI analyze them as a single knowledge base."

Real-World Use Cases

Scenario 1: Project Documentation Management

A new project launches with these documents:

• Requirements (20 pages)
• Specifications (50 pages)
• Design docs (30 pages)
• API spec (15 pages)
• Test cases (50 pages)

Integrate all 165 pages into one NotebookLM:

1. Upload all documents from Google Drive
2. Ask AI in notebook:
   "Give me an overview of this project"

   → AI creates integrated summary from all docs

3. Ask detailed questions:
   "How are API authentication methods and frontend authentication flow defined?"

   → Automatically extracts API specs and shows related sections

This lets new team members get up to speed in minutes instead of hours.

Scenario 2: Academic Research Organization

Consolidate 50 research papers:

Question 1: "What are the mainstream hypotheses in this field?"
Question 2: "What's the difference between hypothesis A and B?"
Question 3: "What's new since 2024?"

AI automatically analyzes across multiple papers,
showing trends and alternative perspectives

Scenario 3: Competitive Analysis

Integrate competitor information:

• Official website pages
• Latest 10 press releases
• Industry reports
• User review collections

Question: "What are the strategic differences between competitors?"
AI automatically analyzes each company's differentiation

Implementation Steps

1. Access NotebookLM: notebooklm.google.com
2. Create new notebook
3. Upload documents, PDFs, YouTube links, etc.
4. After auto-indexing completes, start Q&A

3. Jules — Async Code Review Assistant

What It Does

Jules is a tool where "AI automatically performs code review work asynchronously for development teams."

Via GitHub integration, it automates:

• Automatic PR analysis
• Code quality checks
• Security vulnerability scanning
• Test coverage inspection
• Automatic comments and suggestions

Use Cases

When a New PR is Created

Developer A pushes code for Feature X and creates a PR.

Traditional method:
1. Team lead reviews PR (30 min)
2. Add multiple comments
3. Developer A makes fixes (1 hour)
4. Re-review (15 min)
→ Total: 1.75 hours

Using Jules:
1. Jules auto-analyzes when PR is created
2. Immediately adds comments and suggestions
3. Developer A reviews and fixes (15 min)
4. Team lead does final check only (5 min)
→ Total: 20 minutes

Workspace Integration

Jules deeply integrates with Google Workspace:

• Google Docs co-creation: AI automatically enforces style consistency and grammar
• Sheets data analysis: AI proposes analysis for complex datasets
• Slides improvements: AI auto-generates slide coherence and visual suggestions

4. Veo 3 — Text-to-Video Generation

Pro vs Ultra Differences

Real Usage Examples

Marketing Video Production

Prompt:
"Create an intro video for new product 'SmartWatch Pro'.
Content:
- Simple, elegant design
- User wearing on wrist
- Real-time health metrics display
- Stylish black and silver
- About 1 minute"

Veo 3 output:
- High-quality 720p video
- Product 3D rendering
- Natural backgrounds
- Photorealistic quality

Use this video across YouTube ads, SNS, and sales pitches immediately.

5. Deep Search — AI-Powered Automatic Research

Differences from Traditional Search

Real Usage Examples

Topic: "2026 No-Code Development Tool Trends"

Deep Search process:
1. Crawl multiple tech blogs
2. Check GitHub trends
3. Collect industry reports
4. Compare multiple platforms
5. Create integrated report

Output:
- Market size growth rates
- Major players (Bubble, FlutterFlow, Retool, etc.)
- Use cases
- 2026 predictions
- Detailed source list

Research that traditionally takes 2-3 hours completes in 5 minutes with Deep Search.

6. Deep Think (Ultra Only) — Deep Reasoning for Complex Problems

What It Does

Deep Think has AI think through complex problems step-by-step, not just give quick answers.

Example Usage

Problem: "How will white-collar salaries change in Japan 2026-2030 with generative AI?"

Deep Think process:

1. Problem decomposition
   - What factors affect salaries?
   - Variables in AI adoption speed
   - Japan-specific labor market factors

2. Multiple hypothesis evaluation
   - Hypothesis A: Salary decline (AI replaces simple tasks)
   - Hypothesis B: Salary increase (productivity gains = higher skill demand)
   - Hypothesis C: Polarization (high-skill up, low-skill down)

3. Evidence evaluation
   - Historical tech innovation examples
   - Current corporate adoption status
   - Labor law constraints

4. Synthesis
   - Most likely scenario
   - Risk factors
   - Countermeasures

Regular Gemini gives shallower answers. Deep Think prioritizes thinking depth.

7. Project Mariner (Ultra Only) — Browser Automation

What It Can Do

Project Mariner "automatically executes complex web tasks."

Can run up to 10 simultaneous tasks, automating:

• Form filling and submission
• Website crawling and data collection
• Screenshot capture
• Button clicking and navigation
• Data extraction

Real Usage Example

Task: "Collect company information from 100 companies"

Traditional method:
1. Create company list
2. Visit each company website one by one
3. Record company overview, location, employee count
4. 100 companies × 30 min each = 50 hours

Using Project Mariner:
1. Upload company list (spreadsheet)
2. Instruct: "Extract overview, location, employee count from each site"
3. AI automatically processes 10 companies in parallel
4. Complete in about 2 hours
→ Time saved: 96%

8. Stitch — Integration of Multiple Google AI Services

What It Does

Stitch is an API framework that programmatically integrates multiple Google AI services (Gemini, NotebookLM, Deep Search, etc.).

Developers can automate complex workflows like:

Example workflow: "Fully automated market analysis"

1. Deep Search collects latest competitor info
2. NotebookLM consolidates information
3. Gemini generates analysis and recommendations
4. Veo turns analysis into video report
5. Auto-populates Google Sheets

All runs automatically daily

Implementation Example (Python)

from google.ai import stitch
 
# Define Stitch workflow
workflow = stitch.Workflow(
    name="market-analysis",
    steps=[
        stitch.DeepSearch(query="Competitive analysis 2026"),
        stitch.NotebookLM(name="Competitor info"),
        stitch.Gemini(prompt="Provide analysis recommendations"),
        stitch.Veo(prompt="Turn analysis into video"),
    ]
)
 
# Execute
result = workflow.execute()

9. Flow — No-Code AI Workflow Construction

Google Sheets Automation Example

With Flow, build AI-powered automation workflows within Google Sheets.

Example: Sales management sheet

Columns: Customer name, sales amount, purchase pattern, AI recommended action

Flow configuration:
- Sales > ¥1M → Ask Gemini for "sales strategy proposals"
- Auto-classify customers into segments from purchase pattern
- New customer added → AI auto-creates customer profile

Result: Sales team only does "data input + confirm AI suggestions"

Feature Availability by Plan

Great News for Japanese Students

Google AI Pro is offered completely free for 1 year to Japanese university students.

Eligible: Students at accredited Japanese universities Duration: 1 year Benefits: All Pro features unlimited

Students can register now and actually use multiple AI tools throughout their university years, significantly boosting competitiveness after graduation.

Looking back: The Power of Google AI Ecosystem

Google AI's nine services aren't just "convenient tools"—they're an integrated productivity platform:

• Research: Deep Search
• Knowledge management: NotebookLM
• Writing and analysis: Gemini
• Video production: Veo 3
• Development support: Jules
• Complex reasoning: Deep Think
• Web automation: Project Mariner
• Integration API: Stitch
• Automated workflows: Flow

Combining these definitely expands individual capability.

I recommend starting with the free version, then considering upgrading to Pro (¥2,900/month).

In Japan especially, Google AI Pro is reasonably priced, and getting access to 9 services for ¥2,900/month is an excellent investment.

Experience the power of the Google AI ecosystem.