Yardstick Research tear-sheet / AI sales cohort
Salesforce-Einstein-Copilot
Identity
- Founded: Einstein Copilot GA'd in 2024 (Salesforce founded 1999; Einstein AI launched 2016; Copilot conversational layer shipped 2024) [THIRD-PARTY - widely documented]
- HQ: San Francisco, CA (415 Mission Street) [VENDOR-CLAIMED - https://www.salesforce.com/]
- Legal entity: Salesforce, Inc. (NYSE: CRM) [VENDOR-CLAIMED]
- Funding: Public company - market cap ~$230B as of May 2026 [THIRD-PARTY - public markets]
- Headcount: ~72,000 employees globally (FY2025 annual report) [THIRD-PARTY - SEC filing]
- Recent news (last 12 months): Agentforce (the renamed Einstein Copilot + autonomous agent framework) shipped GA in February 2025. Atlas reasoning engine added frontier LLM support (GPT-4o + Claude Sonnet 4) behind the Salesforce Trust Layer. Einstein Copilot for Sales rebranded into Agentforce for Sales. Salesforce Revenue Cloud + Copilot integration GA'd. [VENDOR-CLAIMED + THIRD-PARTY - multiple sources]
- Archetype: Conversational AI agent embedded natively in Salesforce CRM - uses Atlas reasoning + frontier LLMs (GPT-4o, Claude Sonnet 4) with full access to the Salesforce data graph, surfacing deal summaries, next-step recommendations, and email drafts without leaving the CRM interface.
Total score: 45.8 / 100
Scoring: equal-weight mean of 6 dimensions × 100, less pricing-transparency penalty (none = 0 pts).
- Stage fit:
- Foundation (<$5M ARR): no - requires an active Salesforce Sales Cloud subscription as prerequisite. Foundation-stage orgs not already on Salesforce face a CRM migration cost before any Copilot value accrues.
- Pilot ($5-$50M): conditional - viable if the org is already on Salesforce Sales Cloud Starter or Professional. Einstein Copilot is included in Enterprise and above; Pilot-stage orgs on lower tiers pay an add-on fee.
- Scale ($50-$500M): yes - core ICP for any Salesforce-committed org at this stage. Native CRM integration means zero data sync overhead; Atlas reasoning with Trust Layer addresses enterprise AI-governance requirements.
- Optimization ($500M+): yes - Salesforce's enterprise compliance posture (FedRAMP High, HIPAA BAA, ISO 27001, SOC 2 Type 2) and the breadth of the platform API make this the AI copilot of record for large Salesforce-anchored enterprises.
- One-line verdict: The native AI co-pilot for Salesforce-committed enterprises. Zero integration overhead, frontier-LLM quality (GPT-4o + Claude Sonnet 4) behind the Trust Layer. Structurally irrelevant to non-Salesforce orgs. Scores low on deliverability and setup time because the CRM implementation is a prerequisite, not a feature.
Headline numbers
| Metric | Value | Evidence |
|---|---|---|
| Free tier? | No standalone free tier; included in Salesforce Enterprise and above | [VENDOR-CLAIMED - https://www.salesforce.com/sales/pricing/] |
| Salesforce Sales Starter | $25/user/month (annual) - no Copilot | [VENDOR-CLAIMED - https://www.salesforce.com/sales/pricing/] |
| Salesforce Sales Professional | $80/user/month (annual) - limited Copilot | [VENDOR-CLAIMED - https://www.salesforce.com/sales/pricing/] |
| Salesforce Sales Enterprise | $165/user/month (annual) - Copilot included | [VENDOR-CLAIMED - https://www.salesforce.com/sales/pricing/] |
| Salesforce Sales Unlimited | $330/user/month (annual) - full Einstein + Agentforce | [VENDOR-CLAIMED - https://www.salesforce.com/sales/pricing/] |
| G2 score | 4.3/5 across 17,000+ Salesforce Sales Cloud reviews | [THIRD-PARTY - G2] |
| Customer count | 150,000+ customers globally (Salesforce platform) | [VENDOR-CLAIMED - https://www.salesforce.com/] |
Dimension scores
Equal-weight scoring - 6 dimensions, each 16.7% of the base score.
| Dimension | Score | Weighted | Evidence |
|---|---|---|---|
| Personalization quality | 3/4 | 12.5 | Atlas reasoning engine with GPT-4o + Claude Sonnet 4 backing produces CRM-context-aware deal summaries, email drafts, and next-step recommendations. Quality is strong for CRM-data-grounded personalization; does not match conversation-grounded specificity of Gong. [MEASURED - D1, yardstick-data] |
| Deliverability infrastructure | 0/4 | 0.0 | Einstein Copilot is a CRM AI layer, not a sending platform. Email drafts are composed within Salesforce and sent through connected email; no native deliverability stack. [MEASURED - D1, yardstick-data] |
| Ease of data integration & accuracy | 4/4 | 16.7 | Native access to the full Salesforce data graph - every object, every custom field, every relationship - with no sync overhead. AppExchange has 7,000+ integrations; Agentforce + Einstein APIs are the widest enterprise SDK surface in the cohort. [MEASURED - D1, yardstick-data] |
| Cost-per-seat efficiency | 1/4 | 4.2 | Einstein Copilot is included in Sales Enterprise ($165/user/month) and above. The base CRM cost is the dominant factor; Copilot itself does not add a separate per-seat fee at the Enterprise tier. However, $165/user/month as the entry point for full Copilot access is the highest CRM-entry price in the cohort. [MEASURED - D1, yardstick-data] |
| UI heuristics | 3/4 | 12.5 | The Copilot sidebar in Lightning Experience is well-integrated; reps do not need to leave the CRM. The conversational interface (ask a question, get a cited answer) is the UX differentiator. Underlying Salesforce UI complexity remains a non-trivial admin and rep onboarding burden. [MEASURED - D1, yardstick-data] |
| Setup time | 0/4 | 0.0 | Einstein Copilot requires a running Salesforce Sales Cloud Enterprise instance as prerequisite. Standard Salesforce enterprise implementation: 3-6 months. Copilot activation on a running instance: 1-2 weeks. The prerequisite implementation timeline is the structural bottleneck. [MEASURED - D1, yardstick-data] |
| Base total | 45.8 | ||
| Pricing-transparency penalty | - | 0 | None - pricing published at https://www.salesforce.com/sales/pricing/ |
| Headline score | 45.8 |
Pricing detail
[VENDOR-CLAIMED - https://www.salesforce.com/sales/pricing/]:
- Sales Starter: $25/user/month (annual) - basic CRM, no Copilot
- Sales Professional: $80/user/month (annual) - limited Copilot features
- Sales Enterprise: $165/user/month (annual) - Einstein Copilot included, full customization
- Sales Unlimited: $330/user/month (annual) - full Einstein suite + Agentforce, unlimited AI actions
- Einstein 1 Sales: $500/user/month (annual) - complete platform with Data Cloud and all AI capabilities
- Copilot add-on pricing: specific per-prompt / per-credit pricing applies for Agentforce autonomous agents beyond Copilot-included usage; see Salesforce pricing calculator
- No permanent free tier; 30-day trial available
Integrations
[VENDOR-CLAIMED - https://www.salesforce.com/]: - Native platform (complete access): All Salesforce Sales Cloud objects and custom fields via Salesforce data graph - AppExchange: 7,000+ partner integrations - AI models (via Trust Layer): GPT-4o (OpenAI), Claude Sonnet 4 (Anthropic), Salesforce native xGen models - Sales engagement: Outreach, Salesloft (deep integrations via AppExchange) - Marketing automation: Salesforce Marketing Cloud, Pardot - Cohort integrations: salesforce-sales-cloud-einstein, agentforce
Editorial assessment
Einstein Copilot's position in this cohort follows from one architectural fact: it is the only AI copilot in the report with zero integration overhead for Salesforce organizations. Every other AI tool in this cohort syncs data from Salesforce into its own system and runs inference there. Einstein Copilot runs inference on the Salesforce data graph natively - no sync, no latency, no data residency question. For a 100-rep Salesforce Enterprise org evaluating AI copilots, that eliminates an entire layer of implementation and compliance risk.
The Atlas reasoning engine running GPT-4o and Claude Sonnet 4 behind the Salesforce Trust Layer is the 2025 product story: frontier-class LLM access with enterprise data governance attached. The Trust Layer prevents customer data from being used to train external models, enforces access controls at the row-level-security layer, and logs every AI action for compliance audit. For financial services, healthcare, and public-sector Salesforce orgs, that combination is hard to match outside the Salesforce platform - third-party tools typically cannot meet Trust Layer compliance depth without significant custom integration work.
The zero scores on deliverability and setup time are architectural realities, not product defects. Einstein Copilot is not a sending platform (zero deliverability score is correct); it is a CRM intelligence layer that surfaces insights and drafts content for reps to send through their existing channels. The setup-time zero reflects the Salesforce implementation prerequisite: a Salesforce Enterprise instance that takes 3-6 months to stand up is the starting condition, not the finishing condition. Buyers who are already on Salesforce Enterprise experience a very different time-to-value profile - Copilot activation on a running instance is approximately 1-2 weeks. The rubric scores the full-buyer journey including the prerequisite, which is the right frame for a cohort buyer who is choosing between tools.
The cost-per-seat score (1/4) reflects the $165/user/month entry point for full Copilot access at the Enterprise tier - the highest CRM-layer entry price in the cohort. The value argument exists - the architecture eliminates a sync layer - but the dollar figure is a real barrier for cost-sensitive buyers.
Best for
- Stage: Scale ($50-$500M) and Optimization ($500M+). Conditional for Pilot if already on Salesforce Enterprise.
- Company profile: Salesforce-committed enterprise organizations, 25+ reps, where the CRM is the authoritative system of record and where AI governance / Trust Layer compliance is a procurement requirement.
- Skip if: You are (a) not on Salesforce or actively migrating away, (b) a Foundation or Pilot-stage org that has not yet invested in a full Salesforce implementation, (c) primarily a conversation-intelligence buyer (Gong is the right call), or (d) comparing AI copilots on personalization quality alone without weighting integration-depth.
Right-of-reply
Salesforce Einstein Copilot received this tear-sheet seven calendar days before publication of the Yardstick Research 2026 Yardstick Report, including all measured numbers, sample outputs, and editorial assessment. Salesforce was given the opportunity to flag factual errors - incorrect pricing, misquoted feature availability, outdated screenshots, factual misstatement in the editorial assessment. Salesforce was not given the opportunity to request a score revision, dispute the rubric or its weights, withdraw from inclusion, negotiate ranking placement, or suggest changes to the editorial assessment beyond factual correction. Where a vendor flagged a factual correction, the correction was applied if verified and noted here; where a vendor disputed scoring, the dispute is recorded in the appendix but the score stands. Silence from the vendor during the right-of-reply window was treated as no objection.
Sources
- https://www.salesforce.com/
- https://www.salesforce.com/sales/pricing/
- https://www.salesforce.com/sales/ai-sales-assistant/
- https://help.salesforce.com/s/articleView?id=sf.einstein_copilot.htm