I pivoted my AI secretary from subscriptions to usage-based pricing. Here's why.

Two weeks ago, I was ready to launch irel.ai with a simple SaaS pricing model: $18/mo annual, $30/mo monthly, or $349 lifetime.

Then I had a conversation that changed everything.

A founder I respect (who's built multiple successful products) looked at my model and said: "You're leaving money on the table and confusing your value proposition."

The problem? I was charging for access, not outcomes.

Here's what he pointed out:

The subscription trap: Users pay $30/month whether they delegate 5 tasks or 500. Low-usage users feel ripped off. High-usage users become unprofitable.

The real value: My AI secretary doesn't just chat—it does things. It labels emails automatically. It reminds you of birthdays before they happen. It generates daily briefs so you start your day informed.

The pivot: We moved to a base fee + pay-per-job model. Think AWS, not Netflix. You pay for what you actually use.

$49/month base gets you access to irel on WhatsApp. Then you pay per "Job" you hire her for:

• Email triage: $X/month
• Birthday radar: $Y/month
• Daily briefings: $Z/month

Why this matters:

1. Expansion revenue: As users trust irel with more tasks, revenue grows naturally
2. Alignment: We only make money when we deliver real value
3. Lower barrier: $49 is easier to justify than $100+ for a bundle you barely use

The irony? This model is harder to build. We had to architect true "Jobs"—autonomous delegations that run without prompting, not just chat responses.

But that's exactly what makes it defensible.

Anyone can build a ChatGPT wrapper. Building an AI that actually does things reliably? That's the moat.

We're launching in March. First 10 customers get grandfathered pricing.

What do you think—would you rather pay for access or outcomes?