What is the right Agentforce attach goal for 2027?

Direct Answer

Target 28-32% Agentforce attach by end of 2027 — balancing Marc's implicit 35-45% bull case with executable ops. This assumes post-Sept 2024 launch acceleration (currently 8-15% estimated Q4 FY26), requires 4 non-negotiable conditions: (1) product-market fit beyond 5-seat pilots (multi-org tooling), (2) integration economics that don't cannibalize net new (pricing clarity), (3) field comfort with AI-in-workflow (sales leader co-sell cadence), (4) attach-specific compensation for ASMs (quota relief or carry credit).

Why 35%+ Is Achievable

• Replace-not-add narrative: Agentforce positioning as Copilot+Slack+HubSpot replacement (not additive seat tax) unlocks land-expand inside existing contracts; 35% assumes 3-4 users-per-org attach vs current 0.5-1.
• AI competency tax: Sales leaders fear Copilot/Claude commoditization; Agentforce-as-diff positions Salesforce as domain moat — higher attach when risk-of-replacement drives renewal discounts.
• Revenue math: Each 1pct attach = ~$70-90M annual (off ~$26B sub base); 35% = +$2.3-3.1B incremental; Marc's org can absorb aggressive field incentives to hit that ceiling.
• Einstein AI momentum: Existing Einstein loyalty (forecasting, lead-scoring) primes install base; Agentforce is Einstein-for-execution, not new stack — lower switching friction than Copilot adoption.
• Workflow stickiness: Agents running deal review, forecast hygiene, activity logging create lock-in beyond traditional org structure; reduces churn risk if attach > 25%.

Why 18% Is the Bear Case

• Pilot purgatory: Sales teams adopt Agentforce in proof-of-concept (5-seat), declare success, stall at adoption — common AI trap. 18% attach = perpetual pilot, limited expansion revenue.
• Pricing friction: If Agentforce seats cost $50-80/mo incremental (not bundled), CFOs reject org-wide deploy; attach stays pegged to innovation-budget buyers, not core ops.
• AI skepticism tax: Sales reps (esp. high-tenure) still distrust outcome prediction; Agentforce perceived as job-threat vs. productivity tool — manager-down adoption fails without exec sponsorship contracts.
• Einstein fatigue: Orgs already burned by forecasting accuracy (Einstein still ~60-70% precision) view Agentforce as "same dataset, new UI" — skepticism bleeds into attach metrics.
• Competitor optionality: Copilot for Sales + Gong + Outreach agents mature in parallel; attach commoditizes if Salesforce doesn't own deal-outcome control (which requires Einstein accuracy first).

What Salesforce Should Execute

1. Attach-as-commission currency: Design 2027 ASM comp to include Agentforce attach bonus (e.g., +2% carry on orgs >25% attach) — direct P&L incentive, not just org MBO.
2. Multi-agent licensing model: Bundle Agentforce in Core + Professional (not Premium-only) at 2-4 included seats; monetize expansion seats at margin-positive rate — lowers perception of tax.
3. Industry-specific playbooks: Ship Agentforce configs for 8-10 verticals (tech, healthcare, financial services) with pre-built agent logic for deal review, forecast scrub, activity enforcement — reduces buyer education time.
4. Pavilion + Bridge Group co-sell: Run quarterly attach webinars (Pavilion pods, BG members) positioning Agentforce as rev-per-rep multiplier — peer credibility drives land, not just sales org messaging.
5. Klue competitive battle cards: Embed Agentforce vs Copilot/Outreach comparison in sales plays; equip reps to close attach against competing AI layers — defensible if Salesforce shows win rates.
6. Force Management outcome guarantees: Offer Salesforce-funded Force Management coaching for orgs attempting >20% attach — lowers customer risk, improves stickiness above pilot threshold.
7. Einstein accuracy sprint: Publish Q4 FY26 forecasting/deal-score precision targets; if Einstein hits 78%+ accuracy, position Agentforce as downstream beneficiary — trust-building before attach sale.
8. Champify seller-awards integration: Embed Agentforce agent suggestions into Champify seller rewards (gamification for high-touch reps using agents); attach becomes status symbol, not cost.

Attach Growth Roadmap

Bottom Line

28-32% attach by end of 2027 is operator-realistic — ambitious enough to unlock $2B+ revenue, grounded in 4 executable field motions (compensation redesign, licensing reform, playbook drops, vendor co-sell). Marc's 35%+ is the bull case contingent on Einstein accuracy sprint + competitor silence (neither guaranteed).

The bear case (18%) happens if Salesforce treats Agentforce as SKU, not leverage point — which is precisely why ASM comp redesign and Pavilion/Bridge co-sell are table-stakes, not bonus plays. De-risk by Q2 FY27 attach tracking (should be 15%+) or forecast reset required.