A complete guide to Salesforce Flex Credits in 2025

Salesforce has shaken things up with its AI pricing. They've moved away from the straightforward (but often debated) flat-rate model for their Agentforce platform. Instead of a single price per conversation, they're now using a more detailed, pay-as-you-go system called Salesforce Flex Credits.

This change is all about giving companies more flexibility and connecting costs directly to what the AI actually accomplishes. But while this solves some old problems, it brings up a few new ones, especially for customer support teams who really need a predictable budget.

So, is this new model the right fit for you? This guide will walk you through it. We’ll cover what Flex Credits are, how they work, the full pricing details, and an honest look at the pros and cons you should be thinking about.

What are Salesforce Flex Credits?

Salesforce Flex Credits are essentially a pay-as-you-go currency for the Agentforce platform. Instead of paying a flat fee for every conversation your AI has, you now pay for the specific "actions" it performs.

The old way was a bit like paying a cover charge to enter a store, no matter what you bought. A simple question like "What are your hours?" cost you the same as a long, complicated troubleshooting session. It was simple, but it didn't always make sense.

The Flex Credits system changes that. Costs are now linked to the individual tasks the AI handles, whether that’s updating a customer record, pulling an answer from your knowledge base, or scheduling an appointment. According to Salesforce's official announcement, the goal is to better match your AI spending with the work being done, making it a system based on value.

How Salesforce Flex Credits work

So, how does this all work in practice? Understanding how you buy, spend, and track these credits is the key to figuring out if this model makes sense for your team.

The basics of the Salesforce Flex Credits system

On the surface, it's pretty simple. You buy credits in bulk, they're sold in packs of 100,000 for $500, and then your AI agents "spend" them as they complete tasks. A typical AI action in your live environment uses 20 Flex Credits, which comes out to $0.10 per action.

The rate can change depending on where the action happens; for instance, a test in a sandbox environment only costs 16 credits.

To give you a better idea, here’s what credit usage could look like for a few common support scenarios, based on examples from Salesforce:

Use Case	Actions Required	Number of Actions	Flex Credits Consumed	Total Cost (USD)
Simple Case Management	Identify Customer, Get Cases, Add Comment	3	60	$0.30
Field Service Appointment	Identify Customer, Get Work Types, Get Time Slots (x2), Answer with Knowledge	6	120	$0.60
Employee Onboarding Query	Answer Question with Knowledge	1	20	$0.10

Managing and monitoring your Salesforce Flex Credits

With a pay-as-you-go model, you obviously need a good way to keep an eye on your spending. Salesforce provides the Digital Wallet for this. It's your main hub for tracking your credit balance. It gives you a near real-time look at your usage, sends you email alerts when you're running low, and lets you see everything across both your live and sandbox environments in one place.

It even has a "Usage Tagging" feature, so you can assign costs to specific agents or departments to get a clearer picture of your ROI.

While these tools give you a visibility, it's important to remember that forecasting your future spending can still be a big challenge. Even with great tracking tools, the unpredictable nature of customer support means it can be tough to set a stable, predictable budget.

The Flex Agreement and other ways to buy Salesforce Flex Credits

Salesforce also rolled out the Flex Agreement, which is a pretty clever idea. It lets companies convert their existing user licenses into Flex Credits, or the other way around. This gives you the freedom to move your budget between your human agents and your "digital labor" as your needs change, without getting stuck in a rigid contract.

On top of that, there are a few other ways to get started with Agentforce:

• Agentforce User Licenses: For a monthly fee per user, you can give your employees metered access to internal AI agents.

• Agentforce 1 Editions: These are premium packages that come with a big bundle of Flex Credits included for the year.

• Future Models: Pay-as-you-go and pre-commit plans are apparently on the roadmap for later in 2025.

Salesforce Flex Credits pricing explained

With a few different ways to buy, it helps to see all the pricing options in one spot. Here’s a full breakdown of the Agentforce pricing structure.

Product/Model	Price	Key Features	Best For
Flex Credits	$500 / 100,000 credits	Pay-per-action ($0.10/action). Managed via Digital Wallet.	Teams wanting granular, usage-based pricing for varied use cases.
Agentforce User License	$5 / user / month	Metered access for all employees to use AI agents. Requires Flex Credits.	Companies deploying employee-facing agents company-wide.
Agentforce Add-ons	$125 / user / month	Unmetered AI usage for employees in Sales, Service, etc.	Departments wanting predictable costs for internal-facing AI agents.
Agentforce Industries Add-ons	$150 / user / month	Same as above, but with industry-specific AI capabilities.	Regulated or specialized industries (e.g., Finance, Health).
Agentforce 1 Editions	from $550 / user / month	All-in-one bundle including the add-on, 1M Flex Credits/year, and more.	Enterprises looking for a comprehensive, premium AI package.

It's also worth mentioning the Salesforce Foundations offer. If you're an existing customer on an Enterprise Edition or higher plan, you can get 100,000 Flex Credits for free to help you get started.

The pros and cons of Salesforce Flex Credits for support teams

So, what does all this really mean for a support leader trying to figure out if this is the right move? Let's break down the good and the not-so-good.

The good: Flexibility and better value

You get more control over costs. The biggest win here is that simple tasks cost less than complex ones. This helps line up what you're spending with the actual work your AI is doing, which just feels right.

It scales better across different departments. If you're planning to use AI for more than just customer service (like for IT, HR, or sales), this model can grow with you more effectively than a one-size-fits-all price per conversation.

It allows for strategic shifts. For larger companies, the Flex Agreement is a big deal. Being able to move budget between software licenses and AI usage gives you a lot of freedom to adjust your strategy without being punished by your contract.

The tricky part: Complexity and budget surprises

Budgeting can become a real headache. This is probably the biggest drawback. Pay-as-you-go pricing makes forecasting your costs tough. How can you accurately predict next month's support ticket volume? A sudden surge in customer issues could leave you with a surprisingly large bill, which is exactly the kind of uncertainty finance teams hate. For teams that need that predictability, some solutions like eesel AI offer a compelling alternative with transparent, tier-based pricing. You pay a flat monthly fee for a set volume of interactions, which gets rid of the risk of a surprise bill after a busy month.

Setup can be a heavy lift. As some analysts have noted, the pricing is only one piece of the puzzle. For many companies, the bigger challenge is getting their data in order. Salesforce can be incredibly complex, and making sure your data is clean and organized enough for an AI agent to use effectively can turn into a massive project. This is where tools like eesel AI's AI Agent try to offer a different path. It's built for a quick setup, with one-click integrations for helpdesks like Zendesk and knowledge bases like Confluence, so it can start learning from your existing information in minutes, not months.

The entry point might be too high for some. A minimum purchase of $500 for a block of credits could be a hurdle for smaller teams or for those who just want to run a small-scale test before they commit a real budget. In contrast, eesel AI has a free simulation mode that can help. You can test your AI on thousands of your own past tickets to get a solid forecast of its performance and automation rate before you spend anything. This lets you build a business case and roll out automation with confidence.

This video provides a detailed breakdown of the new Salesforce Flex Credits model, including what a Flex Agreement is and how the credit system works.

Are Salesforce Flex Credits right for you?

So, what's the verdict? Salesforce Flex Credits are definitely a more logical and flexible way to price AI than the old flat-rate system. Tying costs to the actual value you get is a smart move, but it’s not without its own set of complications.

This new model seems like a great fit for large companies that already have solid data practices and the resources to manage a consumption-based budget. The flexibility to shift investments will be a huge plus for them.

But for teams that need predictable costs, a quick setup, and an easy way to get started, the complexity and uncertainty of this model might be a dealbreaker. The last thing anyone wants is to spend months on a complicated project only to get hit with an unpredictable bill.

If you're looking for a powerful AI support solution that's simple to set up, has predictable pricing, and lets you prove its value risk-free, you might want to explore eesel AI's AI Agent. You can simulate its performance on your own data and see what it can do in just a few minutes.