What are Skills? How to Configure and Use Skills?

I. Function Introduction

The skills enable AI employees to complete tasks such as information recording, customer tagging, advertising product identification, and proactive follow-up based on customer sources, conversation content, and preset rules.

Currently, AI employees support the following skills:

Proactively follow up with customers: If a customer does not reply to the AI's last message within a set time, proactively send a follow-up message to continue following up with the customer.
Collect customer information: Identify and record information provided by customers or their current status during sessions, and save them as user-level or session-level attributes depending on the scope of use. Eligible user-level attributes can also be synchronized to SaleSmartly customer profile fields as needed.
Update customer tags: Determine customer characteristics based on conversation content and record them as HelpKnow customer tags. They can also be synchronized to SaleSmartly visitor tags as needed.
Learn from advertising information: Determine the product a customer is inquiring about based on the source of the customer's advertisement.

The four skills address different problems and can be used individually or activated simultaneously as needed for business operations.

This article mainly introduces the functions, selection methods, configuration entry points, and activation rules of various skills. For specific field settings, testing methods, and result viewing locations, please refer to the corresponding skill's tutorial.

This article's link: "What are skills? How to configure and use skills?"

II. Skill Types and Selection

You can first select skills based on your business objectives:

If a customer does not continue to respond, you need to proactively follow up and send messages according to the follow-up time and user scope: use "Proactive Customer Follow-up".
You need to record specific details such as customer purchase quantity, budget, and contact information: use "Collect Customer Information".
To add standard categories to customers based on conversation content, use "Update Customer Tags". If you need to tag and categorize customers in SaleSmartly, you can sync the results to SaleSmartly Visitor Tags.
When customers inquire via Facebook ads through Messenger or WhatsApp, we want AI employees to directly obtain the product information the customer is inquiring about based on the source ad ID, and accurately reply with product information to improve inquiry and order conversion efficiency: using "Learning Ad Information".

1. Proactively follow up with customers

"Proactive Customer Follow-up" is used to send follow-up messages to AI employees at set times when customers do not continue to respond.
You can set the follow-up time, AI follow-up requirements, and follow-up users. You can also limit the follow-up scope through custom attributes, contact attributes, or system attributes.
In addition to generating follow-up content, the AI can also be instructed to invoke pre-configured operations such as transferring to human assistance or updating customer tags or attributes during the follow-up call.
The same callback rule can be triggered at most once per session; if multiple rules apply to the same session, each rule can be triggered once. Each callback will consume 1 message quota.

For detailed setup instructions, please refer to: "How to Set Up AI-Powered Employees to Proactively Follow Up with Customers?"

2. Collect customer information

"Collecting customer information" is used to identify and record specific information or current status during a conversation, such as customer name, mobile phone number or email address, company name, purchase quantity, customer budget, and customer mood.
When creating an attribute, you need to select the attribute level according to its usage scope:
- User level: Used to record long-term valid customer information, such as company name, customer identity, and contact information. Once recorded, it can be retrieved in subsequent new sessions with that customer.
- Session-level: Used to record information relevant only to the current session, such as the quantity purchased, the product inquired about, or the after-sales issue. This information needs to be collected again when a new session begins.
Once a new value is detected for the same attribute, the new value will be used to overwrite the original attribute value.
Attributes are stored in HelpKnow by default. Only user-level "Custom Text" and "Number" attributes can be mapped to SaleSmartly customer profile fields; session-level attributes and "Option" type attributes do not support mapping.

For detailed setup instructions, please refer to: "How to Set Up AI Employees to Collect Customer Information?"

3. Update customer tags

"Update Customer Tags" is used to determine customer characteristics based on the conversation content and record the determination results as HelpKnow customer tags.
The tag values and corresponding judgment descriptions need to be configured in HelpKnow in advance. The AI staff will determine which tags need to be added based on the existing tag rules and will not generate new tag values on the spot during the conversation.
Customer tags are added appended. A single customer can have multiple tags at the same time, and adding a new tag will not overwrite or remove existing tags.
To synchronize the judgment results to SaleSmartly, you can select "Map Tags (SaleSmartly)" in the tag rules. Map Tags is an optional field:
- If no mapping label is selected: the judgment result will only be recorded in HelpKnow.
- Select Mapping Tags: After a customer initiates an actual session through the SaleSmartly channel and triggers the tag rules, the judgment result will also be synchronized to the corresponding SaleSmartly visitor tag.
- Visitor tags used for mapping need to be created in SaleSmartly beforehand.

For detailed setup instructions, please refer to: "How to Set Up AI Employees to Update Customer Tags?"

4. Learn about advertising information

When a customer contacts you via Facebook ads in Messenger or WhatsApp, HelpKnow can obtain the customer's referral ad ID.
After enabling "Learning Advertising Information", the AI employee will search for the corresponding product information based on the advertising ID, obtain the product information that the customer is inquiring about, and reply with the product name, price, image or description.

For example, after a customer clicks on a bicycle advertisement, even without explicitly mentioning the product name, the AI employee can retrieve the corresponding product information and respond.

Customer: How much is this one?

AI Employee: Hello! This bicycle is priced at $48.00. Are you planning to ride it yourself or give it as a gift? I can provide you with more product details.
Before using this feature, you need to fill in the ad ID and corresponding product information in the "Ad & Product Mapping Table". This feature only works for Messenger and WhatsApp channels integrated with SaleSmartly.

For detailed setup instructions, please refer to: "How to Make AI Employees Learn Advertising Information?"

5. Selection of Attributes and Customer Tags

Both attributes and customer tags can record information from a session, but they differ in how they are saved and their applicable scenarios:

User-level attributes: These are used to store single, long-term valid values or states for customers, such as company name, customer identity, and contact information. They can still be read in a new session.
Session-level attributes: Used to store a single value or status that is only relevant to the current session, such as the quantity purchased this time, the product currently inquired about, or the after-sales issue.
Customer tags: Used to record multiple customer characteristics that can coexist, such as bulk purchasing, price sensitivity, and urgent delivery time.

Once a new value is detected, the attribute will overwrite the original value; customer tags are appended, so adding a new tag will not overwrite an existing tag.

For example, you can use the session-level attribute "Purchase Quantity" to record the specific quantity of this purchase, and use the "Bulk Purchase" tag to mark customer characteristics.

6. Examples of skill combination usage

Suppose a customer clicks on a Facebook ad for a product and starts asking questions, stating in the conversation:

I'd like to purchase 200 units, with a budget of approximately $5,000. Are there any discounts available for bulk purchases?

After the AI employee replied with product and price information, the customer did not respond further. The four skills can each complete the following tasks:

Learn advertising information: Obtain information about the products that customers are inquiring about based on the source ad ID.
Collect customer information: Record the quantity of this purchase as "200 pieces" using a session-level attribute, and record the customer's budget as "$5,000".
Update customer tags: Add a "bulk purchase" tag; if the customer's bargaining behavior meets the preset judgment criteria, a "price sensitive" tag can also be added.
Proactively follow up with customers: After the set follow-up time has elapsed, proactively ask customers if they still need to know about pricing, discounts, or order placement methods.

Each of the four skills handles its corresponding task and they do not substitute for one another. You can activate only the skills you need based on your actual business objectives.

III. Skill Configuration and Activation

Enter the AI employee that needs to be configured, click "Configure" in the left menu, and then find the "Skills" area.

1. Configuring skills

Different skills require different configurations:

Collect customer information: Use system attribute templates or create attributes according to business needs, and set attribute types and attribute levels; attributes that meet the mapping conditions can also select the corresponding SaleSmartly customer information fields.
Update customer tags: After enabling the feature, set the tag value and judgment description; if you need to synchronize with SaleSmartly, then select the corresponding mapping tag.
Learn about advertising information: After enabling the feature, upload an "Ad & Product Mapping Table" containing the ad ID and product information.
Proactively follow up with customers: Create follow-up rules and set the rule name, follow-up time, AI follow-up requirements, and follow-up users.

Simply enabling or entering a skill does not mean the configuration is complete. You also need to set return visit rules, attributes, tag rules, or advertising product mapping tables according to the corresponding skill.

For specific fields, testing methods, and results viewing locations, please refer to the corresponding skill-specific tutorials.

2. Apply the configuration.

Different skills take effect in different ways:

Collect customer information: Customer attributes take effect automatically after creation, editing, deletion, or activation/deactivation; no publishing is required.
Update customer tags: You need to publish after enabling or disabling the feature; after adding, editing or deleting tag rules, click "OK" for the changes to take effect automatically.
Learning advertising information: After modifying the function switch or re-uploading the mapping table, you need to publish the AI employee configuration.
Proactive customer follow-up: Rules for creating, editing, deleting, or starting/stopping follow-up visits take effect automatically and do not need to be published.

If the "Publish" button in the upper right corner of the page is clickable, it means there are still changes that need to be published.

IV. Precautions

Attributes use a single-value overwrite method, while customer tags use a multi-value append method. Please select the appropriate method based on the data you need to record.
When creating attributes, select either user-level or session-level based on the scope of the data. Only user-level custom text and numeric attributes can be mapped to SaleSmartly customer profile fields.
The Q&A test can verify whether HelpKnow customer tags are generated, but it cannot verify whether SaleSmartly visitor tags are synchronized.
"Learning Ad Messages" only apply to Messenger and WhatsApp channels integrated with SaleSmartly.
Proactive follow-ups cannot be triggered in the "Q&A Test". For testing purposes, please create a real conversation through websites or social media channels that are already integrated with SaleSmartly.

V. Frequently Asked Questions

1. Why is there no result after activating the skill?

Please first confirm that the corresponding skill has been configured as required:

Proactively follow up with customers to see if they have created and enabled follow-up rules, whether the set follow-up time has been reached, and whether the current user or session meets the filtering criteria.
Check if customer information has been created and the corresponding attributes have been enabled.
Update customer tags to check if tag values have been configured and determine the description.
Check if the advertising information has been uploaded to the advertising and product mapping table.
Has the required changes already been published?

If the configuration has already taken effect, you also need to confirm whether the test session contains information that can trigger the corresponding skill.

2. Can I use multiple skills at the same time?

Yes. For example, AI employees can retrieve corresponding product information based on the source ad ID, record the purchase quantity using attributes, add a "bulk purchase" tag, and proactively send follow-up messages when the customer does not respond further.