Video summary
Bài 2: Xây dựng thương hiệu dễ nhớ cùng Gemini
Main summary
Key takeaways
Overview
This video is a training livestream (“Gemini Academy”, part of the “Breakthrough with AI” series) focused on using Google’s AI tools—especially Gemini—to help small/medium businesses build a memorable brand. It includes hands-on demos, brand-building workflows, and guidance on how to write effective AI prompts/statements.
Key technological concepts & tool features
1) Accessing and using Gemini (and related tools)
Participants are guided to access Gemini via:
- Phone: Apple Store / Google Play
- Laptop/desktop:
jami.google.com(and other Google links)
A recurring theme is “statement-based thinking”—how you instruct the AI.
2) Extracting images + using prompts for personalized outputs
A demo shows:
- Taking a photo (phone/computer)
- Using a prompt to generate an image/story based on extracted details
- Iterative editing, such as changing descriptive text and regenerating outputs
3) “Gemini Live” + interactive assistant behavior
Gemini is emphasized as capable of:
- Supporting multimedia tasks (text + interpretation)
- Providing chat-based interactive help
- Sharing/exporting outputs (e.g., captions/subtitles, share links)
4) NotebookLM (research + summarization workflow)
NotebookLM is presented as a research/grounding tool that uses user-approved sources (documents/links). It differs from Gemini’s general assistant style:
- Gemini may browse more freely
- NotebookLM focuses on materials you provide
Demos include:
- Uploading files or adding Drive links
- Summarizing text into structured outputs
- Generating presentations, mind maps, podcasts/audio, slides/infographics, videos
5) Prompt engineering: “Statements” vs “Commands”
A major portion teaches how to structure prompts for better brand results:
- Command (basic) vs high-quality command/statement (better)
-
A repeated framework of 5 key factors:
- Role
- Mission / goal
- Task breakdown
- Target audience
- Style + format (tone, structure, output format)
It also teaches a “looping” workflow:
- Draft → test → edit → repeat until satisfied
6) Brand building with Gemini: end-to-end pipeline
The course demonstrates a workflow to produce brand assets:
- Brand naming
- Logo + slogan creation
- Brand story / storytelling
- Brand personality definition
- Brand positioning via environment scanning
- Brand guidelines / consistency system (“Canvas”)
- Content planning and internal training materials
Product/feature walkthroughs in the brand workflow
A) Naming (idea assistant / game-like workflow)
Participants use an “idea assistant” interface to generate multiple options matching constraints:
- Short, meaningful, easy pronunciation/spelling
- Uniqueness and domain availability considerations
B) Logo creation with “Nano” (Nano Banana referenced)
Participants provide:
- Role (e.g., designer)
- Context (e.g., “logo for brand X”)
- Style constraints (e.g., minimalism, vintage/nostalgia, Vietnamese character)
- Color specification and symbol concepts
The demo emphasizes iterative refinement, such as:
- Regenerating with adjusted size/position/colors (e.g., lotus flower too big → shrink/shift)
It also stresses specificity with domain terms, such as:
- logo/icon/vector, brand name inclusion, design style keywords
C) Slogan creation with “brand expert” + evaluation
Gemini is used to:
- Generate multiple slogan candidates
- Then evaluate/select the best one
The demo includes scoring logic (e.g., “good idea” vs “needs revision”) and improvements for:
- Grammar
- Tone
D) Brand story creation (300–500 words) + polishing
Story requirements include:
- Authenticity and consistency with brand personality
- Inclusion of target audience
- Emotional/empathetic framing
Outputs can be repurposed into:
- Voice/audio readings
- Mind maps
- Video/podcast scripts
- Slides/infographics
E) Turning brand assets into multimedia content with NotebookLM
NotebookLM generates:
- Mind maps, videos, podcasts/audio, slides, infographics
- Support for editing and changing design colors/styles
Demos show exporting/embedding content for presentations and training.
F) Brand personality definition + consistency across touchpoints
Participants define brand personality using archetype-like traits (examples mentioned):
- Friendly, cheerful, modern, energetic, caring
Once defined, all outputs should match it across:
- Posts and responses
- Visuals
- Tone of voice
- Language consistency
G) Brand positioning via camera-based “live” scanning
A feature is demonstrated where participants:
- Record a space with a camera around the setting
- Receive Gemini feedback such as brand positioning/atmosphere description
The goal is to reduce manual effort (no need to type long descriptions).
H) “Canvas” as a structured brand guideline document
A Canvas concept compiles:
- Brand strategy foundation
- Story insertions
- Logo/slogan/positioning
- Guidelines for consistent output
Canvas examples include:
- Brand story formatted for social posts
- Email templates
- Landing page-like content suggestions
- “Internal training kit” plans for staff
Review/tutoring emphasis (what the lesson explicitly teaches)
The tutorial repeatedly focuses on:
- How to write effective prompts (role/mission/task/audience/style/format)
- How to run an iterative prompt loop until outputs meet requirements
- How to ground outputs using NotebookLM with reliable sources
- How to produce a full brand system: name, logo, slogan, story, personality, guidelines, content plan, internal training materials
- How to keep brand consistency operationally by training staff and using standardized materials
Main speakers / sources (as mentioned in subtitles)
- Diễm (main trainer/host throughout)
- Ms. Thi Le (head; opening remarks on NIC/Google collaboration)
- Ms. Pham Tran Hong Diem (lecturer/companion name appears in subtitles)
- Nguyễn Thi Lệ Quyên (credited in the scholarship/acknowledgements section)
- Brandon Thanh Do (Google SEA & Australia market development; closing Google AI/search/ads perspective)
- Google / National Innovation Center (NIC) (collaborating organizations)