🗂️ AI Email Categorization · AI Auto-Categorize + Manual Email Categorization

Hundreds or thousands of emails in your inbox — how do you spot the real inquiries at a glance? Laifaxin has two ways to categorize:

Category	Who categorizes	Purpose
🤖 AI Auto Categorization	System itself, no action needed	Auto-sort into 5 categories (Inquiry / System / Ad / Reject / Risk)
🏷️ Your Custom Categories	You (pick a name + a color)	Segment by your own dimensions: project / customer / urgency

System entry: Email

---

I. 🤖 AI Categorization · Zero-Config Auto-Identification

All emails arriving in the inbox auto-sort into 5 AI categories, no configuration needed — sales reps can tell at a glance which are real inquiries and which are spam promos.

1. 5 AI Categories

Category	English	Color	Typical Email	Recommendation
🎯 Inquiry	Inquiry	🟣 Purple	Real business inquiries / customer questions / partnership intent	Top priority, reply immediately
⚙️ System	System	⚫ Gray	Bounce notices (Mail Delivery) / IT system notices	Check the reason, adjust sending strategy
📢 Ad	AD	🔵 Blue	Third-party promo / marketing emails	Bulk delete or tag as "Ad"
🚫 Reject	Reject	⚪ Gray	Customer unsubscribe / refusal emails	Add to blacklist to avoid future contact
⚠️ Risk	Risk	🔴 Red	Phishing / scam / abnormal emails	Don't click links, handle carefully

💡 3 Tricks for Efficiency

1. Check "Inquiry" first — real business inquiries concentrate here, faster response = higher conversion
2. Check "System" for bounce reasons — bounces aggregated, judge mailbox quality + adjust templates
3. Block "Risk" directly — phishing / scam emails, don't click links + add to blacklist

2. Inquiry AI Hint · Tells You Why It's Classified as an Inquiry

When AI identifies a real inquiry, a green-highlighted banner appears at the top of the email, telling you the AI's basis for judgment:

✨ [Real Inquiry] The party precisely responded to our CTA, replied "1" to get the customer credit, classified as business inquiry.

This transparent judgment lets you understand at a glance why AI classified this as an inquiry — not a black-box guess, but based on the customer's precise response to your CTA (Call-to-Action).

3. Correction Mechanism · AI Gets Smarter with Use

AI isn't 100% perfect — it occasionally misjudges. Click the 【Correct】 button on the right of the banner to give the right category back to the system. AI learns in the background, and next time the same type of email is identified more accurately.

⚠️ Saw AI misclassify? Please report

• Real ad classified as inquiry → click 【Correct】 → pick "Ad"
• Real inquiry classified as ad → click 【Correct】 → pick "Inquiry"

Each correction makes AI smarter, benefiting all users.

---

II. 🏷️ Manual Colored Categories · Custom by Business Dimension

AI categorization covers general scenarios, but you may want to segment by business dimensions (by project / customer / urgency etc.) — for that, use manual colored categories.

Imagine using it this way

• 🎯 VIP Customers — Separate category for key customer emails
• 📦 Orders / Contracts — Group order confirmations and contract files
• 💰 Payment Notifications — Centralize payment / invoice emails
• 🎪 Trade Fair Customers — Post-fair follow-up customer emails
• ⏰ This Week Urgent — To-do emails that must be handled this week

---

III. Create and Manage Manual Categories

1. Create a New Category Tag

Steps:

• 1️⃣ Enter create mode: On the email list page, click the 【+】 button next to the 【Categories】 area in the top-left
• 2️⃣ Configure category: Enter category name (e.g. "VIP Customers") + pick an eye-catching color → Save

2. Modify or Delete Existing Categories

Click the small gear ⚙️ icon next to the category tag, then pick 【Edit】 or 【Delete】 from the dropdown.

💡 Impact of editing / deleting categories

• Edit name / color → all linked email tags auto-sync update
• Delete category → only removes the "tag" itself, emails are not deleted

---

IV. Apply Categories to Emails

1. Apply Existing Category

Open an email → in the action area above the email click 【Category】 → pick an existing category tag.

2. Create New Category and Apply Immediately

While processing email, find a need for a new category? Just:

• Above the email, 【Category】 action area → pick 【New Category and Tag】
• Follow Create New Category flow → new category is immediately applied to the current email

3. View Categorization Result

After an email is linked to a category, the email list shows the colored tag — clear at a glance:

---

📋 Pitfalls Summary

Pitfall	Consequence	How to avoid
🔴 Phishing email in "Inquiry" not detected	Clicked malicious link / downloaded virus attachment / leaked financial info	AI is not antivirus — for "Inquiry", still check sender domain + whether content is on-topic; verify suspicious links / attachments by reverse-checking identity
🔴 Not correcting AI mistakes	AI keeps learning the wrong way, all future similar emails misclassified	When you see a misclassification click【Correct】 immediately → pick the right category
🔴 Clicking risk email links	Phishing site steals credentials / virus download	For emails AI marks "Risk" never click links, first verify the sender
🟡 Thinking emails are gone when deleting category	Mistakenly believing important emails are lost	Deleting a category only deletes the tag, the email is still in its original folder
🟡 Mixing AI categorization with manual categorization	An email has multiple category tags, confusing	AI auto-identification vs. user-built are two parallel systems, they don't conflict
🟡 Building a pile of manual categories never used	Too many tags becomes a management burden	Start with 3 - 5 core categories, expand as needed

---

❓ FAQ

Q1 · What's the difference between AI categorization and manual categorization?

Dimension	🤖 AI Categorization	🏷️ Manual Colored Categories
Who builds	System auto-identifies	User builds manually
Categorization standard	Fixed 5 categories (Inquiry / System / Ad / Reject / Risk)	User defines
Config needed?	❌ 0 config, out-of-the-box	✅ Manual build + link needed
Best for	General: judge email nature	Business segmentation: project / customer / urgency
Can coexist?	✅ Yes, an email can show both types of tag simultaneously	✅

Q2 · Can AI mistakes be corrected?

✅ Yes. Open the email → click the 【Correct】 button on the right of the banner → pick the right category → submit. AI learns in the background, more accurate next time.

Q3 · How many manual categories can I create?

Create as needed, we recommend starting with 3 - 5 core categories (like "VIP Customer / Order / Urgent / Complaint"). Specific upper limit depends on Laifaxin plan, usually enough for daily use.

Q4 · What's the impact of editing / deleting a category on linked emails?

Operation	Email itself	Email tag display
Edit name / color	Unchanged	Auto-sync update
Delete category	Unchanged (email not deleted)	Email no longer shows that tag, still in original folder

Q5 · What is the "CTA" mentioned in the Inquiry AI hint?

CTA = Call-to-Action, the phrasing in your cold email designed to make the customer take a specific action. Example:

• ❌ Vague: Let me know if you're interested.
• ✅ Clear: Reply 1 to get our latest catalog. (→ customer reply "1" = real intent)

When AI detects the customer precisely responds to your CTA (such as replying "1" or clicking the link you provided), it classifies as a real inquiry.

Q6 · Can phishing emails appear in "Inquiry"? How to identify?

⚠️ Yes. AI mainly judges from semantics ("looks like an inquiry"), but some phishers mimic inquiry phrasing. 3 ways to identify:

Check	Real Inquiry	Phishing
Sender mailbox	Real company domain (e.g. john@acme.com)	Free mailbox / spoofed domain (e.g. john@acme-corp.gmail.com)
Content match	Mentions your specific product + quantity + spec	Vague "interested" + nudges you to click a link / download an attachment
Attachments / links	PDF quote sheet / real order screenshot	.exe / .zip / short URL / bank or cloud-drive spoof

When you get a suspicious inquiry:

• 🔍 Reverse-verify: Send a separate confirmation email to the email publicly listed on the company's website
• 🔒 Don't click links / don't open attachments: Copy the link to VirusTotal for a scan first
• 📞 For large orders, request phone / video to verify identity

---

💡 Learning Tips

Scenario	Tip
🎯 Check AI Inquiries first (but watch for phishing)	Daily prioritize handling AI-marked "Inquiry" emails — but AI is not antivirus, "Inquiry" can also contain phishing. Check sender domain + content relevance before replying
📊 Use "System" to analyze bounces	Bounces concentrate in "System". Periodically check bounce reasons to adjust templates
🚫 Block "Risk" directly	Phishing / scam emails — don't click links, add to blacklist to cut off future harassment
🔄 Correct AI mistakes	Every correction makes AI smarter, accuracy improves significantly within 3 months
🏷️ Only use 3 - 5 core manual categories	Don't build piles of useless tags, trim to the necessary few for maximum efficiency
🌈 Use color smartly	Different importance levels of manual categories use different colors for quick visual identification

---

📚 Related Features

Topic	Link	Description
📜 Email List	email-list	8 list operations + 3-step bulk processing
📬 Receive New Emails	check-new-emails	System mailbox + bound mailbox receive config
📨 Email Account	email-account	Bind personal / corporate mailbox
💡 Compose New Email	compose-new-email	1-on-1 reply to inquiry emails
🏷️ Tags and Views	contacts-tags-views	Contact-level tags (different from email-level categories)
🚀 Mass Email	email-mass-sending	Pinpoint mass-send replies in the system mailbox with AI categorization

---

⚠️ Safety Warning · Phishing Emails Can Mix Into "Inquiry"

AI mainly judges at the semantic level "this looks like an inquiry", but some phishers use inquiry-style phrasing:

• 🎣 "I'm interested in your products, please click this link to see my specs" (lures you to click a malicious link)
• 🎣 "Please confirm your bank details for our pending order" (financial info phishing)
• 🎣 "Open this attachment for inquiry details" (virus attachment)

After getting an inquiry, always verify:

• ✅ Is the sender mailbox a real company domain (abc@company.com ≠ abc@gmail.com)
• ✅ Does the content match your product + customer base (steel seller getting "want to buy clothes" = phishing)
• ✅ Be cautious with attachments / links, when suspicious, reverse-verify the sender's identity with a separate email

For detection details, see ❓ FAQ · Q6 How to identify phishing.

---

🔗 Permanent link: https://laifa.xin/zhinan/email-categorization