Agentforce Just Hit $1.2B ARR in 14 Months: What CROs Should Audit Before Q3 Planning

The number that changes your Q3 planning conversation is $1.2 billion. And it arrived faster than almost anyone expected.

Salesforce disclosed its first-quarter fiscal 2027 results on May 27, 2026, with Agentforce annual recurring revenue (ARR, the annualized revenue run rate from active subscriptions) reaching $1.2 billion, a 205% increase year over year. The platform launched in September 2024. Reaching $1.2 billion ARR in 14 months puts it in rare company among enterprise software product ramps.

But for Chief Revenue Officers (CROs) and Sales Operations leaders, the ARR headline is the second most important number in the report. The first is 3.8 billion.

Quick Take: Agentforce's Agentic Work Unit (AWU) volume grew 111% quarter over quarter. That's not a usage metric. It's a billing signal. CROs who don't have a unit-economics model for AI agent consumption before Q3 planning are going to be surprised by what shows up in their tech stack budget.

What Salesforce Actually Reported

Salesforce's Q1 FY27 results beat analyst consensus across the board. Total revenue reached $11.13 billion (up 13% year over year). Non-GAAP earnings per share of $3.88 outpaced the $3.12 consensus by 24%. The company recorded a non-GAAP operating margin of 34.8%, a record level for the business. Salesforce raised its full-year FY27 revenue guidance to $45.9 billion-$46.2 billion, with second-quarter guidance set at $11.27 billion-$11.35 billion (representing 10-11% growth).

Within those headline numbers, the AI platform metrics stood out. Agentforce combined with Data 360 reached approximately $3.4 billion in ARR, up over 200% year over year. Agentforce alone contributed $1.2 billion; Informatica Cloud contributed $1.1 billion. This is a meaningful step up from the approximately $800 million in Agentforce ARR reported earlier in 2026, representing roughly a 50% increase in a single quarter.

The AWU number tells the adoption story in operational terms. 3.8 billion Agentic Work Units (AWUs, the unit Salesforce uses to meter agent task completions across Agentforce and Slack) have been delivered to date, growing 111% quarter over quarter. On the data side, Data 360 ingested 52 trillion records in Q1 (up 136% year over year), with 35 trillion arriving via Zero Copy (up 277% year over year). Zero Copy means Salesforce is reading data from external warehouses without moving it, which reduces friction for customers whose data lives in Snowflake, Databricks, or BigQuery.

Key Facts

• Agentforce ARR: $1.2 billion, up 205% year over year (Salesforce Q1 FY27 earnings, May 27, 2026)
• Agentic Work Units delivered to date: 3.8 billion, growing 111% quarter over quarter (Salesforce Q1 FY27)
• Data 360 records ingested in Q1: 52 trillion (up 136% year over year), with 35 trillion via Zero Copy (up 277% year over year) (Salesforce Q1 FY27)

Why the AWU Number Matters More Than the ARR Number

ARR tells you Salesforce is winning deals. The AWU count tells you those deals are actually being used in production, and at scale.

Most enterprise software platforms struggle to convert license revenue into genuine usage. A deal closed is not a workflow changed. The fact that Agentforce delivered 3.8 billion work units at 111% quarter-over-quarter growth means customers aren't just buying licenses. They're running agents on real workflows at volumes that compound each quarter.

That has a direct implication for how you budget. Traditional seat-based software has a predictable cost curve: you pay per user, you add seats when you hire, you reduce seats when you cut. Consumption-based AI agent billing doesn't work that way. Your AWU spend grows with the number of workflows you automate, the frequency those workflows run, and the complexity of each task the agent handles. A single sales process automation could trigger hundreds of AWUs per deal.

If you don't know your AWU cost per deal, your AI spend can quietly double inside a quarter. That's what the 111% quarter-over-quarter growth number means in practice for any organization that has Agentforce in production. For guidance on structuring AI-native sales workflows, see What Is an AI Sales Operator? 4 Patterns Working Together.

The 4-Point CRO Audit

This is "The AWU Audit Stack": four questions every CRO and Sales Operations leader should have answered before Q3 planning begins.

Question 1: Do you have a unit-economics model for AWU consumption?

If you're running Agentforce in any capacity, ask your Sales Ops team for a breakdown of AWUs by use case. How many AWUs does a typical deal cycle consume? How many per qualified lead? Per support ticket routed through Slack? Most teams don't have this view yet, which means they're flying blind on one of the fastest-growing line items in their tech stack. The goal isn't to minimize AWU usage. It's to know your cost per outcome so you can make rational trade-offs.

Start here: pull your last 30 days of AWU consumption from Salesforce and map each cluster of usage to a specific workflow. Then run the math on AWUs per deal and AWUs per qualified lead. If you're still in pilot mode, use Salesforce's reference customer benchmarks to project what production volumes would cost at your current deal velocity.

Question 2: Which Agentforce skills are actually in production inside your org?

Agentforce adoption is uneven. Many organizations closed deals in 2025, configured a few agents, and then let the platform sit while teams worked through change management. The competitive pressure is shifting. If your peers are deploying agents at scale (and the 3.8 billion AWU count suggests many are), the window to catch up without a capability gap is narrowing.

Audit your current production use cases against the highest-ROI patterns being adopted by reference customers. Salesforce's field team and customer success organization maintain a ranked list of use case patterns by industry and deal type. Request that list before your Q3 planning session. Time-box three pilots against the top patterns. If a pilot doesn't show measurable ROI in six weeks, kill it and move to the next one. For context on how leading sales ops teams are structuring AI adoption, see AI Sales Ops vs. Traditional Sales Operations.

Question 3: How much of your data estate is already connected to Salesforce's Zero Copy layer?

The 277% year-over-year growth in Zero Copy volume is a strong signal that Salesforce is building a data plane, not just an agent plane. Zero Copy means Salesforce agents can operate on data that lives in your warehouse without requiring you to duplicate or migrate it. That's a meaningful architectural shift. If your sales data lives in Snowflake or Databricks, you may already have more Zero Copy connectivity than you realize. And if you don't, setting it up is now a shorter path than it was a year ago.

Have your Sales Ops or Data Engineering team map which data sources are connected today. Then identify two or three high-value data sets (competitive intelligence feeds, product usage data, support ticket history) that aren't yet connected but would meaningfully improve agent quality if they were. See AI Lead Scoring Beyond Rules-Based Models for examples of how real-time data feeds change scoring outcomes.

Question 4: Have you designed your agent vs human workflow decision framework?

The 111% quarter-over-quarter AWU growth means use cases are scaling faster than most organizations have designed for. Reps may be triggering agent work units without understanding the cost implications. Without a clear framework for which workflows are agent-led, which are agent-assisted, and which stay human-only, you end up with an inconsistent mix of automation that neither reduces rep burden nor improves deal quality.

A practical three-tier model: agent-led for high-volume, low-complexity tasks (lead triage, data enrichment, follow-up sequencing); agent-assisted for medium-complexity tasks where human judgment adds value (discovery call prep, deal coaching, pipeline review summaries); human-only for high-stakes, relationship-sensitive interactions (executive business reviews, procurement conversations, contract negotiations). For a framework on structuring this kind of tiered approach, see Pipeline Reviews That Surface Real Risk.

Where Agentforce Fits Versus HubSpot's Trajectory

It's worth putting the Salesforce numbers in competitive context. HubSpot reported that its Customer Agent product surpassed 9,000 customers in Q1 2026 and now accounts for 53% of HubSpot AI credit consumption. That's meaningful scale for a company building into the mid-market, but the unit economics comparison matters: Salesforce is billing AWUs (task completions), while HubSpot is billing AI credits (a broader usage abstraction).

Both models point in the same direction. Consumption-priced AI is becoming the default structure for sales technology at scale. The question isn't which platform you prefer. It's whether you have a financial model that accounts for consumption-based billing before your Q3 numbers are set.

CROs who treat AI agent costs as a fixed line item in their FY budget will be wrong. The variable is how fast your team actually uses the agents. And based on the 111% quarter-over-quarter AWU growth, the correct assumption is: faster than you planned.

For a broader perspective on how AI tools are changing the commercial tech stack, see Forecasting Accuracy: Getting to Plus-or-Minus 10% by Week 8.

What to Do This Week

Here's a concrete checklist you can take into your next pipeline review or QBR prep session before Q3 planning:

AWU Audit Checklist for CROs and Sales Ops

1. Pull current AWU consumption report from Salesforce (Admin, Org Settings, Agentic Usage). Sort by workflow or agent type.
2. Map top AWU consumers to specific sales workflows. Calculate AWUs per deal closed, per SQL generated, per support ticket resolved.
3. Request Salesforce's current reference customer benchmark deck for your industry vertical. Ask your CSM or AE for the top five use case patterns by ROI.
4. Identify your top three candidate pilots for Q3. Each should have a clear success metric (cost per outcome), a six-week time box, and a named owner.
5. Map your current data warehouse connections to Salesforce Data 360. Identify two high-value data sources not yet connected via Zero Copy.
6. Document your agent vs human decision framework. For each active Agentforce workflow, classify it as agent-led, agent-assisted, or human-only. Close any workflows where the classification is unclear.
7. Build a Q3 AWU consumption forecast. Use your current AWU rate plus planned new pilots. Stress-test at 2x growth (which is conservative relative to the 111% QoQ the platform reported).
8. Present this model at your Q3 planning QBR as a consumption cost line, not a fixed subscription cost.

The organizations that treat agentic AI as a seat-based SaaS contract are going to be surprised by Q3 actuals. The ones that build unit-economics thinking into their planning now are the ones that will use the platform's compounding growth to their advantage rather than against their budget.

---

FAQ

What is Agentforce ARR in 2026?

Salesforce reported Agentforce annual recurring revenue (ARR) at $1.2 billion in its Q1 FY27 earnings release on May 27, 2026, representing 205% growth year over year. The product launched in September 2024, making this one of the fastest enterprise software ARR ramps on record. An earlier report earlier in 2026 cited approximately $800 million in ARR, meaning the platform added roughly $400 million in ARR in approximately one quarter.

What is an Agentic Work Unit (AWU)?

An Agentic Work Unit is Salesforce's consumption billing metric for Agentforce and Slack AI agent activity. Each AWU represents a discrete task completion by an AI agent. Examples include routing an inbound lead, drafting a follow-up email, summarizing a meeting transcript, or pulling a data enrichment record. Salesforce charges customers per AWU consumed, which means total cost scales with the volume of agent activity rather than the number of users. As of Q1 FY27, 3.8 billion AWUs had been delivered to date, with growth of 111% quarter over quarter.

How should sales ops budget for consumption-priced AI agents in 2026?

The key shift from seat-based to consumption-based pricing is that your AI budget now scales with workflow volume, not headcount. Practical budgeting steps: (1) pull your current AWU consumption and calculate cost per deal and cost per qualified lead; (2) build a usage forecast that accounts for new pilots planned for Q3 and Q4; (3) apply a 1.5x-2x growth buffer relative to current consumption, based on the 111% quarter-over-quarter growth rate observed across the Salesforce customer base; (4) separate your AWU budget from your seat license budget and present them as distinct line items in your FY planning model. Organizations that model AI costs as a fixed line item will consistently underestimate spend in high-growth quarters.