August 17, 2026. HighLevel extended the AI Assistant in Workflows to report WhatsApp performance alongside Email and SMS on August 14. You ask about any WhatsApp step in plain language and get delivery and read metrics calculated from live data, with no configuration required. HighLevel notes in the release note that WhatsApp is among the most used actions in Workflows, which is precisely why the numbers it returns are about to start appearing in client reports. Two of them do not mean what a client will assume.
What the Assistant now returns
- Volume and outcome counts for any WhatsApp step. Total sends, delivered, failed and pending, plus calculated delivery and failure rates.
- Read data. Read counts among delivered messages, and the corresponding read rate.
- Automatic period comparison. Name a window such as the last two weeks and the Assistant compares it against the immediately preceding period, showing volume movement and delivery direction.
- Insight led answers. Responses lead with the headline figures, then describe what changed in plain language.
The read rate has a delivered denominator
HighLevel spells it out: read counts are among delivered messages, and the read rate is calculated from those. That is the correct way to measure a read receipt and the wrong number to paste beside an email open rate.
The consequence is arithmetic. A step that sends 1,000 messages, delivers 600 and gets 300 reads reports a 50 percent read rate, while only 30 percent of the intended audience read anything. A step with worse targeting but better delivery can look worse on the same dashboard. The Assistant hands you sends and delivered in the same answer, so compute reads over sends yourself and show both. One is a platform performance metric and the other is a campaign performance metric, and a client will read whichever number is larger as the campaign result.
One interface, three different meanings
The stated benefit is that Email, SMS and WhatsApp now resolve through the same conversational interface instead of separate dashboards. That is genuinely useful and it introduces a comparison the channels do not support. A WhatsApp read is a receipt sent back by the recipient's app. An email open is inferred from a tracking pixel and has been unreliable since mail providers began prefetching images. An SMS has no read event at all. Putting the three in one answer makes them look like one metric family. Name the channel next to every rate in a client facing report, and do not build a blended engagement figure across all three.
The comparison window has the same shape of problem. The Assistant compares your chosen window against the immediately preceding period, which is a fixed rule rather than a setting. Ask for the last two weeks during a launch or over a holiday and the baseline is whatever happened to sit behind it. Put the actual dates in the report rather than the phrase you typed.
What it means for operators
Three moves. First, decide today whether your client reports quote read rate over delivered or over sent, write the definition into the report footer, and never change it mid engagement. Second, remember the Assistant reports at the step level, so delivery and reads say nothing about replies, bookings or pipeline. Reconcile against conversations and opportunities before anyone calls a workflow successful. Third, check the colours on any dashboard the client already has. Failure rate is a metric where a fall is good, and HighLevel only added the option to colour a decrease as positive on August 11, with existing widgets defaulting to increase is positive. We covered that default in our note on the trend badge release. A falling failure rate showing red is a support ticket waiting to happen.
If you want workflow reporting that a client can read without a translator, we build GoHighLevel reporting and automation and wider AI automation for agencies that have outgrown the default dashboards.
Frequently Asked Questions
For any WhatsApp step in a workflow it returns total sends, delivered, failed and pending counts, calculated delivery and failure rates, read counts among delivered messages and the corresponding read rate. It also compares your chosen window against the immediately preceding period.
From delivered messages. HighLevel states that read counts are among delivered messages and the read rate is the corresponding figure. If delivery is poor the read rate will look better than the share of the intended audience that actually read the message, so compute reads over sends as well.
No. HighLevel says no configuration is required. Open the AI Assistant in any workflow and ask for the metrics in plain language, for example the WhatsApp delivery stats for the workflow over the last two weeks.
Not directly. A WhatsApp read is a receipt from the recipient's app, an email open is inferred from a tracking pixel and is unreliable, and SMS has no read event at all. Label the channel beside every rate and avoid building a blended engagement number across the three.