Constant Contact's AI Features: Included, Unmetered, and Undocumented

Constant Contact puts AI on every tier and tells you nothing about it. That is the whole finding, and it scores 2.9 on our rubric for exactly that reason, not because the features are weak, but because this axis measures how much a buyer can establish about what they are agreeing to, and here the answer is almost nothing.

constant contact ai draws 40 searches a month, which is small. The reason to write this anyway is that AI features are now a line in every comparison table, and the difference between vendors is no longer whether they have them.

What is actually included, by tier

The feature table names five AI capabilities and spreads them across the tiers ^[2].

An AI copy generator is on the entry tier, described as part of "Easy to use drag-and-drop email editor, AI copy generator, and email templates" on Lite, Standard and Premium alike ^[2]. So the drafting assistant is available at $12 a month, which is genuinely not true everywhere in this category.

AI content recommendations appear in the Email feature group ^[2]. Subject-line recommendations and subject-line personalisation sit alongside subject-line A/B testing, which places them from Standard upward ^[2].

An AI campaign builder is the headline AI item of the Standard tier, the vendor's own tier summary describes Standard as adding "3 automation templates, AI campaign builder, and automatic resend to non-openers" ^[2].

And the SMS editor includes AI generation, described in the vendor's documentation as part of what you can evaluate before subscribing: the full SMS suite "including campaign creation, AI generation, and message preview" is usable on any plan before buying the add-on ^[6].

That is a reasonable spread for the price. Nothing here is behind a separate purchase.

## Nothing is metered, and that is worth more than it sounds

The single most buyer-friendly fact about AI on this product is a negative one: there is no AI meter.

No credit system, no per-generation charge, no AI add-on on the pricing page, and no AI line in the two billing documents that explain everything else about how this product charges ^{[3] [4]}. The two meters that exist are email sends and SMS segments, and neither of them counts generations.

Set that against the rest of the category. Klaviyo publishes Composer at $20 a month with $5 of usage included and additional usage at $10 per 1,000 credits, and a Customer Agent at $50 a month for 75 resolved conversations. Those are clear terms, and they are also a third and fourth meter on an invoice that already has six.

On a product whose billing already carries one genuine trap, the contact band set by "the highest tiered number of active contacts you've had" ^[3], the absence of an AI meter is a real simplification rather than an omission. A buyer here cannot run up an AI bill, because there is no AI bill to run up.

That is the strongest thing we can say in this article, and it is the reason the axis is 2.9 rather than 2.0.

What is not published, which is the rest of the article

Now the reason the score is not higher, item by item. None of the following appears on any page we could read.

Which model. No provider is named, no version, no family. A buyer cannot tell whether their copy is drafted by a frontier model or a small one, and cannot tell when that changes.

What happens to the content. No statement about whether prompts or campaign content are used for training, retained, or sent to a third-party processor. For an organisation with a data-protection obligation, a charity holding donor records, a practice holding patient lists, that is not a detail.

Any usage ceiling. Nothing is metered, which is good, and nothing is stated about fair use, rate limits or throttling, which means the absence of a meter is a current fact rather than a promise.

Language and country coverage. A product sold in the United States, Australia and beyond, with an SMS feature restricted to two countries by tax identifier ^[5], publishes nothing about which languages its copy generator supports.

And no quality or provenance statement of any kind: no evaluation, no disclosure requirement, no guidance on reviewing output before sending to a list.

## How this compares, and why the gap matters

Klaviyo scores 4.2 on this axis and Constant Contact 2.9, and the difference is not about capability.

Klaviyo publishes rates, names connectors for ChatGPT and for Claude, and operates a documented MCP server, so a buyer can see which model sits where and what each interaction costs. It also leaves the model behind its own assistant unpublished, which is why it is 4.2 rather than higher.

Constant Contact publishes features. The gap of 1.3 points is entirely in specificity.

That matters for a concrete reason rather than a philosophical one. An AI feature whose terms are unpublished is an AI feature whose terms can change without anyone noticing, because there is no published baseline to compare against. The vendor has changed billing mechanics before and documented it, the send allowance and overage fee were introduced with an article that says plainly "we are implementing a monthly email send allowance and an overage fee if the allowance is exceeded" ^[4], dated and explained. The same vendor could introduce AI metering tomorrow and there is no prior statement it would contradict.

So the practical reading is not avoid the AI features. It is do not build a workflow on terms that have never been written down.

The feature that is not called AI and matters most

There is one capability here that does the work people expect from AI, and it is not marketed as AI at all.

"Resend to non-openers" is an automatic behavioural action: the system decides who did not engage and sends them a second copy without being asked. It is the Standard tier's headline feature ^[2], and it is the single most consequential automation in the product.

It is also the one thing in this product that spends your send allowance unpredictably, and the vendor says so in its own words: "Since you don't know how many contacts won't open your original email, you won't know for sure how many resends to non-openers there will be" ^[4]. Those resends count against the allowance while genuinely automated emails do not ^[4], and they land on a different day, so they can fall into a different billing cycle than the original ^[4].

The lesson generalises past this product. The automation that costs you money is rarely the one labelled AI. It is the one that decides how many messages to send.

If you turn it on, do it in a month with allowance headroom, and read your send count on the Billing tab rather than the Reporting tab, because the vendor states those two differ by construction ^[4].

## What the rivals charge, so the comparison is concrete

Worth laying the numbers out, because "AI is included" means different things at different prices.

Klaviyo sells its AI as products. Composer is $20 a month with $5 of usage included on the free tier and additional usage at $10 per 1,000 credits. Customer Agent, which resolves shopper conversations autonomously, is $50 a month for 75 resolved conversations. Neither appears on Klaviyo's pricing page; both are published in a billing support article. Those are two more meters on an invoice that already runs to six, and adding every published entry price takes a fully equipped Klaviyo account to $220 a month before any volume at all.

Constant Contact's equivalent line is zero. The copy generator, content recommendations, subject-line recommendations, the campaign builder and the SMS generation are inside $12, $35 or $80 ^[2].

The trade is legibility for simplicity, and which side you want depends on what you are buying. If you intend to run AI at volume, thousands of generations, autonomous conversation handling, you want published rates, because an unpriced feature at volume is a negotiation rather than a purchase. If you intend to use a drafting assistant a few times a week, you want it in the subscription and no meter to watch.

For most buyers of this category, the second is the realistic case, which is why the absence of a meter is scored as a genuine positive here rather than as a gap.

Why this axis is weighted at 0.083

A note on the scoring, because a 2.9 on AI reads worse than it is in the composite.

Our rubric weights axes by category, and for email the weights are value for money and capability at 0.229 each, ease of use and integrations at 0.146, AI and support and reliability at 0.083 each, and user sentiment and documentation at 0.042. AI is a tenth of the weight of the two leading axes combined.

That is deliberate and it reflects what this category is for. The job a buyer hires an email tool to do is reach a list reliably at a predictable cost; the drafting assistant is a convenience on top of it. A product with an excellent AI story and an unpredictable bill is the wrong purchase, and the weights say so.

In Constant Contact's composite of 3.63, the AI axis contributes 0.24. Raising it from 2.9 to a perfect 5 would add 0.17 to the total — less than a two-tenths improvement on value for money would add. Conversely the ratchet in the contact band, which costs this product most of its value score, is worth several times as much.

So read the 2.9 as what it is: a precise statement that nothing about these features is documented, on an axis that matters less here than almost anywhere else.

## What AI does not replace here

Two things, and both are reasons this product scores well elsewhere.

It does not replace the human on the phone. Live phone and chat with marketing experts is on every tier including the cheapest, along with live 1:1 onboarding ^[2], and in the hundred-review corpus support appears in 27 reviews and splits 18 positive to 8 negative ^[10] — the only product in this category where that theme splits rather than condemns. A five-star reviewer names the person who helped: "Shannon was a huge help in answering my question about contact teir" ^[10].

That combination is this product's actual differentiator, and it is the opposite of an AI story. Where rivals answer a pricing question with a chatbot, this one answers it with somebody who knows the contact-tier rule — which, given that the rule is published in one sentence in a community FAQ ^[3], is a sensible division of labour.

It also does not replace reading your own numbers. No AI feature in this product tells you which contact band you are in or what the next one costs, and those are the two facts that determine what you pay.

How to use what is there, safely

Four practical rules, given what is and is not published.

Use the copy generator for first drafts and treat the output as a draft. There is no published statement about model, provenance or review expectation, so the review is yours.

Do not put confidential or personal data in a prompt. In the absence of any published statement about retention or processing, that is the only defensible default — and it costs you nothing, because the generator's job is phrasing rather than analysis.

Use subject-line recommendations alongside A/B testing rather than instead of it. Both arrive at Standard ^[2], the test is the thing that produces evidence about your list, and the recommendation is a hypothesis.

Try the SMS AI before buying SMS. The vendor explicitly allows the full SMS feature set, AI generation included, without the subscription ^[6] — so draft real messages, then count their segments at "around 160 characters each, or 70 characters when using emoji" ^[5] before choosing a tier, because that product bills segments rather than messages ^[5].

## Five questions for support, and why each one matters

All five are answerable in the 1:1 onboarding session that comes with every tier ^[2], and none of them is answered on any page we could read. Ask them in that call and write down what you are told, because a verbal answer you recorded is better than a published statement that does not exist.

Which provider and model generates copy, and will I be told if it changes? The answer tells you whether output quality is a constant you can build process around, and the second half of the question matters more than the first.

Is anything I type into the generator retained, and is it used to train anything? If your organisation holds donor, patient, member or client data, this is the question your data-protection obligation turns on. A vendor that cannot answer it in a call is telling you to keep that data out of prompts.

Is there any ceiling, fair-use limit or throttle on generations? Nothing is metered today. Ask whether that is a commitment or a current state, because those are different purchases.

Which languages does the copy generator support? A product sold beyond the United States publishes nothing about this, and it is a yes-or-no answer that costs the person on the call ten seconds.

And the one that is not about AI at all: if I turn on resend to non-openers, how do I see what it cost me? The answer is the Billing tab rather than the Reporting tab, by the vendor's own documentation ^[4], and hearing it from support confirms you are looking in the right place.

If the call cannot answer the first four, that is not a reason to avoid the features. It is a reason to use them for drafting rather than for anything you would need to defend, and to treat today's absence of a meter as today's fact.

## The short version

AI here is included, unmetered and unspecified. The copy generator is on the $12 tier, content and subject-line recommendations and an AI campaign builder arrive with Standard, and the SMS editor has AI generation you can use before you subscribe.

Nothing is charged for it. On a product that already bills on a contact band set by your highest-ever count, the absence of an AI meter is a genuine simplification, and it is the reason this axis is 2.9 rather than lower.

Nothing is published about it either: no model, no provider, no retention or training statement, no ceiling, no language coverage. Klaviyo scores 4.2 on the same axis for publishing rates and naming connectors, and the 1.3-point gap is entirely specificity rather than capability.

The automation that actually costs you money is not labelled AI. It is resend to non-openers, it is the Standard tier's selling point, and the vendor states in writing that its volume cannot be forecast.