AI SDR contract terms that raise your bill in month 4

TL;DR

The quoted monthly price only describes month one. Rollover, top-ups, expiring promos, quote-only tiers and duplicate price lists decide the rest.

ai-sdrpricingcontractsrevops

Five clauses change an AI SDR bill after month three: how unused credits expire, what a top-up costs, when a promo discount ends, which features sit behind "talk to sales", and which of two published prices is the contract price. Price each one before you sign, because the number in the demo only describes month one.

I work through order forms for a living, so this is how I read them. It is not a negotiation script. The clauses are where the money moves, and you need to know which ones to ask about.

Which ai sdr contract terms change the price at month three?

The pattern is the same across vendors. The first invoice matches the pricing page. Somewhere between month three and month five, one of the following kicks in, and the invoice stops matching.

  • Credit expiry: unused credits vanish at the end of the billing period, or roll over for a limited time.
  • Top-up pricing: credits bought mid-month cost more than credits in your plan.
  • Promo expiry: an introductory discount ends and the list price resumes.
  • Quote-only tiers: the feature you need turns out to live in a custom plan.
  • Duplicate price lists: the pricing page, the FAQ and the order form disagree.

To spot them in an order form, look for four things: the credit allotment per period, the unit price for extra credits, the discount end date, and the plan name. If any of the four is missing, the clause lives in a linked terms page, and that page can change.

Credit rollover decides whether you pay for idle months

Outbound volume is lumpy. You run a big campaign in one month, then spend two weeks fixing deliverability and sending nothing. Credits that expire at the end of the period mean you paid for that quiet stretch and got nothing back.

One credit-metered vendor just added a one-month rollover on its team plan, which tells me buyers are pushing on this. One month is a start. If your cycle has a quarterly rhythm, a one-month window still burns credits in the slow quarter. Ask what rolls, for how long, and whether top-up credits roll too. Often they don't.

The top-up premium is where growth gets taxed

Top-ups matter most when the agent works. A rep sees good replies, asks for more volume, and the plan runs out on day 18.

A worked example, with assumed numbers: a plan costs $100 for 10,000 credits, so a credit costs $0.01. You burn 13,000 in month four. At a 30% top-up premium, which one workflow-automation vendor publishes on two of its tiers, the extra 3,000 credits cost $39 instead of $30. Small at this scale. Now run the same ratio on a team spending $5,000 a month, and the premium becomes a line item your finance lead notices.

The reason I ask for a written top-up rate card is so that the premium is a number in the contract and not a surprise in the invoice. For better or worse, vendors rarely volunteer it.

Expiring promos make the entry price a fiction

Right now one vendor is running 50% off on every volume band, cutting an entry band from $300 to $150 a month. The offer covers the first two months. Its countdown timer and its structured data disagree on the end date by four days, and the FAQ still shows $300.

Do the arithmetic as a 12-month average, which is my assumption for the example: two months at $150 plus ten at $300 is $3,300, or $275 a month. The anchor in your head says $150. The real number is 83% higher. Budget against the post-promo price, and treat the discount as a one-time credit.

Quote-only tiers and the two-number pricing page

Two patterns here. The first is the fully quote-only vendor, where cost is tied to leads contacted. You can't estimate anything until a sales call, and when your volume grows you don't know what happens to the rate. The second is the cheap-looking per-seat entry price that holds back CRM write-back or signal-triggered automation for a custom annual tier. If the feature is the reason you're buying, the entry price isn't the price.

Then there is the page that shows two numbers. One vendor currently lists a $49.99 team plan while its structured data lists a $29.99 to $199.99 ladder. $49.99 against $29.99 is a 67% jump on the entry price. Maybe it's an experiment, maybe it's a bug. I don't have a way to know which arm you'll land in, and neither do you.

My recommendation: the order form must name the plan, the monthly price, the credit allotment and the top-up rate, all in one place. Screenshot the page the day you decide. If the page and the form disagree, the form wins, so make sure the form is the one you like.

How to model twelve months before you sign

Take your real activity volume, not the demo's. Then run three cases: current volume, double volume, and a quiet quarter. For each, add the plan, any top-ups at the premium rate, the promo's end date, and any feature you need from a higher tier. Our post on forecasting a credit-based bill by cost per completed output walks through the unit-cost part, and the credits versus seats explainer covers why the meters behave differently.

Then add the people cost, which no order form shows. Someone reviews what the agent sends. The comparison in AI SDR versus human SDR cost treats those supervision hours as a line item, and I agree with that framing. A cheap tool that needs a lot of review isn't cheap.

Two asks I'd make, with an honest caveat. First, a cap on the renewal increase. I don't have a great answer on what cap is realistic, since I haven't seen enough signed renewals to quote a number, so ask for one and see what you get. Second, an exit or credit-back right if the vendor misses a deliverability or meeting target. Vendors tend to resist this, but asking tells you how much they believe their own claims. The cost of a bad send is also real, which is why brand risk in a supervised AI SDR belongs in the same conversation.

Where we land on this at Coco

We price in credits, and I'd rather you check us on the same clauses. The published tiers are a free hobby tier with your own API keys after beta approval, a $40 a month founder tier with 5,000 credits, and custom team pricing. Credit estimates usually land within about 20% of actual cost, and the pricing estimator lets you test your own volume. Every spend goes through an approval gate, so a top-up isn't a surprise on the invoice. It's a step you approved.

If you're still deciding between an autonomous tool and a supervised co-worker, what proof to demand before buying is the next read. Whichever way you go, read the ai sdr contract terms first, then the demo. The clauses decide your bill in month four, and the demo never mentions them.

Sources

Get started

Put an AI co-worker on your GTM work

Coco researches, drafts and executes across your GTM stack, and asks before anything goes out.