The HighLevel AI Employee is not one AI worker. It is HighLevel’s bundle of AI products, managed per sub-account and sold three ways: pay-per-use with no monthly fee, Growth at $50 per enabled location per month, and Unlimited at $97 per enabled location per month.
Our verdict: use it for locations that miss calls or messages outside business hours, and skip it where nobody will read transcripts or keep the prompt and Knowledge Base current. Start on pay-per-use, measure one week of real usage, and move a location to Unlimited once its measured pay-per-use AI bill approaches $97 a month. Skip Growth unless your test week proves it: its voice allowance is small, and the rest of the $50 depends on usage you have to measure first.
What The Highlevel AI Employee Includes
Six metered products (Conversation AI, Voice AI, Reviews AI, Content AI, Ask AI and AI Studio) are covered by the plans, three more are free on every plan, and one, Agent Studio, is covered by none. Managed Agents, formerly Super Agents, run on a monthly allowance of runs. The list can change, because the name covers whatever HighLevel currently prices under these three options.
These are AI fees, not platform tiers: the pricing page lists them per enabled location, separate from any HighLevel platform plan, so do not confuse the $97 Unlimited AI plan with a platform price.
| Pay-per-use | Growth | Unlimited | |
|---|---|---|---|
| Monthly fee | None | $50 per location | $97 per location |
| Conversation AI | Token cost | 1,000 agent responses | Unlimited |
| Voice AI | Voice engine + TTS + LLM tokens | 100 minutes | Unlimited |
| Reviews AI | $0.01 per sent reply | Unlimited | Unlimited |
| Content AI | $0.063 per image; $0.0945 per 1,000 words | Unlimited | Unlimited |
| Ask AI, AI Studio | Token cost | Usage included | 3x Growth’s usage |
| Managed Agents | Token cost | 100 runs | 1,000 runs |
Funnel AI, Workflow AI and Email AI are free on every plan (apart from premium Workflow AI actions, covered below), and the Knowledge Base is included at no extra cost.
When Each Plan Pays Off
Unlimited beats pay-per-use once a location’s voice bill reaches $97, which is about 485 to 970 minutes a month at speech-to-speech rates; the table below covers the other setups.
Voice AI is the clearest place to find the break-even, because it is the only metered product with a published per-minute price. Since May 20, 2026 the voice engine costs $0.045 per minute, and you add text-to-speech ($0.015 for OpenAI or Cartesia, $0.035 for ElevenLabs V2.5, $0.17 for ElevenLabs V3) plus LLM tokens. The alternative is speech-to-speech at a flat $0.10 per minute (Google Gemini) or $0.20 per minute (OpenAI).
| Per-minute AI cost | Pay-per-use equals Unlimited ($97) at | Growth plus overage equals Unlimited at |
|---|---|---|
| Speech-to-speech, Gemini, $0.10 | 970 min/month | 570 min/month |
| Speech-to-speech, OpenAI, $0.20 | 485 min/month | 335 min/month |
| OpenAI or Cartesia TTS, $0.06 before tokens | Under 1,617 min/month | Under 883 min/month |
These are our calculations from HighLevel’s published rates, for Voice AI alone. On the third row, LLM token costs push the real break-even lower than shown. Every row excludes Phone System charges.
Why Growth Is Hard To Justify
Most of what Growth includes is cheap to buy separately. Its 100 voice minutes cost $10 to $20 at speech-to-speech rates on pay-per-use, and its unlimited Reviews AI and Content AI cost one cent per review reply and 6.3 cents per image. That leaves the rest of the $50 to Conversation AI (1,000 agent responses) and to the Ask AI, AI Studio and Managed Agents allowances, which are otherwise billed at token cost.
HighLevel publishes token rates rather than a price per response, so the value of those 1,000 responses cannot be calculated from published data. Measure tokens per response during your test week (see the final section) and price 1,000 responses yourself. If 1,000 responses plus 100 voice minutes price out below $50 at pay-per-use rates, Growth costs you money unless you also use its Ask AI, AI Studio or Managed Agents allowances.
Who Each Plan Suits and Who Should Avoid It
Pay-per-use suits testing and light use, Unlimited suits sustained voice volume or a fixed cost ceiling, and Growth fits only a narrow case.
| Plan | Best for | Avoid if |
|---|---|---|
| Pay-per-use | Any location you are testing, and locations with light or unpredictable usage, since there is no monthly fee | You need a fixed monthly cost |
| Growth | A location with steady text conversations and few calls, when your test week shows Growth costs less than pay-per-use | Voice use is heavy, because minutes past the allowance bill at pay-per-use rates |
| Unlimited | Locations with sustained voice volume, and agencies that want a fixed AI cost per client | Usage is light, because the $97 is billed monthly for each enabled location however little it is used: ten locations cost $970 |
What The Plan Price Does Not Cover
Phone charges, Agent Studio, premium workflow actions and anything past a plan’s allowance all bill on top of the flat fee, and “unlimited” has a fair-use ceiling.
Phone calls. Unlimited Voice AI covers the AI portion of a call. Phone System charges stay separate on all three plans, which is why an Unlimited location can still show phone charges.
Agent Studio. It is pay-per-use on every plan, including Unlimited, and web search inside it adds $0.01 per search.
Premium Workflow AI actions. Decision Maker, Intent Detection, Summarize Text and Translate are listed at $0.01 per execution. The pricing page does not say Unlimited waives them.
Growth overages. After the allowance runs out, the location’s AI Usage Limit setting decides what happens. With “Keep AI running, just notify” (or spending limits off), AI continues and bills at pay-per-use rates. With “Block AI at the limit,” AI stops. Every sub-account defaults to “Keep AI running, just notify,” so AI keeps billing past your limit unless you switch a location to blocking.
Ask AI and AI Studio limits. Usage is measured in 5-hour windows. If a plan’s limit is reached, access resets when the window ends. On pay-per-use, the same work bills at token cost.
Managed Agents runs. A run is one chat session, one trigger event or one scheduled invocation. An agent scheduled daily uses 30 runs a month, so three of them consume 90 of Growth’s 100 runs before any chat session or tag trigger fires.
Fair use. “Unlimited” is not uncapped. HighLevel may throttle, limit, require an upgrade or terminate access for excessive use. The numeric threshold is not publicly documented.
What It Does Well and Where It Oversells
The documented strength is cost control on Unlimited, not autonomy. HighLevel describes an AI that answers calls, books appointments, replies to messages and reviews, drafts content and builds automations. The call and message totals on HighLevel’s AI page and the customer testimonials are self-reported, and HighLevel’s pricing and product pages give no benchmark for booking or conversion rates. Any figure for how many calls an agent handles unaided is an estimate until you measure your own transcripts.
The “employee” label oversells the product. Someone has to write the prompt and load the Knowledge Base, and someone should read call transcripts. Speech-to-speech agent prompts are capped at 15,000 tokens, so the prompt is not the place for your whole business.
HighLevel also notes that larger knowledge-base context can raise per-call token use, so a bigger knowledge base is not free. Load what callers actually ask, not everything you know. Treat the AI as a first-line responder, with a human path for everything else.
For Agencies: Rebilling vs Reselling
Rebilling passes usage costs to a client with a markup, or with a fixed rate per message or Voice AI minute. Reselling offers Growth or Unlimited to the client as a recurring subscription. Fixed-rate rebilling applies to Conversation AI and Voice AI only. HighLevel states that fixed-rate rebilling does not change your underlying charge: your agency still owes the real cost.
That makes a fixed per-minute rate a bet on your cost per minute. Price it above your highest-cost configuration; ElevenLabs V3 alone is $0.215 per minute before tokens ($0.045 + $0.17). Reselling Unlimited removes that bet for covered products, because your cost is capped at $97 a month, subject to fair use, however much the client’s call volume spikes.
Frequently Asked Questions
The Voice AI allowance covers inbound calls, outbound calls and the Voice AI widget on both Growth and Unlimited. On Growth they share one pool of 100 minutes a month; on Unlimited they are included.
HighLevel does not publicly state a plan requirement for reselling Growth or Unlimited subscriptions. It does state that rebilling usage requires the $497 per month agency plan, so confirm the reselling requirement in Agency View before you quote a client.
Two calls of equal length differ in cost because the text-to-speech provider, AI model, token use and phone charges vary per call. HighLevel’s own example is a 10-minute call on OpenAI text-to-speech: $0.60 before tokens and phone charges ($0.45 voice engine plus $0.15 speech).
Generating up to four reply suggestions per review is free on pay-per-use, and the $0.01 charge applies only when you send a reply that uses an AI suggestion. A reply you substantially rewrite counts as manual and is not charged. Retrying or resending the same reply does not bill twice.