Salesforce reframed its enterprise platform around three product lines in 2024 and 2025: Data Cloud (the customer data platform), Data 360 (the data unification and governance layer announced in 2024), and Agentforce (the AI agent platform launched in late 2024 and expanded through 2025). The three sit on top of the historical CRM clouds. The pricing is consumption based, the metering is opaque, and the renewal economics are unlike any prior Salesforce purchase. This guide is the buyer side framework for understanding the three products, pricing them realistically, and negotiating a 2026 renewal that does not lock the customer into a multi year commitment they will struggle to consume. Pair this with our Salesforce services overview, the Salesforce Knowledge Hub, and the Salesforce renewal negotiation playbook.
The strategic question for the customer is not whether Data Cloud or Agentforce delivers value. The platform's capabilities are real. The question is whether the consumption based commercial model fits the customer's actual usage trajectory, and what protections must be negotiated to prevent the commitment from becoming a stranded asset. The buyer side answer addresses both at once.
What Data Cloud actually is
Data Cloud is Salesforce's customer data platform, the unified data layer that ingests data from Salesforce and external sources, resolves identity, builds calculated insights, and activates segments into downstream applications. The product was previously branded Customer Data Platform (CDP), Genie, and Data Cloud across successive releases. The current commercial model meters consumption in credits, where each credit represents a unit of work the platform performs. Different operations consume different credits: data ingestion is one rate, identity resolution is another, calculated insights are a third, segment activation is a fourth.
The credit metering is the single most consequential commercial element. Customers commit to credit volumes for the term, drawing down credits against actual usage. Unused credits expire at the end of the commitment. Customers who commit aggressively without precise consumption modeling forfeit unused credits. Customers who commit conservatively without expansion rights pay overage at premium rates when consumption exceeds the commitment.
Data 360: the unification layer
Data 360 is the data unification layer announced at Dreamforce 2024 and refined through 2025. It builds on Data Cloud and adds three capabilities: Zero Copy Federation with major data warehouses (Snowflake, Databricks, BigQuery, Microsoft Fabric); data quality and master data management across federated sources; and data product management, treating curated data sets as managed products with versioning and access control.
The commercial model for Data 360 sits within the Data Cloud credit framework. Federation operations consume credits. Data quality operations consume credits. Master data resolution consumes credits. The customer's credit consumption forecast must therefore include both the Data Cloud usage and the Data 360 usage. Customers who model only Data Cloud consumption underforecast the credit need by 30 to 60 percent in many estates.
Agentforce: the AI agent platform
Agentforce is Salesforce's AI agent platform. The platform enables the design and deployment of AI agents that perform multi step tasks across Salesforce data and business processes. The agents are deployed across customer service (Service Cloud), sales (Sales Cloud), commerce (Commerce Cloud), marketing (Marketing Cloud), and broader business workflows. The 2025 releases expanded the platform with the Atlas reasoning engine, the agent builder, and the Foundations data layer.
Agentforce is metered by conversation in the Service Cloud and Sales Cloud contexts, and by task action in broader business contexts. Salesforce's published list price in 2026 is approximately $2 per conversation for the Service Agent SKU and $0.50 per task action for the broader agent contexts, both with significant discount available in negotiated commitments. The metering is opaque in the sense that a single user interaction often produces multiple conversations or task actions, depending on the agent configuration. Customers who do not measure the metering at agent design time discover the run rate at the end of the first quarter, when the consumption report exposes the consumption pattern.
Consumption units in detail
The three product lines together produce a consumption based commercial model that requires precise consumption modeling. The units are different for each product. The units interact with each other. The customer's commitment must reflect the actual consumption pattern across all three.
| Product | Primary Unit | Common Operations | Cost Trap |
|---|---|---|---|
| Data Cloud | Credits | Ingestion, identity, calculated insight, activation | Calculated insights at high frequency |
| Data 360 | Credits (shared) | Federation, data quality, MDM, data product | Federation queries against external warehouses |
| Agentforce Service | Conversations | Service agent interactions | Multi turn conversations counted per turn |
| Agentforce Business | Task actions | Multi step automation in business workflows | Background actions not visible to user |
Consumption modeling
The consumption model for the three products together must be built before the commitment is sized. The model uses three inputs. Historical data volumes for ingestion and federation. Forecast user populations and use cases for activation, conversations, and task actions. Salesforce's standard credit and conversation cost factors for each operation. The output is a credit and conversation forecast over the commitment term, with sensitivity bands for the major variables.
Customers who run the modeling exercise produce a commitment that lands within 80 to 110 percent of actual consumption across the term. Customers who do not run the modeling exercise commit on Salesforce's account team's recommendation, which is consistently aggressive relative to actual consumption.
List pricing in 2026
Salesforce's published list pricing for the consumption based products is structured by tier. The tiers below are representative of 2026 pricing, with discount typically available in the 25 to 55 percent range for negotiated commitments and multi year terms.
- Data Cloud Starter.$108,000 per year for 1 million credits.
- Data Cloud Growth.Negotiated, typically 5 to 50 million credits with discount.
- Data Cloud Enterprise.Negotiated, typically 50 million credits and above with deeper discount and price holds.
- Agentforce Service Agent.$2 per conversation at list, with significant volume discount available.
- Agentforce Business Agent.$0.50 per task action at list, with significant volume discount.
- Foundations.Included in Data Cloud Growth and Enterprise tiers; the foundational data layer for Agentforce.
The list prices are starting points. The negotiated price reflects volume, term length, the customer's discount history, and the customer's strategic value to Salesforce. Discounts of 50 percent on Data Cloud and 60 percent on Agentforce are achievable for committed multi year terms with volume.
Commitment economics
The customer's commitment to Data Cloud, Data 360, and Agentforce sits in the Salesforce master subscription agreement (MSA) or as a Data Cloud and Agentforce addendum to the existing CRM commitment. The commitment economics determine the true cost of the platform. Three structural choices matter most.
The ramp commitment
Salesforce's standard commitment is a multi year ramp with annual escalation. Year one commitment is lower than year three commitment, reflecting expected adoption growth. The ramp shape is negotiable. Customers should structure the ramp to match the realistic adoption curve, which is typically slower in year one than Salesforce's standard ramp implies. A flatter ramp protects the customer from year one over commitment.
Overage protection
Salesforce's standard contract prices overage at the list rate per credit or conversation, which is significantly higher than the committed unit rate. The customer should negotiate overage protection: the contractual right to convert overage into pre commit at the next anniversary at the committed unit rate, rather than paying overage at list. Customers without overage protection face cost spikes when consumption exceeds forecast.
True down rights
Salesforce's standard contract does not allow reduction of committed quantities mid term. The customer should negotiate true down rights at annual anniversaries when adoption falls short of plan. The standard true down protection is the right to reduce committed quantities by 10 to 20 percent at each anniversary, with notice. Customers who do not negotiate true down rights find themselves paying for unused commitments year over year.
Data and AI governance
Data Cloud, Data 360, and Agentforce together create governance requirements that exceed traditional Salesforce CRM governance. The customer's data flowing through Data Cloud is subject to privacy regulations across jurisdictions. The AI agents in Agentforce make decisions that affect customers and create accountability requirements. The data products in Data 360 cross organizational boundaries and require access controls. The governance work is not optional. It is a precondition to deployment at scale.
The governance framework
The framework runs across three areas. Data classification, with explicit classification of every data set ingested into Data Cloud and federated through Data 360. Access control, with explicit data product permissions and AI agent data access scopes. AI accountability, with explicit logging of agent decisions, prompts, and outputs, and a defined human review process for agent actions that affect customers. Customers who run the governance work in parallel with the deployment achieve scaled adoption. Customers who treat governance as a year two project find the deployment stalled at year one.
The 2026 renewal stance
Customers approaching renewal in 2026 face a Salesforce account team that is heavily incentivised to convert CRM customers to Data Cloud and Agentforce commitments. The account team's standard motion bundles Data Cloud and Agentforce into the CRM renewal at a price discount that appears generous against list. The buyer side discipline evaluates the bundle on its own merits, not on the discount headline.
Seven negotiation levers
The buyer side renewal stance uses seven levers. Each lever is independently negotiable. Customers who use all seven preserve significant value. Customers who address one or two pay headline discount and accept structural concessions.
- Modeled commitment.Commit to the modeled consumption, not the account team's forecast. Document the model.
- Flatter ramp.Year one commitment matches year one realistic consumption, not year three forecast.
- Overage protection.Right to convert overage into pre commit at unit rate.
- True down rights.Annual right to reduce committed quantities with notice.
- Discount preservation.Multi year price hold that preserves the negotiated unit rate through term.
- Foundation entitlement.Foundations included in the Data Cloud commitment, not as separate priced add on.
- Termination protection.Right to terminate consumption based products at anniversary with notice, separate from CRM term.
Pattern study: a 2,400 user financial services firm
A North American wealth management firm we advised was 18 months into a Data Cloud pilot when its 2026 renewal cycle opened. Salesforce's renewal proposal bundled CRM, Data Cloud, and Agentforce into a five year ramped commitment of 47 million dollars total, with the headline discount described as 38 percent off list. The customer's actual Data Cloud consumption in the pilot phase was 8 million credits per year, against a year five commitment in the proposal of 95 million credits per year.
The defense had four steps. We modeled the consumption forecast across Data Cloud, Data 360, and Agentforce: a year three credit forecast of 35 million credits, a year five forecast of 60 million credits. We negotiated the ramp to match the modeled forecast, with overage protection and true down rights. We negotiated Foundations as included rather than priced. We negotiated termination protection on the Agentforce piece, recognising the customer's adoption was less certain than Data Cloud. The renegotiated commitment was 28 million dollars total over five years, with annual true down and contracted unit rate preservation. The customer's avoided over commitment was 19 million dollars.
For more Salesforce patterns see our case studies library, Canadian financial institution case study, and Finnish energy company case study.
The consumption model is the contract. Customers who do not model their actual consumption commit on Salesforce's forecast, which is consistently aggressive relative to reality.
Closing thought
Data Cloud, Data 360, and Agentforce are strategic platforms with real value for the right customer. The commercial model is unforgiving for customers who commit without precise consumption modeling. The buyer side discipline is to model the consumption, structure the commitment to match, negotiate the protections, and treat the renewal stance as a negotiation rather than a default acceptance. Customers who follow that discipline capture the value. Customers who do not subsidise Salesforce's product transition.
Redress Compliance is independent and 100 percent buyer side. We do not partner with Salesforce. We have advised on Data Cloud, Data 360, and Agentforce negotiations across financial services, manufacturing, retail, healthcare, and the public sector. If you are evaluating the consumption based products or planning a 2026 renewal, the next step is a confidential briefing.
