Incident sheet · /enterprise-ai

Filed
Read
3 min · 633 words
Severity
Routine

Agentic AI pricing splits three ways as buyers gain leverage

Subscription per user leads AI payment preferences at 36.4%, Infotech finds, as Salesforce reworks Agentforce pricing and Horizon Media negotiates consumption disputes weekly.

What buying party balloons taught me about buying agentic AI
FeedWhat buying party balloons taught me about buying agentic AI — Pexels / pixabay

Event log

  1. Subscription per user tops preferred AI/software payment models at 36.4% in Infotech's Global IT Leadership Survey.
  2. Open source with support contract follows at 23.1%, consumption pricing at 22.1%, one-time purchase at 13.9%.
  3. Salesforce has changed Agentforce pricing multiple times since the per-conversation model launched in 2024.
  4. Salesforce price-per-resolution caps voice interactions at 10 minutes before charges increase.
  5. Horizon Media negotiates consumption-cost disputes up to dozens of times a week with each client.

Subscription per user remains the top payment preference for AI and software solutions among IT leaders, at 36.4%, according to Infotech's Global IT Leadership Survey. The finding anchors a widening procurement debate over how enterprises should pay for agentic AI (autonomous software agents that perform multi-step tasks with minimal human input).

Infotech principal research director Brian Jackson frames the choice through three vendor business models, each with distinct cost and risk profiles:

  • Flat-rate pricing. A set price for compute capacity. Cheapest and most predictable, but buyers do the integration work and may exhaust capacity before reaching the outcome.
  • Consumption-based pricing. A metered rate on inference compute (the processing used to run AI models). Guarantees capacity for all workloads, but budgets are less predictable and buyers may pay for wasted runs.
  • Outcome-based pricing. Payment only for delivered results. Least effort and guaranteed outcomes, but the most expensive option.

The model a vendor offers reflects where it started, Jackson wrote. Software-as-a-service vendors are layering consumption metering onto subscriptions. Business process outsourcers are pushing outcome-based deals.

What do the survey numbers show?

After per-user subscription at 36.4%, Infotech's survey found open source software with a support services contract ranked second at 23.1%. Consumption-based or utilization-based pricing followed at 22.1%. A one-time purchase rounded out the top four at 13.9%.

How is Salesforce pricing Agentforce?

Salesforce has changed Agentforce pricing multiple times in recent months as the cloud CRM vendor shifts toward AI-driven CRM. Customer-facing agent consumption carries a per-conversation price, an approach launched alongside Agentforce in 2024, or a newer price-per-resolution option.

Internal and customer-facing agents can also draw on flex credits, with a standard number of credits required per action. Per-conversation pricing applies only to customer interactions. Price-per-resolution is available only for Help Agent, a pre-packaged agent for resolving customer inquiries.

Salesforce calls price-per-resolution outcome-based pricing and shifts risk to itself: no charge applies when the customer escalates to a human agent. The definitions carry limits, including a cap of 10 minutes of voice interaction before the meter ticks up.

Negotiation levers for buyers include:

  • Volume discounts on flex credits committed in advance, with steeper discounts for larger commitments and longer contracts.
  • Where to pay the per-seat add-on for unmetered use.
  • Whether unused credits roll over to the next period.

Where does outcome pricing break down?

Horizon Media shows the friction. The advertising agency launched agentic buying capabilities on its Blu platform in July, packaging agentic AI, data sources and analytics tools so clients build audience groups and deliver ads themselves.

Pricing is flexible and customer-driven. Clients running their own analysis incur variable costs and dispute outcomes, producing constant negotiation — as many as "dozens of times a week with each client," one executive said.

Clients argue they should not pay when they perceive no value delivered. The agency, which already paid inference costs and data calls to answer the prompt, responds that users need more training for productive results. Most disputes settle with the client paying most, but not all, of the consumption costs.

Where should buyers apply pressure?

Outcome verification separates easy cases from hard ones, Jackson wrote. A balloon is inflated or it is not. Agentic AI outcomes resist such clean verification, and the contract's dispute space is exactly where buyer leverage sits. He advised concentrating negotiation on the definition of an outcome and on what precisely is metered.

Enterprise buyers placing agentic AI at the core of operations face harder service relationships than consumers. The CFO will review the bill, and the vendor's definition of "done" will matter as much as its price, Jackson wrote.

via techtarget.com (Source)

Filed under

  • agentic-ai
  • ai-pricing
  • consumption-based-pricing
  • outcome-based-pricing
  • software-procurement
Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Senior reporter covering money, funding and market shifts around enterprise information technology and the leaders who run at CIO Monitor.

199 articles

The newsroom digest

One email a day with the stories that matter. No noise.

By subscribing you agree to the privacy policy.

Cross-references

‹ Previous articleNext article ›