Insights

What happens to my automation if I cancel the subscription?

In most cases, the automation stops that day and everything it was doing stays behind on the vendor's platform. The workflows, the configurations, the customer history it collected — those were never yours to take. What you owned was the login.

It's worth walking through the actual exit, because almost nobody does it before signing up. Cancellation day: the missed-call texts stop sending. The follow-up sequences stop firing. The data — every customer interaction the system logged — sits in an account you can no longer open. Some platforms let you export a file, and a file is better than nothing, but data without the workflow that used it is a filing cabinet without the office.

That was the short walk. Here's the long one — the full inventory of what stops, what leaves with you, and the before-you-sign homework that turns cancellation day from an amputation into an errand.

What exactly stops working, and how fast?

Trace one customer through your rented system and count the touchpoints. A call comes in, rings out, the system texts back — that's touchpoint one, gone at cancellation. The reply gets read and answered — two, gone. The lead gets logged — three. The estimate follow-up fires on day three and day seven — four and five. The review request goes out after the job — six. One customer, six touches, all of them stopping the same afternoon, simultaneously.

Speed varies by vendor and it's checkable in the terms: some cut access the day you cancel, some run out the billing period, a few offer a read-only grace window. What almost none do is keep anything running. And there's a subtler timing trap on the way in: annual contracts in a category where the tools themselves turn over quarterly mean the exit door only opens once a year, whatever your business needs in month four.

What actually leaves with me?

Four items to check, in descending order of how often they survive the exit.

The contact list — usually exportable, usually the headline of the vendor's "your data is yours!" claim. Read what the export actually contains. The activity history — the conversations, the follow-up records, which customer was texted what and when. This is the operational memory of your business, and it survives far less often than the contact list; a bare list of names without the history is a phone book, not a relationship. The phone number — the sleeper issue of the whole category. If your textback system ran on a number the vendor provisioned, every customer who saved that number is now texting into a dead line — or worse, into a line the vendor reassigns. Numbers you brought are generally portable by right; numbers they provisioned are a negotiation. The workflow itself — the logic, the timing, the wording, the routing rules. This one essentially never leaves. It was the product. Rebuilding it elsewhere means starting from a memory of how it behaved.

Run those four against any tool you currently pay for. It's the same audit that tells you whether you're renting — ten minutes now versus discovering the answers on the day they're unfixable. And if the audit comes back ugly, that's not automatically a reason to cancel; it's a reason to start exporting what you can on a regular schedule, so the day you do leave, you leave with something.

What's the before-you-sign version of this homework?

Five questions, asked while you still have leverage — vendors answer differently before the sale than after. Show me the export: which file formats, and does it include conversation history? Whose number will this run on — can I bring mine, and can I port yours out later, in writing? What happens to my configuration if I cancel — can I get a written copy of my own workflows, wording, and timing rules? How long after cancellation does my access last? If you shut down or get acquired, what's the wind-down commitment? That last one isn't paranoia; in a category this young, vendors churn as fast as their tools. And get the answers in writing — an email reply is fine. Sales-call answers about exits have a way of differing from the terms of service, and the version that governs cancellation day is never the spoken one.

You'll notice something about the answers: they cluster. Vendors happy to put all five in writing are selling you a service. Vendors who go vague are selling you a wall around your own operations. Either can be worth buying — renting has honest uses, especially for experiments and edge functions — but you deserve to know which purchase you're making at the price you're paying, because the two are priced very differently when you count the exit.

None of this is a scam. It's the subscription model working exactly as designed — you were paying for access, and access is what ends. The mistake isn't subscribing; plenty of businesses rent tools sensibly. The mistake is building your operations around something you rent without knowing you're renting it.

So ask the exit question before you're in: if I stop paying tomorrow, what still works, and what walks out the door? If the answer is "nothing still works," you've priced that subscription wrong. The alternative worth knowing exists: a system built for your business where the process and the data stay yours, no matter what sits underneath — and "yours" has three specific, checkable tests.

Which of your current tools would survive their own cancellation day — and which functions of your business are too close to the spine to leave hostage — that's the sorting that pays, and it's the work I do. Run the four-item check; then bring me what you find: book a conversation.

← All insights

Got a bottleneck in mind?