Introduction
Salesforce has jumped into generative AI with offerings like Einstein GPT and the newer Einstein Copilot (now evolving into Agentforce). These promise to boost productivity across CRM functions using AI-generated content and insights.
However, they come at a significant cost and with complex licensing conditions. CIOs, CFOs, and procurement leaders must carefully scrutinize these pricing models.
The enthusiasm around AI is high, but so are the price tags and prerequisites associated with Salesforce’s AI features. Read our overview of Salesforce AI & Automation Licensing.
Salesforce’s strategy is to monetize generative AI as premium add-ons or upgraded editions on top of its core cloud products. Unlike basic CRM features, the AI capabilities are not simply included for free – they require additional investment.
This means organizations need to budget for AI features separately and understand exactly what they are paying for. Below is a breakdown of the pricing and structure for Einstein GPT, Einstein Copilot, and the overall AI Cloud package, along with key considerations for enterprise buyers.
Einstein GPT Add-Ons and Limited Credits
Salesforce Einstein GPT was first introduced in early 2023 as an add-on feature to existing clouds, rather than a standalone product.
These add-ons enable generative AI in Sales Cloud and Service Cloud through features branded as Sales GPT and Service GPT.
Each comes with a per-user monthly fee and an included allotment of “Einstein GPT” usage credits.
- Sales GPT – Priced at $50 per user per month, this add-on is part of the Sales Cloud Einstein offering. It enables sales teams to auto-generate emails, call summaries, and other sales-related activities with the aid of AI. The $50 user fee includes a limited number of GPT usage credits for generating content. If users exceed the included credit quota (for example, through very frequent AI-generated emails or notes), the organization needs to purchase additional capacity via expansion packs.
- Service GPT – Similarly priced at $50 per user per month, this add-on attaches to Service Cloud Einstein. It provides customer service agents AI assistance with drafting chat responses, email replies, and case summaries. Like Sales GPT, the subscription includes a certain number of AI generation credits. Once the preset credit limit is exhausted by an agent’s AI usage, further generations incur extra costs (through paid expansion packs or usage-based charges).
Notably, Salesforce initially restricted Einstein GPT availability to customers on its highest-tier “Unlimited” editions of Sales Cloud and Service Cloud.
In practice, only organizations already on those top-tier plans could even access GPT features. This gating strategy led some customers to consider expensive edition upgrades or special negotiations just to pilot Einstein GPT.
The concept of Einstein GPT credits is central to these add-on licenses. Salesforce did not initially specify exactly how many credits each $50 license provides, nor what constitutes a single “credit” of AI generation (it’s analogous to token counts in other AI services). Essentially, each user license comes with a certain amount of AI processing capacity included.
If an organization’s users collectively go beyond their allotted GPT credits, Salesforce offers Enterprise Expansion Packs to buy additional AI capacity.
In other words, the $50 add-on opens the door to generative AI for a user, but heavy usage can drive the cost higher. This usage-based element made budgeting tricky – finance teams had to anticipate how much their staff might use the AI and what the overage fees could be.
Read how to manage and control costs, AI Usage Limits and Overage: Navigating Salesforce’s AI Consumption Model.
Einstein Copilot and Einstein 1 Edition Pricing
Later in 2023 and into early 2024, Salesforce expanded its generative AI offerings beyond the initial Sales GPT and Service GPT features. They introduced Einstein Copilot, a more advanced conversational AI assistant that works across the Salesforce platform.
Einstein Copilot allows users to ask questions and receive AI-generated answers or actions within Salesforce, effectively acting like a chatbot that can perform tasks in any connected app (sales, service, marketing, etc.). With this broader capability came a new pricing approach.
Salesforce launched new “Einstein 1” editions of Sales Cloud and Service Cloud to package Einstein Copilot and related AI capabilities. The Einstein 1 Sales Cloud and Einstein 1 Service Cloud editions are premium bundles priced at around $500 per user per month (list price).
This is a dramatic jump compared to standard CRM licensing (for context, Sales Cloud Enterprise Edition is roughly $165 and Unlimited Edition is about $330 per user). It’s essentially an upsell bundle: the Einstein 1 edition includes everything in Unlimited plus the Einstein Copilot functionality, Data Cloud integration, advanced analytics (Tableau), Slack collaboration tools, and other AI-powered features – with generative AI as the centerpiece.
For organizations that did not want to switch their entire subscription to the $500 Einstein 1 editions, Salesforce also offered Einstein Copilot as a separate add-on. This Copilot add-on could be attached to existing Sales Cloud or Service Cloud licenses (Enterprise or Unlimited) at roughly $75 per user per month. By paying this fee on top of the base license, a customer could give certain users access to Einstein Copilot’s AI assistant capabilities without a full edition upgrade.
Even the $75 Copilot add-on is substantial. For example, an Enterprise Edition user paying $165 jumps to $240 per month with Copilot, and an Unlimited Edition user at $330 rises to about $405 with Copilot included. In either case, the company is spending a significant premium solely to enable generative AI features on top of core CRM functionality.
Salesforce clearly places a high monetary value on these AI capabilities. The company justifies the steep pricing by pointing to potential productivity gains and the secure, enterprise-grade nature of its AI (e.g. the Einstein Trust Layer that prevents sensitive data from leaking to external models). However, savvy enterprise buyers have viewed these add-ons with skepticism.
They know Salesforce is using Copilot to boost its revenue per user, and many insisted on pilot programs to confirm that Copilot’s benefits outweigh its costs before deploying it broadly.
It’s also important to see Salesforce’s push for Einstein Copilot in strategic terms. By bundling Data Cloud and Slack with the Einstein 1 editions, Salesforce encourages customers to invest in those products as part of the AI journey.
A company paying for these new AI capabilities might also be increasing its dependence on Salesforce for data management and collaboration. CIOs and CFOs must weigh the value of this integrated approach versus the flexibility (and cost savings) of using alternative solutions for things like data warehousing or team chat.
AI Cloud Starter Pack and Upfront Costs
When Salesforce first announced its overarching AI Cloud suite (encompassing all the generative AI features plus the infrastructure and trust layer), it floated a large-scale entry package for enterprises. The AI Cloud Starter Pack was advertised at around $360,000 per year as a flat annual price.
This hefty “starter” price drew considerable attention in the industry. For that sum, Salesforce bundles a collection of components aimed at jump-starting AI deployments: a Data Cloud environment (to unify and store customer data for AI), a generous allowance of MuleSoft automation integrations, Einstein GPT capabilities across multiple clouds, Tableau analytics, Slack licenses, and even some Salesforce professional services (such as an AI readiness assessment).
The idea behind the $360K/year starter pack is to give big customers a one-stop shop to infuse AI across their Salesforce landscape. For a large company, $360,000 annually might be a manageable investment if it covers a wide range of users and use cases with all those advanced tools.
However, not every organization can justify paying such a sum upfront just to kick off an AI initiative. This pack is clearly aimed at the largest, most AI-forward enterprises – those willing to make a significant budget commitment to be early adopters of Salesforce’s AI capabilities.
Industry reaction to the initial AI Cloud pricing was mixed. Salesforce argued that the bundle could deliver value commensurate with its cost by enabling major efficiency gains across departments (sales, service, marketing, etc.).
Skeptics countered that a number that large would deter many customers, and indeed some firms started exploring alternative AI solutions due to Salesforce’s lofty entry fee.
In response, Salesforce soon introduced more modular ways to adopt their AI features (for example, the per-user add-ons and later the unlimited usage plans), rather than forcing a massive all-in-one purchase.
The $360K sticker price ultimately underscored Salesforce’s confidence in the transformative value of its AI offerings – but it also reminded customers that this transformation would not come cheap.
2025 Update – Agentforce and Unlimited AI Usage
By mid-2025, Salesforce revised its generative AI pricing approach, partly due to customer feedback and competitive pressures. Einstein Copilot and the earlier GPT add-ons evolved into a new offering branded as Agentforce.
With Agentforce, Salesforce introduced an “all-you-can-use” model for AI, aiming to simplify how companies pay for these capabilities and remove the uncertainty of usage overages.
The Agentforce add-on is priced at roughly $125 per user per month and grants that user unlimited generative AI usage within their Salesforce applications.
In practice, a user can invoke Einstein GPT and other AI-driven “agent” features as often as needed without the organization worrying about running out of credits or incurring overage fees.
For teams expecting heavy use of AI (say, sales reps who generate dozens of emails or service agents handling countless AI-suggested replies), this flat per-user pricing provides much more cost predictability than the old credit-based approach.
This shift to a fixed per-user fee for unlimited AI is a notable change. It eliminates the need to meter each prompt or charge for extra tokens, although it comes at a higher base cost per user.
For small teams or light users, $125 could increase costs compared to the $50 add-on model. But for larger deployments where users would have easily exceeded the previous credit allotments, the unlimited model prevents runaway expenses.
It also indicates Salesforce wants AI to become ubiquitous in daily CRM use – the pricing encourages enabling it broadly, not just for a few power users.
In addition to the $125 plan, Salesforce rolled out a new top-tier offering called Agentforce 1 at about $550 per user per month.
Agentforce 1 includes not only the unlimited AI usage per user but also the core CRM license itself and a huge pool of AI Flex Credits (roughly 1,000,000 credits per organization).
Those Flex Credits support extremely large-scale AI workloads beyond normal user interactions – for example, autonomous AI bots performing background tasks or massive batch processes that generate content.
Agentforce 1 is essentially an all-inclusive package for AI-heavy operations. It bundles the standard cloud license, the generative AI features, and a significant consumption allowance for AI and automation.
In most cases, that 1,000,000 credit allotment ensures the company won’t run out of AI capacity under normal conditions. Should an organization somehow exhaust that pool, they can purchase additional credits as needed.
This hybrid model – fixed per-user licensing combined with optional consumption for extreme scenarios – does add some complexity for procurement.
Teams will need to understand how credits translate to real usage and monitor any extraordinary workloads. On the upside, for the majority of customers, the unlimited user licenses mean they can adopt AI features without constantly watching a usage meter.
Strategic advise, Salesforce Einstein and AI Cloud: A CIO’s Playbook for Regulated Industry Clouds.
Comparison with Competitor AI Pricing
Enterprise software buyers will naturally compare Salesforce’s AI pricing with other options in the market. Salesforce is not the only CRM or business software vendor adding generative AI, but each is taking a different approach to monetization.
Here’s how a few major competitors stack up:
- Microsoft – Microsoft has been embedding AI “Copilot” features across its product suite. In the Microsoft 365 productivity apps, Copilot is offered at about $30 per user/month as an add-on. In the Dynamics 365 CRM line, Microsoft has included many AI functions in existing licenses, with any additional AI upgrades so far priced much lower than Salesforce’s offerings. In short, Microsoft’s strategy keeps AI add-ons relatively affordable (likely subsidized by its broader cloud business), positioning AI as a common productivity booster rather than a premium upsell.
- HubSpot – HubSpot, which caters to mid-market CRM customers, has introduced generative AI tools (like content assistants under the “Breeze AI” name) and is including them at no extra charge in its paid plans. If you subscribe to HubSpot’s professional or enterprise tiers, the AI features are generally built-in instead of sold separately. HubSpot is using AI as a value-add to enhance its platform appeal, rather than as an immediate revenue generator.
- Zoho – Zoho’s CRM AI assistant (Zia) takes a bring-your-own approach for generative AI. Customers can connect their own OpenAI API key to enable features like AI-generated emails or summaries. Zoho itself doesn’t charge a specific fee for these capabilities – the customer only pays the usage fees to OpenAI or whichever AI service they use. This means a company can experiment with AI in Zoho at relatively low cost, paying only for what they use externally, though it requires trusting a third-party AI provider with their data prompts.
When evaluating Salesforce’s AI pricing against these alternatives, companies need to weigh cost versus benefit carefully. Salesforce’s solution is undoubtedly expensive, but it offers deep integration with the company’s CRM data and strong security/compliance measures via the Einstein Trust Layer.
For a highly regulated industry or a business that runs heavily on Salesforce, that enterprise-grade, tightly integrated approach might justify the premium.
On the other hand, some competitors are bundling core AI features at little to no extra cost, which can be attractive if budget is a primary concern and if those tools meet the business’s needs.
CIOs and CFOs should assess how critical Salesforce’s advanced AI capabilities are to their strategy – and whether similar results could be achieved through cheaper alternatives or in-house initiatives.
Read about our Salesforce Negotiation Services
