Salesforce
Q2 FY27 (August 26) was a headline beat - revenue $11.3B (+11% YoY, including $456M from acquired Informatica), cRPO $33.5B (+14%), non-GAAP operating margin 34.1%, FY27 guidance raised to $46.1-46.4B - but the AI metric got harder to read, not easier. Effective Q2, Salesforce redefined Agentforce ARR to include 'our AI offerings, Slackbot and Headless 360' and reported more than $1.5B (+240% YoY) with no restated Q1 baseline published, so the move from Q1's $1.2B is not a like-for-like comparison; Data 360 ARR of $2.4B similarly absorbs acquired Informatica Cloud, making the combined $3.9B AI ARR figure partly definitional and partly inorganic. All are company-defined, unaudited metrics. That leaves the July adoption-depth critique - roughly 23,000 of 150,000 customers actually using Agentforce - unresolved rather than answered, with headline revenue growth of 11% the slowest in Salesforce's public history. The offsetting positive is cRPO growing 14% against revenue's 11%, the one forward indicator pointing to re-acceleration.
Scenarios
150K-customer install base and 30+ years of CRM workflow data provide integration moat. Agentforce $1.2B ARR at 205% YoY is the fastest commercial ramp of any prior Salesforce product launch; Slack MCP 1M users in 6 weeks signals developer adoption at scale.
AI commoditizes CRM data management via MCP/tool use. Per-seat revenue contraction may outpace usage-based upsell. 70% of production Agentforce deployments stall due to data quality or technical debt. The 28.6T-token figure is consumption, not outcomes — without NRR confirmation, the activity could be expensive without being valuable.
Key Factors to Watch
- ●Agentforce ARR $1.2B Q1 FY27 (+205% YoY, +50% QoQ from $800M) — named earnings print; fastest commercial ramp in Salesforce product history
- ●Slack MCP 1M users in 6 weeks; 50%+ of Agentforce bookings from existing-customer expansion — distribution-to-AI conversion is operational
- ●70%+ of Fortune 100 joint Agentforce + Data Cloud customers; 28.6T tokens Q1 (+152% QoQ) confirms platform depth of use
- ●AWU and token metrics are activity-not-outcomes — $1.2B ARR is the only anchor number that matters; watch NRR and churn in Q2 print
- ●Per-seat CRM pricing erosion window 18-24 months; Agentforce at ~10% of total revenue — growth spectacular, share still minority of $45.9B FY27 base
Score History
| Date | Score | Direction | Note |
|---|---|---|---|
| 2026-08-22 | 6.9 | Positive | Score 7.0->6.9 (ai_adoption_maturity 9->8). Two July sell-side downgrades (KeyBanc Jul 9; Morgan Stanley Jul 21, PT cut 35% to $185) found only ~23,000 of 150,000 customers (~15%) actually using Agentforce despite the ARR headline, confirming the 'activity not outcomes' bear case with primary customer research; Salesforce responded by forcing Agentforce default-on for all orgs in August. |
| 2026-05-30 | 7.0 | Positive | Score 6.7→7.0 (md 7→8, aam 8→9). Q1 FY27 print confirms Agentforce $1.2B ARR (+205% YoY, +50% QoQ from $800M); 3.8B AWUs (+111% QoQ); 28.6T tokens (+152% QoQ); Slack MCP 1M users in 6 weeks. md 7→8 on durable Slack-MCP distribution lock-in. aam 8→9 on AWU acceleration + 50%+ existing-customer expansion bookings. Materiality gate cleared by named earnings event with 50% sequential ARR beat. |
| 2026-04-10 | 6.7 | Positive | Score 6.6→6.7 (rounding drift correction — dimensions unchanged, formula recomputed to match stored value) |
| 2026-03-08 | 6.6 | Positive | Score 6.9→6.6 (reweight + dr 7→6). dr corrected: Agentforce 29K deals + usage-based pricing = active disruption response, not passive target |
| 2026-03-08 | 6.9 | Positive | Score 6.6->6.9 (are 8->7, md 5->7). External research cross-ref: AI revenue only 7% ($2.9B/$41.5B), but 21.7% CRM market share and Einstein Trust Layer = strong moat (md was 5, far too low) |
| 2026-03-08 | 6.6 | Positive | Blind spot review: sector corrected from customer-service to software-development, ai_revenue_exposure 7→8, ai_adoption_maturity 7→8 |
| 2026-03-08 | 6.3 | Positive | Initial assessment from batch 1 research |
Software Peers
Last researched: 2026-09-19
This is research and analysis, not financial advice. Scores reflect AI impact potential, not investment recommendations.