GoHighLevel AI Pricing Explained: Plans, Rates, Scenarios

"Are AI Employees an additional cost?" is one of the most common questions in GoHighLevel communities right now, and the confusion is understandable. HighLevel restructured its AI billing in May 2025, splitting the old single AI Employee price into two tiers and a pay-per-use option, then layered a summer promotion on top. The result is a pricing matrix that is fair once decoded, but genuinely hard to decode from the official docs alone.

This guide lays out every current rate, explains how the pieces stack, and runs worked scenarios so an agency can decide, sub-account by sub-account, whether to pay per use, buy Growth, or buy Unlimited. All figures below were verified against HighLevel's official AI pricing documentation in July 2026. Rates change, so always confirm against the source before committing budget.

The two layers of GoHighLevel AI pricing

Every GoHighLevel AI cost sits on top of a platform subscription. The base plans on the official pricing page are Starter at $97 per month with 3 sub-accounts, Unlimited at $297 per month, and Agency Pro (previously SaaS Pro) at $497 per month. None of these base plans includes unlimited AI usage by default.

AI billing is the second layer, and it is decided per sub-account, not per agency. Each sub-account (each client location, in agency terms) independently chooses one of three modes:

1. Pay-per-use. No monthly AI fee. Every AI action bills at published usage rates.
2. AI Employee Growth, $50 per month per sub-account. Includes 1,000 Conversation AI responses, 100 Voice AI minutes, and unlimited Reviews AI and Content AI. Usage beyond the included amounts bills automatically at pay-per-use rates without interrupting service.
3. AI Employee Unlimited, $97 per month per sub-account. Unlimited Conversation AI, Voice AI, Reviews AI, and Content AI, plus triple the Ask AI and AI Studio allowance of the Growth tier.

One naming trap deserves its own paragraph: the $297 platform plan and the $97 AI add-on are both called "Unlimited." They are unrelated products. A sub-account on the $97 Starter platform plan can buy AI Employee Unlimited, and an agency on the $297 Unlimited platform plan still pays separately for AI. When someone quotes "$97 for unlimited AI," they mean the add-on, not the platform plan.

For a full breakdown of the base plans themselves, see the GoHighLevel pricing guide.

What pay-per-use actually costs, tool by tool

Pay-per-use is the default when no AI Employee plan is active, and it is also what Growth overages bill at. The published rates as of July 2026:

Voice AI: three components per call

Voice AI is the tool most likely to surprise on a bill, because a single call stacks three separate meters:

• Voice engine: $0.045 per minute of call duration (effective May 20, 2026; per the official rate card, usage before that date was billed under an older bundled model).
• Text to speech: depends on the voice provider. OpenAI and Cartesia voices run $0.015 per minute, ElevenLabs V2.5 runs $0.035 per minute, and ElevenLabs V3 runs $0.170 per minute.
• LLM tokens: the model powering the conversation bills per million tokens, ranging from $0.05 to $3.00 per million depending on the model selected.

HighLevel's own worked example: a 10 minute call on the cheapest voice stack costs $0.60 in engine plus TTS charges, before token costs. On the cheapest configuration, a practical planning figure is roughly $0.06 per minute plus a small token cost. Choose an ElevenLabs V3 voice and the same call runs $0.215 per minute before tokens, more than three times the cost, purely from the voice choice.

Phone system charges (the underlying telephony) bill separately on every call, on every plan, including AI Employee Unlimited. The upcoming Stack Insiders breakdown of Voice AI economics will run deeper worked examples; for setup and a real test of the feature, see the Voice AI setup guide.

Conversation AI: token-based

Chat and SMS AI bills by model tokens. On ChatGPT-5, input tokens cost $1.25 per million and output tokens $10.00 per million. On ChatGPT-5 Mini, the rates drop to $0.25 and $2.00 per million. In practice, text conversations are cheap: short SMS exchanges measure in thousands of tokens, so even an active sub-account's monthly Conversation AI token bill on the Mini model typically lands in single digit dollars. The 1,000 included responses on Growth matter less for the token savings and more for predictability.

The rest of the AI lineup

• Reviews AI: $0.01 per review response.
• Content AI: $0.063 per generated image, $0.0945 per 1,000 words of text.
• Workflow AI Premium actions: $0.01 per execution, with bulk plans from $10 to $50 per month for high-volume automations.
• Funnel and Website AI: free, capped at 1,000 prompts per day per location.
• Agent Studio: pay-per-use on every plan. No AI Employee tier includes it, which surprises many buyers.
• Ask AI and AI Studio: bundled allowances with Growth and Unlimited, with Unlimited carrying triple the allowance.

Worked scenarios: which mode fits which sub-account

The right answer differs per sub-account, which is exactly why HighLevel bills it that way. Three common profiles, using the official rates above and the cheapest voice configuration:

Scenario 1: low-volume local client, pay-per-use wins

A client whose AI does light chat duty: a few hundred Conversation AI responses a month on ChatGPT-5 Mini, no Voice AI, occasional review replies. Token costs land under $5, review responses cost pennies, and there is no monthly fee. Total: well under $10 per month. Both AI Employee plans would lose money here. Pay-per-use exists precisely for this profile.

Scenario 2: active chat plus modest voice, Growth wins

A client running Conversation AI as a real front desk (approaching 1,000 responses) plus roughly 100 minutes of Voice AI answering. At pay-per-use rates the voice alone is about $6, and tokens stay cheap, so raw usage still prices below $50. What Growth buys at this volume is headroom and a flat, predictable line item that is easy to pass through to the client. Push meaningfully past the included 100 voice minutes, though, and overage charges start stacking on top of the $50, which is the signal to model the next tier.

Scenario 3: voice-heavy operation, Unlimited wins

A sub-account where Voice AI answers most inbound calls. On the cheapest voice stack at roughly $0.06 per minute, pay-per-use crosses the $97 mark at around 1,600 minutes per month, before token costs. That is about 53 minutes of AI calls per day. Any serious after-hours answering or missed-call recovery setup clears that easily, and on premium ElevenLabs V3 voices the breakeven collapses to roughly 450 minutes per month. For these accounts, $97 flat is the obvious buy, with only telephony charges variable.

The general rule: pay-per-use for pilots and light users, Growth for steady chat-first accounts, Unlimited the moment Voice AI becomes the product. Review each sub-account quarterly, because usage drifts.

The agency math: rebilling and the Agency Pro requirement

For agencies, AI pricing is not just a cost line, it is a margin opportunity. HighLevel lets agencies rebill AI usage to clients with a custom markup multiplier, or resell AI Employee plans at prices the agency sets. A client paying the agency $297 per month for an "AI receptionist" that costs the agency $97 in AI Employee Unlimited plus telephony is a standard arrangement in the ecosystem.

The catch: rebilling AI usage requires the agency to be on the $497 Agency Pro plan. Agencies on Starter or the $297 plan can pay for client AI usage but cannot mark it up through the platform's rebilling system. For an agency packaging AI services across even three or four clients, the rebilling margin alone can justify the plan upgrade.

Summer of AI: what is free right now, and the fine print

Through August 31, 2026, HighLevel's Summer of AI promotion makes five AI tools free: Ask AI, AI Studio, Workflow AI, Funnel AI, and Email AI. It also offers 30 day AI Agent trials, extendable to 60 days with an AI Employee plan. July's featured theme is Voice AI. Full mechanics are covered in the Summer of AI breakdown.

The promotion is a genuinely good testing window: run real volume through a trial sub-account, read the usage numbers, and buy the tier the data supports rather than guessing.

How to read your own AI bill

Three practical checks before and after enabling any AI tier:

1. Check the wallet, not just the plan. Usage charges draw from the sub-account wallet. Overages on Growth bill automatically without interrupting access, so the plan price is a floor, not a ceiling.
2. Isolate telephony. Voice AI charges and phone system charges are separate meters on the same call. When a "Voice AI bill" looks wrong, the telephony line is often the real driver.
3. Watch the voice provider setting. The gap between $0.015 and $0.170 per minute in TTS is the single largest controllable cost in the entire AI lineup. Premium voices are a deliberate quality upgrade, not a default anyone should inherit unknowingly.