AI Message Suggestions in HighLevel

Updated: February 4, 2026  | 
Author: Bill Stilwell

This post covers how Suggestive mode works, how it differs from Auto-Pilot, how to train the AI for better suggestions, and how to use it effectively across your team’s inbox workflow.

Reading time: about 8 minutes.

What Are AI Message Suggestions in HighLevel?

AI Message Suggestions is what HighLevel calls Suggestive mode inside the Conversation AI feature.

When an inbound message arrives in the Conversations inbox, the AI reads the message and generates up to three suggested replies. Those suggestions appear in the message composer – your team member reviews them, clicks the one that fits, edits it if needed, and sends.

No message ever goes out without a human approving it first. That is the defining difference between Suggestive mode and Auto-Pilot.

Enable it at Settings, then Conversation AI in the sub-account, and set the bot mode to Suggestive.

Suggestive Mode vs. Auto-Pilot

Both modes use the same AI bot, the same training data, and the same underlying response logic. The difference is who controls the send.

Suggestive mode drafts reply options for a human to review and approve. The AI helps – the human decides.

Every message your contact receives was seen and approved by a person before it was sent.

Auto-Pilot mode sends replies automatically, with no human involvement until a handoff is triggered. The bot handles the entire conversation on its own.

Suggestive mode is the right choice when conversations involve sensitive topics, high-value relationships, or situations where a wrong automated reply would cause real harm. Auto-Pilot is appropriate for high-volume, lower-stakes inbound conversations where speed matters more than nuance.

Some teams use both: Suggestive during staffed hours for human oversight, Auto-Pilot outside those hours so no message sits unanswered overnight. Switching between modes is done in the same settings screen and takes effect immediately.

How Does It Work in the Inbox?

When an inbound message arrives in a conversation where the AI bot is active and set to Suggestive mode, the AI analyzes the message and generates draft replies.

The suggestions appear in a panel above or alongside the message composer. Each suggestion is clickable – clicking it populates the composer with that text so you can review and edit before sending.

If none of the suggestions are appropriate, you can ignore them and type a manual response. The AI suggestions are an accelerator, not a replacement for judgment.

The suggestions draw from the AI’s training data – your website content, FAQ pairs, and any custom training responses you have added. A well-trained bot produces suggestions that are more accurate, more specific, and require fewer edits.

Training the AI for Better Suggestions

The quality of suggestions is directly tied to the quality of training data. A bot with minimal training generates generic suggestions.

A well-trained bot generates replies that sound like they could have been written by a knowledgeable staff member.

Go to Settings, then Conversation AI, then the Bot Training tab. There are two main training tools.

The Web Crawler lets you enter your website URL or specific pages. The AI indexes the content and uses it to inform suggestions.

This is the fastest way to give the bot a foundational understanding of the business, its services, pricing, and FAQs.

The FAQ pairs section lets you add specific question-and-answer pairs. Write each question the way a real contact would ask it – conversational, not formal.

The more closely the training FAQ matches the actual language of inbound messages, the more useful the suggestions will be.

Test your training using the Bot Trial tab. Enter sample messages and review the suggestions before going live.

If the bot consistently produces poor suggestions for a specific type of question, add that question as a new FAQ pair and test again.

Which Channels It Works On

Suggestive mode works across any channel where Conversation AI is enabled in the sub-account. You enable each channel individually in the Conversation AI settings.

Supported channels include SMS, email, Facebook Messenger, Instagram DM, Live Chat, and Google Business Chat where available. The suggestions appear in the same Conversations inbox message composer regardless of which channel the original message came in on.

Enable Suggestive mode only for the channels your team actively monitors. A channel with suggestions enabled but no one checking it is no different from a channel with no AI at all.

What Can You Do With It?

• Cut response time without removing human judgment: Your team goes from reading a message and composing a reply from scratch to reading a message and selecting or editing a ready-made draft. Response time drops significantly while every reply stays human-approved.
• Maintain quality control over AI-generated responses: Unlike Auto-Pilot, Suggestive mode ensures a person reviews every reply before it reaches the contact. For businesses in sensitive industries – healthcare, legal, financial – this is often a non-negotiable requirement.
• Onboard new team members faster: A new staff member handling the inbox for the first time has AI-suggested replies as a guide. The suggestions show them what a knowledgeable response looks like, which accelerates their ability to handle conversations independently.
• Handle high-volume inboxes without hiring: A growing business with increasing inbound message volume can absorb that volume more efficiently when each message arrives with a ready-made draft rather than requiring full manual composition.
• Combine with Auto-Pilot for 24/7 coverage: Use Suggestive during business hours for human-reviewed responses. Switch to Auto-Pilot after hours so contacts who message overnight get an immediate, intelligent reply rather than waiting until morning.
• Reduce the mental load of inbox management: Reading a message and evaluating three suggested responses is cognitively lighter than reading a message and composing a reply from scratch. Over the course of a busy inbox day, that difference adds up.

Key Definitions

Use Cases by Industry

Marketing Agencies

An agency manages inbound leads for a home services client through HighLevel. The client’s office manager handles the Conversations inbox but is not a trained sales communicator.

With Suggestive mode enabled and the client’s services and FAQs loaded into Bot Training, the office manager opens each new message and sees three ready-made reply options. She selects the most appropriate one, sometimes adds a personal detail, and sends.

Response quality is consistent regardless of her individual communication skills.

Result: The agency’s client looks professional in every conversation without the agency having to write every reply manually – and the office manager gains confidence handling inbound sales conversations.

Medical and Healthcare

A medical practice uses HighLevel to manage inbound appointment inquiries and patient questions. Auto-Pilot is not appropriate because some messages involve sensitive health information requiring careful human handling.

Suggestive mode gives the front desk team AI-drafted replies for routine scheduling and FAQ-type questions – “what are your hours,” “do you accept my insurance,” “how do I reschedule” – while keeping a human in the loop to review every response before it reaches the patient.

Result: The front desk handles higher message volume without sacrificing the careful human review that the practice’s patient relationships require.

Real Estate

A real estate agent receives 30 to 50 inbound messages per week from leads at various stages. Writing individualized replies to every message is time-consuming and often gets deferred.

Suggestive mode provides draft replies for each message that the agent reviews and personalizes with specific property details or follow-up questions. The agent spends 10 seconds choosing and lightly editing rather than 2 to 3 minutes composing from scratch.

Result: Response time drops from hours to minutes for most messages – and leads who receive fast, relevant replies progress through the pipeline at a higher rate.

E-commerce and Online Businesses

An e-commerce business receives inbound messages across Live Chat, Facebook Messenger, and SMS. The customer service team uses Suggestive mode to handle questions about orders, returns, shipping, and product details.

The bot is trained on the full FAQ from the website, return policy, and shipping information. Most common customer questions produce accurate suggestions that require minimal editing – the team handles three times as many conversations per hour as they could manually.

Result: Customer service capacity scales without adding headcount – and response times are faster across all channels simultaneously.

Legal and Financial Services

A law firm uses HighLevel to manage inbound inquiries. Every reply to a potential client carries professional and ethical weight – the firm cannot use Auto-Pilot for any of these conversations.

Suggestive mode provides the intake coordinator with draft replies to common initial inquiry questions. The coordinator reviews every suggestion before sending and adds any case-specific details the AI could not know.

The drafts save 60 to 90 seconds per message without removing the coordinator’s judgment from the process.

Result: The intake team handles more inquiries per day with less cognitive load – while maintaining the careful human oversight that professional responsibility requires.

Who Is This For?

How to Enable and Use AI Message Suggestions

Step 1: Open Conversation AI settings

Go to Settings in the sub-account left navigation, then select Conversation AI.

This is where Suggestive mode, Auto-Pilot mode, channel settings, and bot training are all configured.

Step 2: Set the bot mode to Suggestive

In the Conversation AI settings, select Suggestive as the bot mode.

This enables AI suggestions in the inbox without activating automatic sending. No message will go out until a human approves it.

Step 3: Enable channels

Select which channels the AI should generate suggestions for – SMS, email, Facebook Messenger, Instagram DM, Live Chat.

Enable only the channels your team actively monitors. There is no benefit to suggestions on a channel nobody checks.

Step 4: Add bot training data

Go to the Bot Training tab. Enter your website URL in the Web Crawler section to let the AI index your content.

Add FAQ pairs for the most common questions your contacts ask. Write each question conversationally – the way a real contact would type it, not the way a business would phrase it formally.

Step 5: Test using Bot Trial

Go to the Bot Trial tab. Enter sample inbound messages that reflect what your contacts typically ask.

Review the suggestions generated. If they are inaccurate or off-topic, add the relevant FAQ pair or additional training content before going live.

Step 6: Go live and handle real conversations

Once Suggestive mode is active, open the Conversations inbox and handle real inbound messages using the AI suggestions.

Review the quality of the first 20 to 30 suggestions to identify patterns where the AI consistently misses the mark – these are training gaps to address.

Step 7: Use suggestions as a starting point

Click any suggestion to populate the message composer. Edit before sending if needed – adjust tone, add specific details, or correct anything inaccurate.

Ignore suggestions that are significantly off and type a manual response. The AI is an accelerator, not a requirement.

Step 8: Refine training based on gaps

When suggestions are consistently poor for a specific type of question, add that question as a new FAQ pair in Bot Training.

More training data covering the actual questions contacts ask produces more accurate suggestions over time.

Step 9: Consider a hybrid mode schedule

If your team is staffed during defined hours, consider using Suggestive during those hours and switching to Auto-Pilot outside them.

This gives you human oversight when staff is available and automated replies when they are not – covering the inbox around the clock without requiring anyone to work outside business hours.

How Does It Connect to HighLevel?

• Conversation AI: AI Message Suggestions is the Suggestive mode of Conversation AI. Both modes live in the same settings section and use the same training data – switching between them changes only who controls the send, not how the AI generates responses.
• Notification Settings: Notification Settings ensures team members are alerted when new inbound messages arrive so they see the AI suggestions promptly and respond quickly rather than letting conversations sit.
• Mobile App: The Mobile App surfaces AI suggestions in the Conversations section on iOS and Android – team members working away from their desks can review and approve AI drafts from their phone.
• Tag-Based Automation: When a human responds to a conversation after reviewing suggestions, Tag-Based Automation can apply tags based on the contact’s replies – automatically triggering next steps in the follow-up sequence based on what the contact said.
• Agent Reporting: Agent Reporting tracks the response time and message counts for each team member. With Suggestive mode active, response times typically improve – comparing pre- and post-Suggestive response time metrics shows the direct impact on team performance.

Common Questions

What are AI Message Suggestions in HighLevel?

Draft reply options generated by Conversation AI’s Suggestive mode. Up to three appear in the message composer when an inbound message arrives.

A human reviews, selects or edits, and sends. Nothing goes out automatically.

How is Suggestive mode different from Auto-Pilot mode in HighLevel?

Suggestive mode drafts reply options for human review before sending. Auto-Pilot sends replies automatically without any human involvement.

Both use the same AI and training data – the difference is who controls the send.

Where do I find AI Message Suggestions in HighLevel?

Go to Settings, then Conversation AI in the sub-account. Set the bot mode to Suggestive.

Suggestions then appear automatically in the Conversations inbox message composer when inbound messages arrive.

Can I edit an AI message suggestion before sending it?

Yes. Click a suggestion to populate the message composer, then edit freely before sending.

The suggestion is a starting point – adjust tone, add specifics, or correct anything inaccurate before the message goes out.

How does HighLevel AI know what to suggest?

The AI uses training data from Conversation AI settings – website content indexed by the web crawler and FAQ pairs you have added manually. More relevant training data produces more accurate and useful suggestions.

Can I use AI Message Suggestions on SMS, email, and other channels?

Yes. Suggestive mode works across any channel where Conversation AI is enabled – SMS, email, Facebook Messenger, Instagram DM, Live Chat. Enable channels individually in Conversation AI settings.

Can I switch between Suggestive mode and Auto-Pilot mode?

Yes. Change the bot mode in Settings, then Conversation AI at any time. Some teams use Suggestive during staffed hours and Auto-Pilot outside them for around-the-clock coverage.

Does AI Message Suggestions work on the HighLevel Mobile App?

Yes. Suggestive mode suggestions appear in the Conversations section of the HighLevel Mobile App – team members can review and approve AI drafts from their phone exactly as they would on desktop.

How do I train the AI to give better message suggestions?

Go to Settings, then Conversation AI, then Bot Training. Add your website URL to the web crawler and add FAQ pairs written in the conversational language your contacts actually use.

Test suggestions using the Bot Trial tab before going live.

Is AI Message Suggestions available on all HighLevel plans?

Conversation AI is priced at $0.02 per suggestion or $49 per sub-account per month for unlimited usage. Check your account settings for current pricing and whether your plan includes AI credits.

To Wrap It Up

Suggestive mode occupies the most useful middle ground in AI-assisted communication: the AI does the drafting, the human does the deciding.

Auto-Pilot is faster and requires less oversight, but it is not the right tool for every business or every conversation. Some situations require a person’s judgment before any reply goes out.

Suggestive mode gives you the speed benefit of AI drafting without giving up that judgment.

The training investment matters. A bot trained on 20 well-written FAQ pairs that cover the actual questions your contacts ask will produce suggestions that require 10 seconds of review.

A bot with no training produces suggestions that take 30 seconds to reject and replace manually. The upfront time spent on training pays back with every conversation handled afterward.

For teams managing high volumes of inbound messages, the hybrid approach – Suggestive during staffed hours, Auto-Pilot after hours – gives you both human oversight when it matters and continuous coverage when it counts.

Here is how to get started:

1. Go to Settings, then Conversation AI in the sub-account
2. Set bot mode to Suggestive
3. Enable only the channels your team actively monitors
4. Go to Bot Training and add your website URL to the web crawler
5. Add 10 to 20 FAQ pairs covering your most common inbound questions
6. Test using Bot Trial before going live – enter sample messages and evaluate the suggestions
7. Go live and monitor the first 20 to 30 real conversations
8. Add FAQ pairs for any question types where the bot consistently misses
9. Review Agent Reporting after 30 days to measure the response time improvement

Train the bot before you go live. A Suggestive mode with no training data gives your team three unhelpful suggestions to ignore rather than one strong one to approve.

Twenty minutes of upfront training changes the experience completely.