Salesforce shifts to flat-rate AI pricing as customers demand predictable costs over usage models

Salesforce AI Embraces Flat-Rate Pricing Through AELA

Salesforce has made a decisive pivot in how it charges for artificial intelligence, moving away from usage-based models toward its Agentic Enterprise License Agreement (AELA), a flat-rate, seat-based structure that promises unlimited use of AI agents for a fixed fee 1

1

. CEO Marc Benioff confirmed the shift during a recent earnings call, acknowledging that customers demanded more flexibility after the company initially floated per-conversation and consumption-based models 3

3

. The AELA covers products including AgentForce, Data Cloud/360, and MuleSoft over two or three-year commitments, fundamentally changing how enterprises budget for AI capabilities .

This approach to AI pricing signals a broader industry trend where market leaders normalize AI as a native capability rather than an incremental add-on. Microsoft has consolidated specialized Copilots into Microsoft 365 Copilot with simple per-user pricing, while Google folded Gemini into Workspace tiers, embedding AI into Docs, Gmail, and Sheets . But Salesforce has pushed further by eliminating charges based on actions, tokens, or conversations entirely, reframing AI agents as economic actors that generate business value rather than utilities that incur variable costs .

Willing to Lose Money for Long-Term Monetization

Chief Revenue Officer Miguel Milano revealed at the Barclays 23rd Annual Global Technology Conference that Salesforce is "okay with losing money on some AI deals" 2

2

. His rationale centers on customer success and long-term relationship value. Milano explained that if a customer deploys AI so extensively under a $5 million AELA that the deal becomes unprofitable for Salesforce, "that customer is the happiest customer in the world. And then I have another 20 years to monetize that customer" 2

2

.

This strategic bet on long-term agent value reflects confidence that AI agents will materially reshape enterprise cost structures and productivity . Milano promised investors that Salesforce's ability to monetize digital labor relationships would be three to four times the business it currently does with customers in CRM, marketing, and data analytics 2

2

. Marc Benioff reinforced this vision, stating customers could expect "three or four times or 10 times more value" from Salesforce products, justifying significantly higher monetization multiples 4

4

.

Predictable Costs Drive Adoption Over Consumption-Based Models

The return to seat-based AI licensing addresses enterprise concerns about unpredictable spending. Jan Cook, senior software licensing expert with Gartner, told The Register that "customers still adopt a cautious approach to investing in GenAI because of the unpredictability of pricing models" 3

3

. The shift allows technology departments to avoid signing blank checks to spiraling usage costs while they evaluate how employees integrate new workflows 3

3

.

However, vendors maintain underlying controls through secondary metrics. Most seat-based licenses include provisioned limits of AI units or credits, with fair use policies that allow vendors to manage potential overuse without losing money on backend calls to large language models 3

3

. Milano confirmed that pay-as-you-go deals remain available for customers who prefer them, though most choose to pre-commit spending on the AI platform 2

2

.

AI Agents as Productive Assets Reshape Investment Logic

For enterprise buyers, particularly CFOs who increasingly govern AI budgets, this pricing evolution shifts AI from a variable-cost experiment to a strategic, multi-year investment . When AI agents are framed as productive assets rather than utilities, buyers must address capital-allocation questions about economic output, ROI, IRR, and the useful life of agents .

This reframing advantages vendors with strong value-attribution clarity who can credibly demonstrate outcomes like case resolution rates or cycle-time reductions . Gartner forecasts that agentic AI could drive around 30 percent of enterprise application software revenue by 2035, surpassing $450 billion, up from 2 percent in 2025 3

3

.

Vendor Lock-In and Strategic Risk Considerations

While flat-rate pricing offers stability, Forrester warned in an August report that organizations face "monumental" efforts retraining employees in new workflows to benefit from AI agent platforms 2

2

. This strategy "dramatically increases vendor lock-in and the strategic risk of your choice," the analyst cautioned 2

2

. The warning echoes concerns about Salesforce's explicit promise to monetize customers over decades, which some view as a double-edged sword for enterprises committing to multi-year agreements.

Cook noted that adoption remains steady rather than explosive: "It's not a sharp spike in adoption, but what we see is a very predictable, steady adoption trajectory for most vendors" 3

3

. Notably, companies aren't significantly reducing headcount due to AI, which supports the viability of seat-based AI licensing. Forrester research found that 55 percent of employers regret laying off workers because of AI, while 57 percent of those in charge of AI investment expect it to increase headcount over the next year, compared to just 15 percent expecting decreases 3

3

.

Salesforce reported a 9 percent year-over-year increase in quarterly revenue and raised its full-year revenue expectations from $41.45 billion to $41.55 billion 4

4

. As the company enters its fourth and final quarter of fiscal 2026, the success of AELA will test whether enterprises view AI agents as transformative business value generators worthy of long-term investment commitments.