For our first eight months, we required a credit card at signup. Standard SaaS practice. The reasoning is well-known: a credit card requirement signals intent, filters out tire-kickers, and simplifies the upgrade flow. We believed it. We were wrong.
Today every new AIVA account starts with ₹500 of free credit and no card. This is the analysis that changed our minds, and the thing we rebuilt once we understood what we'd actually been measuring.
The argument for credit cards
The logic is seductive. Onboarding is expensive — it consumes infrastructure, support time, and setup effort. People who aren't serious enough to enter a credit card aren't worth that cost. And once you have a card on file, converting to paid is frictionless: one click, not a procurement cycle.
This logic is correct in some markets. It's wrong in ours.
Our buyer is a clinic manager, a salon owner, a coaching centre director. They are not evaluating twelve tools in a spreadsheet; they're deciding in one sitting whether the phone problem they've had for two years is solvable.
Asking that person for a card before they've heard the product speak is asking them to commit before you've made an argument.
The broader research on checkout friction points the same way — Baymard's abandonment studies consistently find that payment-step requests are among the highest-drop-off moments in any funnel, and ours was no exception.
What the data said
In month eight, we ran a retrospective on signup cohorts. The question was simple: what best predicts whether a new account becomes a paying customer?
The answer was not whether they'd entered a credit card. It was whether they'd completed first setup — connected AIVA to a real data source and had a real conversation with it.
| Cohort | Converts to paid |
|---|---|
| Completed first configuration | 68% |
| Did not complete configuration | 4% |
| Entered a credit card | no measurable effect |
The credit card was completely uncorrelated with conversion. The implication: the bottleneck wasn't intent, it was activation. People who signed up with a card weren't more likely to convert. They were equally likely to activate — or not — based on factors that had nothing to do with payment information.
Meanwhile, the card requirement was suppressing signup volume. We ran a three-week A/B test removing it. Signups increased 340%.
The maths
After removing the credit card requirement:
- New signups: +340%
- Signup-to-paid conversion rate: 28%, down from 31%
- Net paying customers acquired per month: substantially up
The conversion rate dropped slightly — some of those new signups were genuinely casual. But the net paying customers acquired per month increased sharply, because 340% more people entered the funnel and the overwhelming majority of them still converted at the old rate.
More importantly: customers who converted without a card on file were just as likely to stay paying as those who'd had one. Churn rates were identical. The "quality filter" the card was supposed to provide didn't exist.
We were solving for the wrong problem. The friction we needed to remove wasn't in the payment flow — it was in the activation flow.
Why credit is a better instrument than a trial
Removing the card left us a design question: what replaces it? The default answer in SaaS is a 14-day trial. We deliberately didn't do that.
A time-boxed trial optimises for the vendor's urgency, not the customer's evaluation. A salon owner who wants to test the agent across a full weekend rush shouldn't be racing a countdown that started the day they signed up.
Worse, a trial clock punishes exactly the careful buyers you most want — the ones who plan a proper pilot instead of clicking around for ten minutes.
Credit doesn't have that problem. ₹500 buys roughly 125 minutes of voice, or 250 web chats, or 500 SMS, and it sits there until you use it.
The comparison is spelled out in more detail in what ₹500 of free credit actually gets you, and the same reasoning drives the rest of our model — see pay-as-you-go versus subscription pricing and how we priced voice per minute.
It also sets up an honest failure mode. When credit runs out, the agent stops rather than quietly billing a card nobody remembers adding. Auto-recharge exists for teams that don't want that interruption, and it's off by default.
What we built instead
Removing the card requirement was easy. What came next was harder.
If activation is the real conversion moment, onboarding has to be engineered around reaching it fast. So we rebuilt it: pre-built setup templates by business type, a wizard that connects a data source in four steps, and an email sequence that nudges accounts that haven't completed configuration after 48 hours.
First-configuration rate went from 34% to 61% in six months. That 27-point improvement in activation did more for our growth than any other change we've made — and it's the same finding that showed up independently in what we learned from our first 100 customers, where every fast-succeeding customer had connected AIVA to live systems rather than a FAQ document.
Setup now typically takes an afternoon rather than a project; how long an AI receptionist takes to set up has the honest breakdown.
How to run this test on your own funnel
The specific answer here is ours. The method is portable, and it's cheap enough that there's no good reason not to run it.
Find your activation event first. Not your signup, not your upgrade — the single action that separates accounts that convert from accounts that don't. Ours was completing a first configuration against a real data source. Yours will be something similarly unglamorous and specific.
Segment your last six months of signups by every candidate action and look for the one where the conversion gap is enormous rather than merely visible. A 68%-vs-4% split is the shape you're looking for; a 40%-vs-30% split means you've found a correlate, not a cause.
Then check what your friction is actually filtering. We assumed the card filtered for intent. When we compared cohorts, card-havers and card-less accounts activated at indistinguishable rates.
The filter was catching volume, not seriousness — which is the failure mode of most signup friction, because the people it deters are disproportionately the ones who haven't yet been convinced, and those are exactly the ones you were hoping to convince.
Run it as a real A/B test, for at least three weeks. Shorter than that and you're measuring a novelty spike. And measure downstream churn, not just conversion — the whole argument for the card is that it improves customer quality, so if quality is unchanged the argument has no remaining leg.
Be prepared for the conversion rate to fall. Ours dropped three points. That number is the one that scares teams out of the change, and it's the wrong number to optimise.
What matters is net paying customers per month, and a small percentage of a much larger denominator won ours comfortably.
The lesson
Credit card requirements are a solution to a problem — intent filtering — that most products don't actually have. The problem is activation. Optimise activation, not intent signals.
If you want to test that claim on your own phone line, start free with ₹500 of credit — no card — and see what the pricing looks like only once you've decided it works.
If you're not sure what to test first, our free-trial checklist is a good place to start.