ActiveCampaign AI: What Your Plan Includes, and What It Meters From Zero

ActiveCampaign puts AI at the top of its pricing page and meters it at the bottom of its billing catalogue. Both of those are true at once, and holding them together is the only way to answer what the AI here actually costs.

OUR TAKE: The published AI surface is the widest we have measured in any category. The AI that suggests is included in your plan. The AI that acts is an add-on on all four tiers, including Enterprise, and the meter behind it includes zero units. Price that before you weigh the feature list.

What is included

The first row of ActiveCampaign's feature matrix is Active Intelligence, and the vendor grades it by tier: "Unlimited access" on Enterprise and Professional, "Limited access" on Plus and Starter^[2].

Underneath it, the matrix names a long list. An AI-powered automation builder. An AI campaign builder and an AI campaign follow-up. Spam checking before a send. AI-suggested segments and suggested actions. An AI brand kit, five on the upper tiers and one on Starter. AI content generation and AI image generation. An AI campaign calendar. Account and industry-based send times.

Two more arrive at Professional and above: predictive sending, which the vendor describes as combining "AI and your customer data to find the optimal time to send your email", and predictive content. BotSense sits alongside them.

That is a genuinely broad published surface, and the vendor's own billing catalogue backs it: AI content generation, AI image generation and the AI brand kit all appear as named entitlements beginning at the Plus tier^[3].

## What is metered

Now the other half.

The pricing page lists an AI Activities add-on, described as "Supercharge your workflows with AI-driven automation and intelligence". Under it, two rows: AI translations and AI actions. Both are marked "Available add-on" in all four tier columns, Enterprise included^[2].

In the vendor's billing catalogue there is a matching entitlement named Activities. It carries zero included units on every one of the four tiers, a consumptive limit type, and a ninety-eight-step monthly price ladder that begins at a dollar a month for two hundred activities^[3].

The page never connects those two things or states a price for the add-on. We are reporting the correspondence of names, which is the vendor's own, rather than asserting that one is the billing line for the other. But the shape is unmistakable: the thing the page calls AI Activities is sold by the unit, and no plan includes any.

What that means for a buyer

Three practical consequences.

Your plan tier decides how much of the AI workspace you can touch, and the two lower tiers get "Limited access" to it. So half the ladder has a capped version of the headline feature.

Your add-on decides whether the AI can do anything on your behalf. Translations and actions are the two rows that would change a workflow rather than assist a person writing one, and both are outside every plan.

And your consumption decides the bill after that, at a rate published in the catalogue rather than on the page.

## The integration nobody else publishes

One line in the integrations accordion is worth pulling out, because no other product in our set lists it: "Claude & ChatGPT (via MCP)"^[2].

That is a connector to the model context protocol, which lets an assistant read and act on the account directly rather than through a prebuilt feature. If your AI plan involves your own assistant rather than the vendor's, that line matters more than the entire Active Intelligence section above it.

We have not tested it. We are noting that it is published, on a pricing page, as a named integration, which is itself unusual in October 2026.

What the reviews say, and it is not flattering

This is the axis where our cross-tabulation disagrees most sharply with the marketing.

Among the eighty-four English reviews we read, AI is a theme in 33% of the one and two-star reviews and 15% of the four and five-star ones, a gap of eighteen points^[17]. That is the second-largest gap in the whole table, behind only cancellation and contract.

One reviewer's description, which we quote because it is representative of the complaint rather than because it is a measurement, is that the vendor "recently implemented AI" by replacing interface pages with a text box. Whether that is fair is not something a feature list can settle. What the number says is that the AI push is landing badly with the unhappy end of this user base and only moderately with the happy end.

Set against that, the interface theme shows the same pattern: 21% at the bottom, zero at the top. The two may be the same complaint wearing different words.

## How to evaluate it in fourteen days

The trial is the only free window you get, and it has a specific ceiling: fourteen days, one user, features matching the Professional tier, and one hundred email sends^[14].

Professional-tier features means the trial includes predictive sending and predictive content, which the two lower tiers do not have. If you are planning to buy Starter or Plus, the trial is showing you more AI than you are about to pay for. Check which of the capabilities you liked are in the tier you intend to buy.

The hundred-send ceiling means you cannot evaluate send-time optimisation in any meaningful way. Predictive sending needs volume to predict with.

And the add-on rows will be marked available rather than active, so the actions and translations are not what you are testing.

What to ask before you pay for it

Four questions, none of which the pricing page answers.

What does the AI Activities add-on cost, in dollars, for the volume you expect? The catalogue's Activities ladder starts at a dollar a month for two hundred and runs ninety-eight steps, so the shape of the curve matters as much as the entry price.

What counts as one activity? The catalogue records the limit type as consumptive with a monthly period, which tells you it is a usage meter and not a capacity one, and nothing we could read defines the unit.

What does "Limited access" mean on Plus and Starter, in numbers? The page uses the phrase four times and quantifies it once, for brand kits, where Starter gets one against five.

And does predictive sending work at your volume? It is a Professional-tier entitlement^[3] and a prediction needs data.

## Where the AI sits in the tiers, concretely

The feature matrix grades AI by tier rather than listing it once, and the catalogue backs each grade with a named entitlement. Here is what that comes to.

On Starter, eleven entitlements in total^[3], and none of them is an AI capability. The matrix gives this tier "Limited access" to the Active Intelligence workspace and one brand kit against five above. There is no AI content generation and no AI image generation on this tier at all.

On Plus, AI content generation, AI image generation and the AI brand kit all appear as entitlements for the first time. The workspace is still marked limited.

On Professional, the workspace becomes unlimited and prediction arrives: predictive sending, predictive content, conditional content, and BotSense. Those are the capabilities that act on timing and content without a person choosing each time.

On Enterprise, the AI set is the same as Professional. What Enterprise adds is elsewhere: single sign-on, account activity logs, custom object schemas, and the Salesforce, Microsoft Dynamics and Zendesk integrations^[3].

So the AI ladder has one real step, and it is from Plus to Professional. If prediction is what you want, the entry two tiers do not have it, and the add-on rows do not change that.

The pricing shape, and why it is unusual

Most vendors we have measured put AI in one of two places. Either it is included everywhere as a differentiator, or it is a flat monthly add-on with a published price.

ActiveCampaign does neither. The suggesting half is included and tier-graded. The acting half is an add-on, marked available on every tier including the most expensive, with a consumption meter behind it that starts at zero on all four plans and climbs through ninety-eight published steps^[3].

That structure has one clear advantage for a buyer and one clear risk. The advantage is that you pay for what you use rather than for a flat fee you may not consume. The risk is that a usage meter with no included allowance has no floor to bump against, and the page that is supposed to tell you what things cost does not carry the rate.

It is also worth noting what this is not. The AI here is not gated behind the top tier, which is the pattern we have measured on several products in adjacent categories. Starter is thin on AI because Starter is thin, not because the AI is being held hostage to an upgrade.

Three things we could not establish

We are listing these because the absence is part of the answer.

What one activity is. The catalogue defines the meter as consumptive with a monthly period and sets its included units to zero, and nothing on any vendor page we read says whether one AI action is one activity, or whether a translation of a campaign into four languages is one or four.

What the add-on costs as a product. The page names it, the catalogue meters it, and no document we could reach joins the two with a price.

What "Limited access" permits. The phrase appears four times in the matrix and is quantified once, for brand kits. On the workspace itself it is unqualified.

All three are reasonable things to ask a salesperson, and all three are answers you should have before, rather than after, the fourteen days run out.

## The honest summary

ActiveCampaign's AI is broad, well-integrated into the rest of the product, and partly unpriced. The parts that help a person write are in your plan. The parts that act without a person are an add-on with no published price, metered from zero on every tier.

That is not a scandal. It is a pricing structure, and it is one you can only see from the vendor's billing catalogue rather than from the page that is meant to tell you what things cost.

If you are weighing this product against another on its AI alone, we would offer one piece of advice drawn from the measurement rather than from the marketing. The feature lists in this category have converged to the point where they no longer separate products, and the thing that still separates them is where the meter sits. Here it sits under the actions, at zero included, on every plan. Find out where it sits on the alternative before you compare the two lists of capabilities, because the lists will look much the same and the bills will not.