The HighLevel Release Radar for August 24 to 28, 2026 introduces five separate capabilities: inbound Voice AI spam controls, Ask AI video and voiceover generation, scheduled availability for Forms, Surveys, and Quizzes, conditional Opportunity fields, and beta certificate design.
This guide turns the announcement into practical setup decisions. It explains what each feature does in the HighLevel interface, what to test, and where a separate review or validation step remains necessary. The Release Radar page displays August 28, 2026 as its publication date. An extracted September 2, 2026 update timestamp was not independently confirmed on the visible page.
These features are not one connected automation suite. The reviewed HighLevel material documents native interface workflows, but does not establish public APIs, webhook payloads, or automation modules for all five updates. Use the official support guides below each section for the detailed product behavior.
What the August 2026 HighLevel updates let you do
Each update solves a different operating problem. Voice AI evaluates call behavior, Ask AI generates media, Schedule Visibility applies a known time boundary, conditional fields guide supported Opportunity saves, and Certificate AI creates an editable design starting point.
Known conditions
Use Schedule Visibility for a real opening or closing deadline, and use conditional fields for stage-specific data collection. These are deterministic interface controls, not AI decisions.
Interpretation or generation
Use Voice AI for behavior-based call handling and Ask AI or Certificate AI for generated output. Test uncertain cases and review the resulting media or credential before use.
| Trigger or need | Feature role | Validation or decision gate | Fallback |
|---|---|---|---|
| Unwanted inbound call | During-call detection or after-call analysis | Test ordinary and clearly unwanted calls; set the threshold deliberately | Review uncertain or valuable callers |
| Video or voiceover needed | Ask AI generates a media asset | Check wording, visuals, timing, pronunciation, and format | Content owner revises or approves |
| Known form deadline | Schedule Visibility opens or closes an asset | Test both inactive states and the selected time zone | Campaign owner checks destination pages |
| Stage-specific opportunity data | Conditional fields show or require data | Check visibility and validate external writes separately | CRM Admin resolves rule conflicts |
| Certificate design needed | Certificate AI creates an editable starting design | Check recipient data, wording, layout, and QR verification | Credential approver holds issuance |
Use AI for interpretation or generation. Use explicit rules for known deadlines and required fields, then keep a human responsible for exceptions.
Configure Voice AI spam protection as two separate controls
During-call detection and after-call analysis have different jobs. During-call detection can instruct a Voice AI agent to end a conversation identified as clear unwanted behavior, such as a sales pitch, telemarketing call, or robocall. After-call analysis evaluates completed inbound calls and supports threshold-based repeat-caller handling.
HighLevel does not guarantee that every unwanted caller will be detected. A legitimate or uncertain caller should not be disconnected solely because the system is suspicious. If the hangup prompt is customized, test the changed behavior before relying on it.
- Enable Voice AI: End Call & Spam Protection in Settings > Labs for the sub-account.
- Open AI Agents > Voice AI, select the agent, and open Actions. Configure End call under During the Call.
- Choose whether to enable during-call detection, after-call spam analysis, or both. After-call analysis applies to completed inbound calls.
- Set the spam threshold and choose notifications for the default Admin recipient, custom email recipients, or both. When the threshold is reached within the documented 24-hour evaluation window, HighLevel can place the caller on inbound Do-Not-Disturb.
- Test ordinary callers, unmistakable unwanted-call examples, notification routing, and any customized hangup prompt. The Voice AI administrator owns the configuration and test decision.
The operating choice is immediate ending for unmistakable unwanted behavior, after-call analysis for repeat-spam handling, or both when the purposes are distinct. This is threshold-based inbound DND, not a guarantee that every unwanted number will be identified or blocked.
See HighLevel’s Voice AI smart call ending and spam protection guide for the documented settings and testing details.
Generate Ask AI video and voiceovers with a reviewable brief
Ask AI supports short videos of 4, 6, or 8 seconds, in landscape or vertical orientation, at resolutions up to 4K, with synchronized audio. A request can include a start image, an end image, or both. Prompts can describe camera movement, lighting, style, scene details, and unwanted elements.
HighLevel recommends creating important on-screen copy as an image before animating it to improve text consistency. A 4K render requires confirmation before the highest-resolution generation is committed. The documentation confirms these capabilities but does not independently confirm a specific underlying video provider.
For voiceover, the documented controls include ten selectable voices, instructions for tone, emotion, pacing, and performance, and playback speed from 0.25x to 4x. Each take supports up to 4,000 characters. Longer scripts can be generated in named sections. Audio exports include MP3, WAV, OPUS, and AAC.
- Open Ask AI and choose video or audio generation.
- For video, enter the scene prompt, orientation, duration, and optional start or end image. For audio, provide the approved script and voice instructions.
- Generate the result and review visible text, spoken wording, timing, pronunciation, and output format.
- Export audio in a supported format. Save generated video through the workspace flow available to the account, including the Media Library when supported. Do not assume automatic publishing.
Keep the approved script separate from the resulting audio asset when wording is governed, such as prices, dates, legal language, medical claims, or regulated disclosures. For longer voiceovers, split the script into named sections and approve each take before assembly.
The official Ask AI audio and video guide documents these controls. It does not confirm Google Veo, workflow invocation, public API access, or automatic publishing.
Schedule Forms, Surveys, and Quizzes around a real deadline
Schedule Visibility is a deterministic control for an individual asset’s public availability. It supports opening right away or at a date and time, closing at a date and time or never closing, and one selected time zone for both boundaries. Before opening and after closing, visitors can see a page or be sent to a link.
Closing blocks new submissions but does not delete existing submissions. Existing Forms, Surveys, and Quizzes retain their usual behavior unless scheduling is enabled.
- Go to Sites, open Forms, Surveys, or Quizzes, and select the individual asset.
- Enable Schedule visibility. Set the opening condition, closing condition, and time zone.
- Choose the before-opening and after-closing experiences separately: show a page or open a link.
- Save, preview, and test before opening, at opening, while active, at closing, and after closing. Check the selected time zone and both destinations.
Use a custom page when visitors need an explanation. Use a link when the next action is a waitlist, replacement offer, or follow-on campaign. Testing near daylight-saving changes and from relevant user locations is an implementation precaution, not a separate product setting.
See HighLevel’s Schedule Visibility guide for the settings, inactive-state behavior, and lifecycle checks.
Use conditional Opportunity fields without mistaking them for universal validation
Conditional Custom Fields let Admins show a field, show a folder, or make a field mandatory according to Opportunity conditions. Supported triggers include pipeline, stage, status, and supported dropdown, radio, checkbox, or multi-select values. Rules can combine conditions with AND or OR logic.
- If needed, enable Show & Require Opportunity Fields Conditionally in Settings > Labs.
- Go to Settings > Custom Fields, select Opportunity, and open Conditional Rules.
- Create a rule, select the trigger fields and values, choose AND or OR logic, and add a Show field, Show folder, or Make mandatory outcome.
- Review the preview and save. Only Admins can create, edit, or delete conditional rules.
A practical rule might require proposal amount and expected close date when a human moves an Opportunity into Proposal. Those field names and conditions are an illustrative operating choice, not a HighLevel template.
A mandatory outcome does not automatically reveal a hidden field. Hidden values remain stored, and hidden fields are excluded from mandatory validation until visible. Multiple applicable rules can combine their mandatory requirements.
Conditional validation protects supported web and mobile Opportunity saves. It does not apply to Public API or workflow-based updates. Validate required values before an external write, then reject, quarantine, or route incomplete records for CRM review.
For an external process, define the validation rule independently. For example, before writing a Proposal Opportunity, check that the proposal amount exists, the date uses an allowed format, and both values belong to the intended record. Do not assume that an API or workflow update receives the same interface validation as a human save.
{
"stage": "Proposal",
"required_fields": [
"proposal_amount",
"expected_close_date"
],
"on_missing": "hold_for_crm_review"
}
The JSON above is an illustrative pre-write validation model, not an official HighLevel contract. If concurrent processes can retry or write the same record, enforce uniqueness in the destination database or use a transactional upsert. A read-then-create check alone is not race-safe.
For broader field governance and external-write controls, see CRM systems consulting. HighLevel’s conditional Opportunity fields guide explains the supported conditions, outcomes, roles, and scope.
Treat Certificate AI as a draft designer, not an unattended issuer
Certificate AI is a Beta feature accessed through Memberships > Credentials > Create > Certificate AI. Prompt with the certificate purpose, organization, recipient details, signer requirements, orientation, style, and verification requirements. Certificate AI may ask follow-up questions before generating an initial design.
Open the generated design in Certificate Builder and review its text, frames, images, signatures, backgrounds, canvas, layers, dynamic recipient information, and QR elements as applicable to the design. Preview mapped recipient information and save only after the design is ready. The available Send flow is for issuing a reviewed certificate, not proof that the generated design is correct.
- Recipient name and mapped values match the intended person.
- Completion or achievement date is correct.
- Organization name, branding, signer name, and signer title are approved.
- Certificate wording is complete and authorized.
- Layout remains readable in the final preview.
- QR verification has the intended destination or behavior.
- The design version is distinct from the recipient-specific issuance record.
QR verification can be reviewed and added or adjusted in Builder. Do not assume every generated design contains a correct QR element. Treat the saved design as a template or version and each recipient-specific certificate as a separate issuance. If you track these outside HighLevel, use a unique issuance key that includes the recipient and the source achievement event, not the template ID alone.
Make a credential owner or designated approver responsible for release, especially when a certificate has formal or compliance significance. Consult the official Certificate AI Beta guide for the prompting, Builder, review, and issuance workflow.
A practical rollout sequence for the five updates
Introduce the features in an order that makes failures easy to detect and reverse.
- Set deterministic rules first. Configure form schedules and conditional field rules, then test every schedule boundary and supported save path.
- Test call behavior in a controlled way. Have the Voice AI administrator check ordinary callers, clear unwanted-call examples, notification routing, threshold behavior, and the uncertain-caller fallback.
- Keep generated assets in review. Assign a content reviewer for Ask AI media and a credential approver for certificates. Do not publish or issue directly from an unreviewed generation.
- Name an owner and fallback. Assign a campaign owner, CRM Admin, Voice AI administrator, content reviewer, and credential issuer as applicable. Record who handles exceptions.
- Measure operational checks, not assumed performance gains. Track schedule test pass or fail, incomplete supported UI saves caught, spam handling reviewed, media approval status, and certificate corrections before issue.
If your team stores feature-related records outside HighLevel, preserve the correct row grain. A call is an event, a spam classification is an observation about a call, and a threshold block is an aggregate decision over a caller and a 24-hour window. A media generation is a run, while a certificate design is separate from a recipient-specific issuance.
For concurrent external writers, use a database-enforced unique key or a transactional upsert at the actual grain. For example, a call observation might use a call event ID plus analysis version, a block decision might use normalized caller plus block-window start and policy version, and a certificate issuance might use recipient plus design version plus source achievement event. These are proposed integration controls, not HighLevel-published fields or guarantees.
Retain only the caller, recipient, prompt, or asset data needed for the operating purpose. Assign a named owner for review and exceptions, and verify authentication scope, payload shape, rate limits, and duplicate handling before proposing any external connection.
The Release Radar is an announcement, while the linked HighLevel support pages provide the detailed setup and limitations. Teams planning a broader platform implementation can review GoHighLevel setup as a relevant service option.
