AI Chatbot Pricing Is Broken: Why Per-Resolution Billing Punishes Growing Stores

Picture this: your Shopify store is growing. Sales are up. Your AI chatbot is handling customer questions beautifully. You're finally sleeping through the night. Then you check your support software bill and it's doubled in three months.

Welcome to the world of per-resolution pricing — a billing model that literally punishes you for being successful. The better your AI performs, the more conversations it resolves, the higher your bill. It makes zero sense, yet it's how most major platforms price their AI features.

I'm going to break down exactly how this pricing works, why it's broken, and what to look for instead. Full disclosure: we built FunctionalAI specifically because we hated this model. But let's look at the math objectively.

How Per-Resolution Pricing Actually Works

Let's start with Intercom, since they're the most transparent about their AI pricing. They charge $0.99 per AI resolution. A "resolution" is when their AI handles a customer conversation without needing to hand off to a human.

Sounds reasonable at first, right? You only pay when it works. But here's where it gets ugly.

And that's just for 1,000 conversations. If you're running a decent Shopify store doing $50k-100k/month in revenue, you're probably getting 2,000-3,000+ support conversations. Now you're looking at $2,000-3,000 monthly just for AI resolutions.

Gorgias does something similar but calls them "automated interactions." Same concept, different label. You pay per automation that fires.

Why This Model Is Fundamentally Broken

Here's the thing that bothers me most: this pricing model creates a perverse incentive. The platform makes MORE money when your AI handles MORE conversations. But you — the customer — don't actually benefit from paying more.

Think about it:

• Your store grows → More customers reach out → AI handles more conversations → Your bill increases
• You improve your AI prompts → Higher resolution rate → More "successful" AI interactions → Your bill increases
• You expand to new markets → More conversations in different languages → AI handles them well → Your bill increases

You're literally being punished for success. The better your business does, the more you pay. The better the AI performs, the more you pay. It's backwards.

The Cost of "Free" AI Features

Some platforms advertise "AI included!" in their base plans. Sounds great until you read the fine print.

Usually it means:

• You get 50-100 AI resolutions/month "free"
• After that, it's $0.50-0.99 per additional resolution
• Or you're capped entirely and can't use AI for the rest of the month

So "AI included" really means "you get a tiny taste for free, then pay through the nose if you actually want to use it at scale."

What Fair AI Pricing Looks Like

Okay, so I promised I'd be honest about our bias. We built FunctionalAI specifically because we thought this per-resolution model was ridiculous. So here's how we approached it differently.

Our model: Flat monthly pricing with messages included. Free tier gets 75 messages/month. Paid plans at $24.99/month (400 messages) and $99.99/month (2,000 messages). If you go over, it's $0.05 per additional message.

I'm not saying we're perfect or that our model is the only right way. But I am saying: you shouldn't be penalized for success. Predictable pricing matters. If your store doubles in size, your support costs should be... well, roughly predictable. Not suddenly 10x.

Some other platforms are starting to get it. Tidio offers flat-rate plans with message caps. Zendesk has some all-inclusive tiers. It's not a FunctionalAI-vs-everyone situation — it's about the fundamental pricing model.

The Hidden Costs Nobody Talks About

Beyond per-resolution fees, watch out for these gotchas:

• Seat-based pricing on top of AI fees: You pay per agent seat PLUS per AI resolution. So you're paying twice — once for humans, once for AI.
• API call limits: Some platforms charge extra if your AI needs to look up order info or check inventory frequently.
• Training/setup fees: Enterprise platforms love to charge $5k-10k+ for "onboarding" and "AI training." Translation: they configure some settings for you.
• Storage overages: Upload too many help docs or product catalogs for your AI to reference? That'll be extra.

Always ask for a full breakdown of costs at YOUR expected volume. Don't settle for "it depends" or "we'll work with you on pricing." Get actual numbers in writing.

What to Ask Before Buying an AI Chatbot

Here's my checklist for evaluating AI chatbot pricing. Copy this and use it:

The Bottom Line

Per-resolution pricing made sense when AI was expensive and unreliable. Vendors needed to offset costs and risk. I get it.

But in 2026? AI costs have plummeted. OpenAI models are 90% cheaper than they were two years ago. Claude, Gemini, and other models are dirt cheap at scale. The infrastructure costs are negligible for SaaS platforms.

So why are vendors still charging $0.99 per resolution? Because they can. Because merchants don't do the math. Because "you only pay when it works!" sounds good in a sales pitch.

AI chatbots should help your business grow, not create another variable cost that scales painfully with success. Demand better pricing. Compare options carefully. And don't accept "industry standard" as justification for a model that doesn't serve you.