# Salesforce

> Source: https://aiwiki.ai/wiki/salesforce
> Updated: 2026-06-21
> Categories: AI Companies
> From AI Wiki (https://aiwiki.ai), a free encyclopedia of artificial intelligence. Quote with attribution.

Salesforce, Inc. is an American cloud software company headquartered in Salesforce Tower, San Francisco, best known for its customer relationship management (CRM) platform and, since 2024, for one of the most aggressive incumbent bets on [agentic AI](/wiki/agentic_ai) in enterprise software. Founded in March 1999 by former Oracle executive [Marc Benioff](/wiki/marc_benioff) together with Parker Harris, Dave Moellenhoff, and Frank Dominguez, the company helped establish the software-as-a-service model under the slogan "The End of Software" [1]. IDC has ranked it the world's largest CRM vendor by revenue for 13 consecutive years [2]. Salesforce reported revenue of $41.5 billion for fiscal 2026 (ended January 31, 2026) and $11.13 billion in its first quarter of fiscal 2027 (ended April 30, 2026, up 13 percent year over year), and brands itself "the world's #1 AI CRM," a positioning built on three layers: the [Einstein](/wiki/salesforce_einstein) AI services introduced in 2016, the [Agentforce](/wiki/agentforce) agent platform launched in 2024, and the Data 360 data foundation [3][28].

## Overview

Salesforce sells subscription applications for sales, customer service, marketing, and commerce, plus a low-code development platform and acquired products including [Slack](/wiki/slack), Tableau, and MuleSoft. According to IDC, it held 20.7 percent of the worldwide CRM market in 2024, the largest share of any vendor [2]. In fiscal 2026 the company generated $41.5 billion in revenue, up 10 percent year over year (including a $399 million contribution from the newly acquired Informatica), with a GAAP operating margin of 20.1 percent, operating cash flow of $15.0 billion, and remaining performance obligation of $72.4 billion [3]. In the first quarter of fiscal 2027 (ended April 30, 2026), revenue reached $11.13 billion, up 13 percent year over year, with a non-GAAP operating margin of 34.8 percent and Agentforce annualized revenue surpassing $1.2 billion [28].

Since 2023 Salesforce has reorganized its product story around what Benioff calls the "agentic enterprise." Announcing fiscal 2026 results in February 2026, he said the company had been "rebuilt... to become the operating system for the Agentic Enterprise, bringing humans and agents together on one trusted platform" [3]. Salesforce guided fiscal 2027 revenue to $45.9 billion to $46.2 billion and raised its fiscal 2030 target to $63 billion, with agentic AI and data products positioned as the main growth engine [3][28].

## History and founding

### When was Salesforce founded?

Benioff, then a senior vice president at Oracle, started Salesforce in March 1999 with Harris, Moellenhoff, and Dominguez in a rented apartment on San Francisco's Telegraph Hill, with the idea of delivering enterprise sales software through a website rather than installed licenses [1]. Early backing came from angels including Oracle founder Larry Ellison, who put in roughly $2 million, and CNET founder Halsey Minor [1]. The company listed on the New York Stock Exchange in June 2004 under the ticker CRM, selling 10 million shares at $11 and raising $110 million [4].

Salesforce grew into one of the largest software companies in the world partly through acquisitions, including MuleSoft (2018, $6.5 billion), Tableau (2019, $15.7 billion), and Slack (2021, $27.7 billion) [1]. After activist-investor pressure in 2022 and 2023 it cut costs sharply, including a roughly 10 percent workforce reduction in January 2023, and reset itself around profitable growth.

Its AI lineage began earlier than the current wave. Salesforce bought relationship-intelligence startup RelateIQ in 2014, and in April 2016 acquired the deep learning startup MetaMind, whose founder [Richard Socher](/wiki/richard_socher) became Salesforce's chief scientist [5]. Socher left in 2020 to found You.com and was succeeded as chief scientist by Stanford professor Silvio Savarese [6]. The MetaMind deal seeded Salesforce AI Research and the September 2016 launch of Einstein, which the company described as the first comprehensive AI platform for CRM [5][7].

## Products and technology

### Einstein: from predictive to generative (2016 to 2023)

Einstein, unveiled in September 2016, embedded machine learning features such as lead and opportunity scoring, forecasting, and service chatbots directly into Salesforce's clouds. By early 2023 the company said Einstein was delivering more than 200 billion AI-powered predictions per day [7]. In March 2023 Salesforce announced Einstein GPT, billed as the first generative AI for CRM, combining its proprietary models with external [large language models](/wiki/large_language_model) including [OpenAI](/wiki/openai)'s, to draft emails, service replies, and code inside the CRM [7]. Generative features were governed by the Einstein Trust Layer, which grounds outputs in customer data, masks sensitive fields, and enforces zero-retention arrangements with model providers [8]. A conversational assistant, Einstein Copilot, reached customers in 2024 before being folded into the Agentforce brand [8].

### What is Agentforce? (2024 to present)

Agentforce, unveiled at the Dreamforce conference in September 2024 and generally available from late October 2024, is a layer on the Salesforce platform for building and running autonomous AI agents that handle tasks in service, sales, marketing, and commerce. Benioff called it "the Third Wave of AI, advancing beyond copilots to a new era of intelligent, low-hallucination agents" [8]. Agents are orchestrated by what Salesforce calls the Atlas Reasoning Engine and can reuse existing automation assets (Flows, Apex code, APIs) as agent actions; early adopters cited by the company included the publisher Wiley, which reported a more than 40 percent improvement in case resolution [8]. Launch pricing was about $2 per conversation [8]; in May 2025 Salesforce added consumption-based "Flex Credits" sold in $500 packs of 100,000 credits, with a standard agent action consuming 20 credits, roughly $0.10 [9].

The platform iterated quickly: Agentforce 2.0 (December 2024) repositioned the product as a "digital labor platform" [10]; Agentforce 3 (June 2025) added a Command Center for observability and native support for the [Model Context Protocol](/wiki/model_context_protocol) [11]; and in October 2025 Salesforce rebranded its entire stack as Agentforce 360, with Slack as the conversational interface and Data Cloud renamed Data 360 [12][13]. Alongside that launch Salesforce expanded partnerships with both frontier labs: OpenAI models including [GPT-5](/wiki/gpt_5) became available inside Agentforce 360 with commerce and Slack hooks into [ChatGPT](/wiki/chatgpt), while [Anthropic](/wiki/anthropic)'s [Claude](/wiki/claude) became a preferred model for regulated industries such as financial services, served through [Amazon Bedrock](/wiki/amazon_bedrock) inside Salesforce's virtual private cloud; Salesforce also said it would roll out Claude Code across its engineering organization [12][14].

Company-reported traction has grown fast, though from a small base relative to total revenue:

| Metric (company-reported) | End of FY2025 (Jan 31, 2025) | End of FY2026 (Jan 31, 2026) | Q1 FY2027 (Apr 30, 2026) |
|---|---|---|---|
| Fiscal-period revenue | $37.9 billion, up 9% [15] | $41.5 billion, up 10% [3] | $11.13 billion, up 13% [28] |
| Agentforce deals since launch | about 5,000, incl. 3,000+ paid [15] | more than 29,000 [3] | not separately broken out [28] |
| Agentforce ARR | not disclosed | $800 million, up 169% [3] | more than $1.2 billion [28] |
| Combined AI and data ARR | $900 million "Data Cloud & AI", up 120% [15] | more than $2.9 billion, incl. $1.1 billion Informatica Cloud [3] | $3.4 billion combined AI, Data 360, and Informatica [28] |

Salesforce also introduced "Agentic Work Units" as a usage measure, reporting 2.4 billion delivered across Agentforce and Slack by February 2026 (up 57 percent quarter over quarter) on nearly 20 trillion tokens processed; by Q1 fiscal 2027 cumulative Agentic Work Units had reached 3.8 billion (up 111 percent quarter over quarter) on 28.6 trillion tokens processed to date, and Salesforce said more than 60 percent of fourth-quarter Agentforce and Data 360 bookings came from existing-customer expansion [3][28].

### Data 360 and the Informatica acquisition

Salesforce's data layer has been renamed repeatedly: launched as Genie at Dreamforce 2022, it became Data Cloud in 2023 and Data 360 on October 14, 2025, when it was folded under the Agentforce 360 umbrella [13]. The product unifies customer records across systems, including "zero copy" federation that queries external warehouses without duplicating data, and serves as the grounding layer for agents. In fiscal 2026 Data 360 ingested 112 trillion records, up 114 percent year over year, 53 trillion of them via zero copy [3].

To deepen this layer, Salesforce agreed in May 2025 to acquire data-management vendor Informatica for about $8 billion ($25 per share), its largest deal since Slack, explicitly framed as infrastructure for "agentic AI" through data catalogs, lineage, governance, and master data management [16]. The acquisition closed on November 18, 2025 [17].

### What AI models has Salesforce built in-house?

Salesforce AI Research, which grew out of MetaMind [5], has released a string of open and proprietary models: the [CodeGen](/wiki/codegen) program-synthesis language models, the [BLIP](/wiki/blip) family of [vision-language models](/wiki/vision_language_model) (continued as xGen-MM/BLIP-3), the xGen text model series, and xLAM "large action models" specialized for [function calling](/wiki/function_calling), alongside fine-tuned commercial models such as xGen-Sales that power Agentforce tasks [18].

The xLAM family, introduced in September 2024, is the research lineage most directly tied to the agent business: Salesforce describes large action models as models built "to predict and perform the next action" rather than the next word, optimized for tool use, planning, and execution by AI agents [29]. The first xLAM release spanned five models from 1 billion to 8x22B parameters in both dense and mixture-of-experts architectures, with the 8x22B variant topping the Berkeley Function-Calling Leaderboard and the 7B model outperforming much larger systems including GPT-4 and Claude 3 Opus on that benchmark despite its smaller size [29][30]. In April 2025 Salesforce AI Research released the xLAM-2 series (1B to 70B), trained with its APIGen-MT multi-turn data pipeline; xLAM-2-70b-fc-r and xLAM-2-32b-fc-r took the top two positions on the BFCL v3 leaderboard with overall accuracies of 78.19 percent and 75.83 percent, and xLAM-2-70b-fc-r reached 75.12 percent multi-turn accuracy on [tau-bench](/wiki/tau_bench), surpassing frontier models such as GPT-4o and Claude 3.5 [31]. On the multimodal side, the August 2024 xGen-MM (BLIP-3) release shipped open 4B and 14B pre-trained and instruction-tuned models alongside trillion-token training datasets such as MINT-1T, positioned as a fully open alternative to closed vision-language systems [32].

| Date | AI milestone |
|---|---|
| September 2016 | Einstein predictive AI announced [7] |
| March 2023 | Einstein GPT generative layer, OpenAI partnership [7] |
| August 2024 | xGen-MM (BLIP-3) open multimodal models released [32] |
| September 2024 | xLAM large action models introduced; top of Berkeley Function-Calling Leaderboard [29][30] |
| September to October 2024 | Agentforce unveiled at Dreamforce; general availability [8] |
| April 2025 | xLAM-2 series with APIGen-MT pipeline tops BFCL v3 [31] |
| May 2025 | Flex Credits consumption pricing [9] |
| June 2025 | Agentforce 3 with Command Center and MCP support [11] |
| October 2025 | Agentforce 360 rebrand; Data Cloud becomes Data 360; expanded OpenAI and Anthropic partnerships [12][13][14] |
| November 18, 2025 | Informatica acquisition closes [17] |
| May 27, 2026 | Q1 FY2027 results: Agentforce ARR passes $1.2 billion [28] |

## Funding

As a 1999-vintage startup, Salesforce raised comparatively modest private capital, anchored by Ellison's and Minor's angel money, before its $110 million IPO in June 2004 [1][4]. As a mature public company it now returns cash to shareholders: in fiscal 2026 it paid out $14.3 billion via buybacks and dividends, raised its quarterly dividend to $0.44, and announced a new $50 billion share repurchase authorization [3].

Salesforce is also a significant AI investor through Salesforce Ventures, which launched a $250 million generative AI fund in June 2023, doubled it to $500 million, and added a second $500 million fund in September 2024, bringing committed AI capital to $1 billion. Its portfolio includes [Anthropic](/wiki/anthropic), [Cohere](/wiki/cohere), [Mistral AI](/wiki/mistral_ai), [Hugging Face](/wiki/hugging_face), [Runway](/wiki/runway), and [Together AI](/wiki/together_ai) [19].

## AI and the workforce

Salesforce has become a closely watched case study in AI-driven workforce change, largely because of Benioff's own statements. In December 2024 he said the company would hire no new software engineers in 2025, crediting Agentforce and other AI tooling with raising engineering productivity "by more than 30%," while adding 1,000 to 2,000 salespeople to sell AI products [20]. In June 2025 he told Bloomberg's The Circuit that AI was "doing 30% to 50% of the work at Salesforce now," across engineering, coding, and support [21].

In September 2025 Benioff said Salesforce had cut its customer-support organization from about 9,000 to roughly 5,000 people because AI agents now handle a large share of inquiries, remarking "I need less heads"; the company said many affected workers were redeployed to other roles [22]. Salesforce reported that agents on its help portal handled about 1.5 million customer conversations in nine months, that support costs fell 17 percent, and that interactions were split roughly half and half between humans and agents, with executives acknowledging the technology's limits [22][23]. Because Salesforce sells the very agents it credits for these cuts, its disclosures are cited both as early evidence of AI labor displacement and as marketing for "digital labor," and outside observers cannot fully verify the internal percentages [21][22].

## Reception and significance

Salesforce's agent push is among the largest deployments of agentic AI in enterprise software, and its metrics ($1.2 billion Agentforce ARR, $3.4 billion combined AI and data ARR, 3.8 billion agentic work units) are widely used as a barometer for whether AI agents generate real revenue [28]. The market verdict has been mixed: the stock fell about 31 percent in calendar 2025, making it one of the two worst performers in the Dow Jones Industrial Average that year [24], and by early 2026 it traded near its cheapest forward earnings multiple since the 2004 IPO, around 19 times versus a ten-year average of 47, as investors worried that agents could erode seat-based software pricing and that AI-native rivals such as [Sierra](/wiki/sierra) and [Decagon](/wiki/decagon), along with [Microsoft](/wiki/microsoft) and ServiceNow, would compress its core franchise [24][25]. Reporting record Q1 fiscal 2027 results in May 2026, Benioff said "Agentic AI is the biggest growth opportunity for our customers, and for Salesforce," adding that the company was "the #1 Agentic CRM, with Agentforce now powering every Customer 360 application" [28].

The company has meanwhile pushed into new arenas: in September 2025 it launched Missionforce, a national-security business unit led by Government Cloud chief Kendall Collins, applying AI to defense logistics, personnel, and decision support [26]. Benioff himself drew controversy in October 2025 when he told The New York Times he supported sending National Guard troops to San Francisco ahead of Dreamforce; after backlash that included investor Ron Conway resigning from the Salesforce Foundation board, he apologized and withdrew the suggestion [27].

For the AI industry, Salesforce matters as the clearest test of whether an incumbent application vendor can convert generative AI from a feature into a consumption business. Its fiscal 2027 guidance assumes organic growth re-accelerates in the second half of the year on Agentforce and Data 360 adoption [3][28]; whether that happens will shape how the rest of the SaaS sector, and its investors, price the agent era.

## References

1. Salesforce, "The History of Salesforce," Salesforce News (company history, updated through 2024).
2. Salesforce, "Salesforce Ranked #1 CRM Provider for 13th Consecutive Year," Salesforce News, 2026 (citing IDC Worldwide Semiannual Software Tracker; 20.7% 2024 share figure from the 2025 edition).
3. Salesforce, "Salesforce Delivers Record Fourth Quarter Fiscal 2026 Results," press release, February 25, 2026.
4. Destination CRM, "Salesforce.com IPO Raises $110 Million," June 2004.
5. Drew Olanoff, "Salesforce acquires MetaMind," TechCrunch, April 4, 2016.
6. Khari Johnson, "Salesforce chief scientist Richard Socher leaves to start his own company," VentureBeat, July 2020.
7. Salesforce, "Salesforce Announces Einstein GPT, the World's First Generative AI for CRM," press release, March 7, 2023.
8. Salesforce, "Salesforce's Agentforce Is Here: Trusted, Autonomous AI Agents to Scale Your Workforce," press release, October 29, 2024.
9. Salesforce, "Salesforce Introduces New Flexible Agentforce Pricing to Accelerate the Digital Labor Revolution," press release, May 15, 2025.
10. Salesforce, Agentforce 2.0 announcement, press release, December 2024.
11. Salesforce, Agentforce 3 announcement, press release, June 2025.
12. Salesforce, "Welcome to the Agentic Enterprise: With Agentforce 360, Salesforce Elevates Human Potential in the Age of AI," press release, October 13, 2025.
13. Salesforce Ben, "Salesforce Data Cloud Renamed to 'Data 360' As Part of 'Agentforce 360'," October 2025.
14. Anthropic, "Anthropic and Salesforce expand partnership to bring Claude to regulated industries," October 14, 2025.
15. Salesforce, "Salesforce Announces Fourth Quarter and Fiscal Year 2025 Results," press release, February 26, 2025.
16. CNBC, "Salesforce to acquire data management company Informatica in $8 billion deal," May 27, 2025.
17. Salesforce, "Salesforce Completes Acquisition of Informatica," press release, November 18, 2025.
18. Salesforce, "Salesforce Delivers Next-Generation AI Models to Power Agentforce," Salesforce News, September 2024.
19. Salesforce Ventures, "Salesforce Ventures Expands AI Investment to $1 Billion," September 2024.
20. Salesforce Ben, "Salesforce Will Hire No More Software Engineers in 2025, Says Marc Benioff," January 2025 (citing Benioff on the 20VC podcast, December 2024).
21. Fox Business, "Salesforce CEO reveals AI now handles 30% to 50% of company's work," June 2025 (citing Bloomberg, The Circuit with Emily Chang).
22. Fortune, "Salesforce CEO Marc Benioff says his company has cut 4,000 customer service jobs as AI steps in: 'I need less heads'," September 2, 2025.
23. CX Dive, "Salesforce still sees a place for live customer service agents after massive cuts," September 2025.
24. The Motley Fool, "Prediction: 2025's Second-Worst-Performing Dow Jones Stock Will Beat the Market in 2026," November 30, 2025.
25. Bloomberg (via Yahoo Finance), "Salesforce's Stock Is Historically Cheap as AI Risk Takes a Toll," 2026.
26. TechCrunch, "Salesforce launches 'Missionforce,' a national security-focused business unit," September 16, 2025.
27. CNBC, "Salesforce CEO apologizes for saying Trump should send National Guard to San Francisco," October 17, 2025.
28. Salesforce, "Salesforce Delivers Record First Quarter Fiscal 2027 Results," press release, May 27, 2026.
29. Salesforce, "Actions Speak Louder Than Words: Introducing xLAM, Salesforce's Family of Large Action Models," Salesforce blog, September 2024.
30. Zhiwei Liu et al., "xLAM: A Family of Large Action Models to Empower AI Agent Systems," arXiv:2409.03215, September 2024.
31. Salesforce AI Research, "APIGen-MT and the xLAM-2 Model Series," April 2025 (BFCL v3 and tau-bench results), with arXiv:2504.03601.
32. Salesforce AI Research, "xGen-MM (BLIP-3): A Family of Open Large Multimodal Models," arXiv:2408.08872, August 2024.