AI usage caps enter Labs beta

NEWS / LABS BETA

1. An AI cap needs a service-continuity decision

Announced 30 October · Checked 31 October 2026

What changed

HighLevel says Pro agencies can use its Labs beta to set client usage limits for Voice AI and Conversation AI. The operator chooses what happens at the limit, such as billing additional usage or stopping the action. Other products are not yet covered.

Sub-account admins can see usage against the limit and reset dates on the billing page. Clients who cannot upgrade their SaaS plan can request a higher limit; those who can upgrade see that option instead. Approving a request changes the limit only, not the behaviour when it is reached.

The announcement lists notifications for limit changes and requests. Alerts at 50%, 80% and 100% are described as coming soon. Usage limits in the Enable Rebilling flow and bulk action are also future work. Do not assume those controls or percentage alerts are already available.

Who this affects

Businesses using Voice AI or Conversation AI through an eligible Pro agency. Growthcenter users should confirm their account, beta access and billing terms with the existing operator. Businesses without these products can ignore the new controls; this is not a reason to purchase an AI service.

Our analysis: what it means for your business

Hypothetical example: a plumbing business uses an AI receptionist after hours. Its owner wants a predictable bill, while the dispatcher wants every enquiry to reach someone. A setting that stops an action at a usage limit addresses the first concern differently from a setting that permits paid overages. Neither choice, by itself, establishes what the caller experiences next.

The practical recommendation is to document two decisions separately: who can authorise additional spend, and who handles an enquiry when the AI service cannot continue. Ask the operator to explain the configured behaviour in plain language. A hard cap needs a verified fallback; an overage policy needs an agreed approval boundary. Do not infer an automatic human transfer from a billing feature announcement.

Reset dates matter as much as the size of a limit. A busy week near the end of a usage period may require a different response from the same activity just after a reset. Review the account’s actual usage and billing period before selecting a threshold. This report does not establish the units, prices or commercial terms that apply to any Australian customer account.

Because percentage-based alerts remain future work, assign a person to check usage until the required alerting is available and verified. A request notification is not the same as advance warning that a cap is approaching. The owner should know where that person looks, how often they review it, and who covers an absence.

An approved increase also leaves the cap behaviour unchanged. If the previous choice was to stop the action, raising the threshold does not turn it into an overage arrangement. Keep the new limit and the existing stop-or-bill choice together in the operating record. No account setting was changed for this article, and the beta is not evidence of lower costs or uninterrupted customer service.

What to do next

1. Confirm product eligibility, beta access, usage units, reset date and the configured stop-or-bill behaviour.
2. Name the spending approver and document the customer-service fallback separately.
3. Review usage manually until any required percentage alerts are released and verified in the account.

NEWS / MOBILE 4.27.x+

2. Call details show who answered, not who owns the next step

Announced 30 October · Checked 31 October 2026

What changed

HighLevel’s mobile call-details update shows the teammate who answered, or a phone number when no teammate answered, plus the answering device. It covers regular, IVR and WhatsApp calls in conversation details. Phone numbers are masked; unknown devices have no icon. Equivalent context in call logs is listed as future work.

A separate update displays multiple recordings on a call, each playable individually. Transcripts appear only where the backend already has one for that recording. Another update adds connection-quality bars and Network, Device and Audio information, plus phone-call ratings. WhatsApp call ratings are still described as next.

All three announcements specify version 4.27.x or later, iOS and Android, and HighLevel, LeadConnector and self-serve white-label apps. They do not establish that every Growthcenter user already has the relevant app build or account access.

Who this affects

Teams using these mobile apps with shared ringing, forwarding or transferred calls. Owners using a separate phone system can ignore the app changes. Existing recording settings and access permissions still need to be understood; this announcement does not supply new recording consent or permissions.

Our analysis: what it means for your business

Hypothetical example: a customer rings a cleaning business and the call is assigned to the office manager, but another teammate answers on mobile. The new detail can help identify who actually spoke to the customer. It does not show that the office manager has received the promised quote details or accepted responsibility for sending them.

The useful handoff rule is therefore explicit: record the customer’s requested next step, name its owner and confirm that person has received it. Use the answerer field to resolve uncertainty about the conversation, rather than treating it as an automatic reassignment. After a transfer, ask who is responsible for closing the loop with the customer.

Multiple recordings may help an authorised operator review different legs of an existing call. A missing transcript should not be treated as proof that the conversation never happened, because the vendor makes transcript availability conditional. Review only recordings the business is authorised to access; this feature announcement is not a recommendation to enable new recording.

Connection diagnostics can provide a useful troubleshooting clue when audio breaks up. They are not proof that both parties understood an address, appointment time or price. Confirm critical details through the established customer process. Likewise, a call rating is feedback about the call experience, not evidence of a qualified enquiry or completed booking.

What to do next

1. Verify the app version and account access before adding the new controls to a staff procedure.
2. Separate the person who answered from the person who owns the follow-up.
3. Treat missing transcripts and connection indicators as limited signals; verify the actual customer next step.

NEWS / STABLE PATCH

3. n8n patches email-trigger and execution-access checks

Released 30 October · Checked 31 October 2026

What changed

n8n stable 2.45.7 requires execution:read permission when workflow-scoped routes read execution data. Its release notes also list fixes for IMAP trigger expressions after activation and for handling user data containing an __isError key.

Who this affects

Operators maintaining n8n, particularly workflows that start from an IMAP email inbox or depend on execution-data access. Businesses without n8n can ignore this patch.

Our analysis: what it means for your business

Our recommendation: ask the maintainer to identify the installed version and relevant dependencies, then follow the normal update and verification process. A controlled sample email can help check whether the intended enquiry record and staff handoff occur after activation. Verify permission boundaries separately. The release notes do not establish that a customer workflow failed or that exploitation occurred; no customer installation was inspected.

What to do next

1. Check the installed version and affected workflow paths with the maintainer.
2. Verify an authorised sample email and execution permissions after the planned update.