Splitting the Bill: Why We're Breaking Up Our Credit System

You know that moment when you're trying to explain your phone bill to someone and you realize even you don't fully understand it? That's kind of how we felt about our credit system. And if we're confused by our own pricing, that's... not great.

So we're changing things, because we believe users benefit most when it’s clear what they’re paying for and why.

TL;DR:

• We're splitting Credits into Actions (what your agents do) and Vendor Credits (AI model costs)
• No markup on Vendor Credits—we pass through exact costs
• Vendor Credits now roll over indefinitely while you're subscribed to a paid plan
• Use any LLM models to power your agents with Vendor Credits, or bring your own API keys to bypass Vendor Credits entirely

The Old Way: Credits for Everything

When we started, credits for everything seemed like the perfect solution. One simple unit to rule them all—whether you were testing new models from Claude, getting agents to use the latest GPT, scraping websites, or having your agents do complex workflows. 

It felt simple and flexible… until we realized it actually created confusion about what you were really paying for.

The simple question, "How many credits will this workflow use?" became a source of friction. Instead of building freely, you were building spreadsheets. Instead of deploying with confidence, you were crossing your fingers. 

The New Way: Splitting the Bill with Actions + Vendor Credits

Starting today, we're evolving our single credit system into two crystal-clear components:

1. 🎯 Actions - What your agent does (e.g. sending an email, updating CRM).
2. 💳 Vendor Credits - The AI model usage cost (e.g. OpenAI writing text).

Think of it like an arcade game: you get Actions is the number of moves you get and Vendor Credits powers each move.

What's an Action?

An Action is one unit of work an agent performs. Whether it’s a simple task or a multi-step workflow, if it runs as one tool, it counts as one Action.

• A tool that sends a single email → 1 Action
• A tool that updates your CRM → 1 Action
• A tool that enriches lead data, scores it, and routes it to the right salesperson, updates your CRM and sends an email → still 1 Action

The beauty? Usage becomes predictable, so you can see how work gets done and plan with confidence.

What Are Vendor Credits?

Vendor Credits are simply prepaid AI dollars you can use on any of the latest AI models we support.

• No markup: You pay the model’s exact cost. Nothing added, nothing hidden.
• Never wasted: Unused credits roll over every month, as long as you stay subscribed on a paid plan.
• Flexible: Use them however you like, or connect your own API keys and skip Vendor Credits altogether.

The goal is simple: you always know what you’re paying for, and you never lose value.

How It Works in Practice

Let’s take a Sales Agent working on a new lead:

• Actions: Enrich contact data → Score lead → Update CRM → Send follow-up email = 4 Actions
• Vendor Credits: Used when the agent works with your instruction prompts, data inputs like details of the lead & writing the follow up email.

👉 Before: One big credit cost with no breakdown. You never knew how much came from workflow steps, model usage, input vs output token consumption.
👉 Now: Clear split — you can see 2 clear metrics - the work done (Actions) vs. LLM usage (Vendor Credits).

‍

Why This Actually Matters

This shift means fewer surprises, more control, and lower costs as models improve. Instead of guessing how many credits a workflow will burn, you’ll see exactly what’s an Action and what’s an AI cost. And as new, cheaper, or more efficient models arrive, those savings flow directly to you.

This will allow you to optimize for each:

• Want to save on AI costs? Use a smaller model for simple emails. 
• Want to reduce Actions? Combine steps into a single workflow tool.

What This Means for You

If you're an existing customer: 

• Your current credits will automatically convert into the new two-part system: Actions and Vendor Credits (LLM costs).
• Your workflows keep running exactly as before, no changes needed.
• Your Credits will now go further - for the same agent task as before, expect your credit used to fall as we remove mark ups & passthrough costs as is.

If you’re new to Relevance:

• From day one, you’ll see exactly what you’re paying for—clearly split between Actions and Vendor Credits.
• Start with our credits to try out any model. Later, if you already have subscriptions to LLM providers like OpenAI or Claude, you can connect those accounts directly (by adding your own API keys).

For the power users: 

• Connect your own API keys to bypass Vendor Credits entirely. 
• Use our credits only when you want to test new models you don’t already subscribe to.
• Monitor costs directly with your LLM provider for maximum control.

A Closing Note

Our goal is to make it easy for you to understand and control your costs. When you can see exactly what’s happening — and make changes as models evolve — you’re free to focus on building, not budgeting.

Like you, we’re builders. We believe the future belongs to those who can orchestrate AI agents to do meaningful work. Our role is to make that orchestration simple, predictable, and powerful.

This pricing change won’t grab headlines, but it will give you clarity. Most importantly, it aligns our focus to help you deliver real work using agents & we’re actively incentivized to help you save on costs. We succeed only when your AI workforce delivers value.

FAQ

Q: What happens to my existing credits?
All your credits convert into the new system. They’ll show up as two line items — Actions and Vendor Credits. As an existing paid customer, this will only take into effect in your next billing cycle after 1 Dec 2025. Read more about it here.

Q: Do credits still expire?
No. Vendor Credits roll over indefinitely as long as your subscription on a paid plan is active. Actions are tied to your plan’s monthly allowance & resets as per your billing cycle annually or monthly. 

Q: What if I already pay OpenAI or Anthropic directly?
You can connect those accounts to Relevance and bypass Vendor Credits. (Technically, this means adding your own API keys.

Q: How do I save money under this new system?
Existing LLM model costs have fallen dramatically over the last 24 months. For many everyday tasks, you can use smaller models for lightweight tasks.
‍
Frontier models with thinking can still be expensive, but have powerful agentic capabilities. Be selective on which agents need that power & critically assess the value of the work done by the agent. 

To save on Actions, combine multiple tool steps into a single Action. Since Actions are consumed when a tool is used, these can be simple steps or complex workflows, you have the flexibility to design & optimize. 

Q: Will my workflows break?
No. Everything you’ve built continues to run the same way.

Q: Why make this change now?
As models get cheaper and more efficient, you’ll automatically benefit. We wanted our pricing to move with the industry, not against it. It aligns our focus to help you build powerful & efficient agents as we will only make money when your AI Workforce is productive doing a ton of Actions, and incentivises us to help you reduce LLM costs.

Q: I still have questions about how this works?
Please refer to our support documentation here on all the details the new plan has to offer.