AI Workflow Automation Statistics 2026: 80 Sourced Numbers

37% of respondents can attribute any EBIT impact to their AI use, and that share is essentially flat against the previous wave. The group McKinsey calls high performers, the ones attributing at least 5% of profit to AI, hasn’t moved either. It’s 6%, the same 6% as the previous wave. Set those against the 88% of organizations that McKinsey’s November 2025 wave found using AI in at least one business function, and the shape of the problem is obvious.

Individual people are getting faster. 80% of respondents who use AI in their jobs say it improved their own productivity. That improvement is stopping somewhere between the desk and the income statement, and finding where it stops is what the money in this category should be chasing.

Below are 80 AI workflow automation statistics for 2026, grouped by the question each one answers. Every figure was traced back to the organization that produced it: survey PDFs and press releases, arXiv preprints, peer-reviewed journal articles, SEC-filed earnings releases, and company funding announcements. Where a source is a vendor measuring a market it sells into, I say so on the line rather than burying it in a footnote.

The market-size figures you’ve seen in every other roundup aren’t here. The Sources and Methodology section at the bottom explains exactly why, along with everything else that got cut.

Key Takeaways

• Adoption is near universal and profit impact is flat. 88% of organizations used AI somewhere in McKinsey’s November 2025 wave. In its August 2026 follow-up, 37% of respondents attribute any EBIT impact and 6% qualify as high performers, both flat on the previous wave.
• The productivity gain is real and it’s personal. 80% of AI users say their own output improved. The financial impact at the organization level hasn’t followed.
• Executives and employees describe different companies. 97% of executives say their company deployed AI agents in the past year. 52% of employees say they use them.
• The cleanest productivity number is 15%, not 55%. The peer-reviewed study of 5,172 support agents found 15%. The famous 55.8% figure came from a 95-person vendor-funded lab trial on a greenfield coding task.
• Coding gains shrink on the way to shipped software. Autonomous coding agents raise commits by 240%. That falls to 80% for projects started and 30% for actual releases.
• Agent scaling is a big-company story. 40% of respondents at organizations above $1 billion in revenue are scaling AI agents, up from 27% in the previous wave.
• Most vendor ARR figures are estimates wearing a disclosure’s clothes. Salesforce files its numbers. Zapier and Workato don’t, and what circulates for them comes from third-party trackers.

Top AI Workflow Automation Statistics for 2026

If you carry four numbers out of this page and into your next planning meeting, carry these.

The Four Numbers Worth Quoting

1. 88% of organizations report using AI regularly in at least one business function, up from 78% a year earlier (McKinsey, The state of AI, November 2025 wave, n=1,993 across 105 nations).

2. 37% of respondents attribute at least some EBIT impact to their AI use, about the same share as the previous wave (McKinsey, The state of AI in 2026: On the road to ROI, 25 August 2026, n=1,719).

3. 6% qualify as AI high performers, meaning they attribute at least 5% of EBIT to AI and call the impact significant. That share has stayed flat since the previous survey (McKinsey, August 2026).

4. 80% of respondents who use AI in their roles say it has improved their individual productivity (McKinsey, August 2026).

The Gap Between Deploying And Using

5. 97% of executives say their company deployed AI agents in the past year, while 52% of employees say they actually use them (WRITER, April 2026, n=2,400 split evenly between C-suite and employees).

6. 59% of companies invest more than $1 million a year in AI technology, and 29% report significant ROI from generative AI (WRITER, April 2026).

That 45-point gap between what executives report and what employees report is worth more than either number on its own. WRITER sells an enterprise agent platform, so read the framing with that in mind. A gap that wide is hard to put down to question wording, because the same survey asked both groups at the same time.

How Far Adoption Has Actually Gone

Adoption is the most-quoted and least-useful metric here, because every survey defines it differently. Read these as a range, not a consensus.

The Cross-Industry Picture

7. About one-third of organizations had begun scaling AI enterprise-wide as of the November 2025 wave, against the 88% who used it somewhere (McKinsey, November 2025).

8. 40% of respondents at organizations above $1 billion in annual revenue say they are scaling AI agents, up from 27% in the previous wave (McKinsey, August 2026).

9. 20% of respondents say AI-related operating costs have constrained their use of the technology (McKinsey, August 2026).

10. 66% of organizations report productivity or efficiency gains from AI (Deloitte, State of AI in the Enterprise 2026, published November 2025, n=3,235 leaders across 24 countries, fielded August to September 2025).

11. Insufficient worker skills is the single biggest barrier to integrating AI into existing workflows (Deloitte, November 2025).

12. 34% of organizations are starting to use AI to deeply transform, meaning new products or business models (Deloitte, November 2025).

13. 30% are redesigning key processes around AI, and 37% are using it at a more surface level (Deloitte, November 2025).

14. Worker access to AI rose by 50% during 2025 (Deloitte, November 2025).

15. The number of companies with at least 40% of their AI projects in production is set to double within six months (Deloitte, November 2025).

Deloitte and McKinsey both sell AI implementation work, and both surveys draw on leaders already piloting or implementing generative AI. They’re the largest cross-industry trackers with disclosed methodology, which is why they’re here. Neither is a random sample of the economy.

What Knowledge Workers Report

16. 75% of knowledge workers use AI at work (Microsoft and LinkedIn, 2024 Work Trend Index, May 2024, n=31,000 across 31 countries).

17. 46% of those users started less than six months before the survey, meaning generative AI use nearly doubled in half a year (Microsoft and LinkedIn, May 2024).

18. 90% of AI users say it helps them save time, and 85% say it helps them focus on their most important work (Microsoft and LinkedIn, May 2024, self-reported).

19. 79% of leaders agree AI adoption is a business imperative, while 60% say their organization lacks a plan or vision for implementing it (Microsoft and LinkedIn, May 2024).

That’s a 2024 survey and it’s quoted here as one. Microsoft hasn’t re-run the 75% question with the same methodology since, so anybody presenting that figure as a 2026 measurement is stretching it two years past its date.

What Marketers Report

20. 82% of marketers name reducing the time they spend on repetitive, data-driven tasks as their top AI goal (Marketing AI Institute, 2025 State of Marketing AI Report, May 2025, n=almost 1,900).

21. AI maturity splits 40% Understanding, 46% Piloting and 14% Scaling (Marketing AI Institute, May 2025).

22. 62% cite lack of training and education as a top barrier to AI adoption (Marketing AI Institute, May 2025).

23. 75% of marketing leaders whose organizations invested in AI say the investment yielded a positive ROI (HubSpot, AI Trends for Marketers, January 2025, n=1,000+).

24. 65% of marketing leaders say their team plans to increase investment in AI and automation tools over the course of the year (HubSpot, January 2025).

25. 94% of marketers plan to use AI in their content creation processes, including blog articles (HubSpot, State of Marketing Report 2026, n=3,400 marketers globally).

26. Nearly 30% of marketers report decreased search traffic as consumers turn to AI tools (HubSpot, State of Marketing Report 2026).

The Marketing AI Institute sample comes from its own newsletter, podcast and webinar audience, which the report itself acknowledges skews toward people already committed to AI. HubSpot sells Breeze agents. Both are directionally useful and neither is a population estimate. If that search-traffic figure is the one that worries you, the GEO and AEO statistics page goes deeper on where those clicks went.

Agents In Production: What The Surveys Actually Say

“Agents in production” is the most abused phrase in this category, because the definition moves with whoever is asking.

The Developer View

27. 57.3% of respondents have agents running in production, up from 51% the year before (LangChain, State of Agent Engineering 2025, December 2025, n=1,340, fielded 18 November to 2 December 2025).

28. 67% of organizations with more than 10,000 employees have agents in production, against 50% of organizations under 100 employees (LangChain, December 2025).

29. Customer service is the most common agent use case at 26.5%, followed by research and data analysis at 24.4% and internal workflow automation at 18% (LangChain, December 2025).

30. 52.4% of organizations run offline evaluations against test sets, 37.3% run online evaluations, and 59.8% still rely on human review (LangChain, December 2025).

LangChain surveys its own community, made up of developers already building agentic applications. Treat 57.3% as the ceiling for that population, not a figure for the economy. LangChain’s working definition of an agent, any LLM that affects control flow, is also broad enough to cover a lot of things nobody would call an agent in a budget meeting.

What The Telemetry Shows

31. Directive, automation-style task delegation on Claude.ai rose from 27% to 39% of conversations across three sampling waves (Anthropic Economic Index, September 2025, drawn from around 1 million Claude.ai conversations and 1 million API transcripts).

32. Augmentation has overtaken automation as the most common Claude.ai pattern, at 52% augmentation against 45% automation (Anthropic, March 2026 report).

33. On Anthropic’s first-party API, the top 10 O*NET tasks rose from 28% to 33% of traffic between August 2025 and February 2026 (Anthropic, March 2026).

Anthropic’s index is product telemetry from one vendor’s users, so it describes how people use Claude rather than how the workforce uses AI. The augmentation and automation shares are published as a chart rather than as a number in the report text, which is worth knowing before you quote them. It earns a place because the automation-versus-augmentation split is measured behavior rather than a survey answer, and that’s rare in this category.

What The Productivity Research Actually Measures

Most “hours saved” figures in this category are self-reported and definitionally vague. A handful of studies aren’t.

The Peer-Reviewed Causal Estimates

34. 15% average increase in issues resolved per hour among 5,172 customer-support agents given access to a generative AI conversational assistant (Brynjolfsson, Li and Raymond, Quarterly Journal of Economics, 2025, vol. 140(2), pp. 889 to 942).

35. Less experienced and lower-skilled workers improved both the speed and the quality of their output, while the most experienced and highest-skilled workers saw small gains in speed and small declines in quality (Brynjolfsson, Li and Raymond, 2025).

36. The earlier NBER working-paper version of the same study, with 5,179 agents, reported a 14% average gain and a 34% improvement for novice and low-skilled workers (NBER WP 31161, April 2023, revised November 2023).

37. Gains were largest for moderately rare problems, where human agents have less baseline experience but the system still has adequate training data (Brynjolfsson, Li and Raymond, 2025).

38. AI assistance improved the experience of work on several measures, including customers being more polite and less likely to ask to speak to a manager (Brynjolfsson, Li and Raymond, 2025).

Almost every roundup quotes “14% on average, 34% for novices” from this study. Those are the working-paper figures from 2023. The peer-reviewed version published in 2025 reports 15% on a slightly different sample and replaces the single novice number with a speed-versus-quality split. If you’re going to call a study peer-reviewed, quote the peer-reviewed numbers.

39. 26.08% increase in completed tasks among developers given an AI coding assistant, pooled across three randomized controlled trials and 4,867 developers at Microsoft, Accenture and an anonymous Fortune 100 company, with a standard error of 10.3% (Cui et al., Management Science, 2025).

40. Less experienced developers had both higher adoption rates and greater productivity gains in those trials (Cui et al., 2025).

The Vendor-Funded Lab Number

41. 55.8% faster task completion for developers using GitHub Copilot, with p=0.0017 and a 95% confidence interval of 21% to 89% (Peng et al., February 2023, n=95 professional programmers).

42. The task was implementing an HTTP server in JavaScript from scratch in a lab setting, and the authors were at GitHub, Microsoft and MIT (Peng et al., February 2023).

43. Effects skewed toward less experienced developers, the same direction as the support-agent study and the three-company field trials (Peng et al., February 2023).

The 55.8% figure is three years old, vendor-funded, based on 95 people, and measures a greenfield task nobody actually does at work. It’s a ceiling for novice work on a clean slate. Line it up against the field trials and the pattern is hard to miss: the larger and more realistic the study, the smaller the measured gain.

Where The Gains Stop

44. Autocomplete, interactive coding agents and autonomous coding agents raise coding activity, measured as commits, by cumulative effects of 30%, 180% and 240% respectively (Demirer, Musolff and Yang, NBER WP 35275, issued May 2026 and revised September 2026, matched event study on more than 500,000 GitHub developers with AI usage telemetry).

45. That 240% cumulative effect falls to 80% for the number of projects and to 30% for actual releases (Demirer, Musolff and Yang, 2026).

46. The estimated elasticity of substitution between AI and human effort is 0.23, which indicates strong complementarity rather than replacement (Demirer, Musolff and Yang, 2026).

47. Across four major software marketplaces the same study found a sharp increase in new apps but no increase in total usage (Demirer, Musolff and Yang, 2026).

This is the closest thing the category has to a measurement of where the gains actually go, and almost nobody quotes it. Task-level gains are enormous. They attenuate at every step toward shipped, used output, because the human review steps around the automated ones didn’t get faster. That’s the same shape as McKinsey’s 80% individual productivity next to 37% EBIT impact, measured a completely different way.

The Self-Reported Numbers

48. A majority of marketing leaders say AI tools save them one to two hours per workday (HubSpot, January 2025).

49. 79% of marketers agree that AI and automation tools can help them spend less time on manual tasks (HubSpot, January 2025).

Self-reported time savings are the weakest evidence on this page and the most widely quoted. Nobody measured a clock. People were asked how much time they thought they saved. I keep them here because they’re what marketing teams are actually deciding on, not because they measure anything.

Spend, Returns, And The Redesign Variable

Buying agents doesn’t move EBIT. Changing how the work is structured appears to.

What Redesign Is Worth

50. AI high performers are roughly three times more likely than their peers to have fundamentally redesigned individual workflows (McKinsey, November 2025, Exhibit 11).

51. Workflow redesign has one of the strongest contributions to enterprise-level impact in McKinsey’s relative-weights model across 31 organizational variables (McKinsey, November 2025).

What The Spending Buys

52. 79% of executives acknowledge struggling with issues including lagging ROI, strategy gaps and internal power struggles, a double-digit increase over the prior year’s wave (WRITER, April 2026).

53. 92% of the C-suite admit they are actively cultivating a new class of “AI elite” employees (WRITER, 7 April 2026).

54. 75% say their company’s AI strategy is “more for show” than real guidance (WRITER, 7 April 2026).

If you track one lever from this entire page, it’s that 3x redesign figure. High performers aren’t simply buying more AI. They’re restructuring the work around it, and the survey model puts the impact there. The 82% of marketers who name repetitive work as the thing they most want reduced are describing a workflow problem before a tooling problem. I’ve written up what that looks like for a content team in content marketing automation and where the review steps belong in human in the loop content marketing.

Where AI Workflow Automation Breaks

The most trustworthy statistics in this category are the ones about what isn’t working, because nobody has a commercial reason to inflate them.

Negative Consequences And Job Fear

55. 51% of organizations report at least one negative consequence from AI use (McKinsey, November 2025).

56. 30% specifically cite inaccuracy as a negative outcome (McKinsey, November 2025).

57. 39% of respondents expect their employer to cut jobs because of AI in the coming year, up from 32% in the previous wave, while 43% still expect little or no AI-related change in total employment (McKinsey, August 2026).

58. 53% of marketers believe AI will eliminate more marketing jobs than it creates, the highest reading since the question was first asked and up 13 points from 40% in 2023 (Marketing AI Institute, May 2025).

59. 38% of CEOs report a high or crippling amount of stress around AI strategy, and 64% fear losing their job if they fail to lead their organization through the AI transition (WRITER, April 2026).

60. 60% of companies plan to lay off employees who will not adopt AI (WRITER, April 2026).

Inaccuracy at 30% is the number worth sitting with if you automate anything that publishes. The AI hallucination statistics page has the model-level error rates behind it.

The Klarna Walk-Back

61. Klarna’s OpenAI-powered assistant handled 2.3 million conversations in its first month, about two-thirds of the company’s customer-service chats, across 23 markets and more than 35 languages (Klarna, February 2024).

62. Klarna described that volume as the equivalent work of 700 full-time agents and projected a $40 million profit improvement for 2024 (Klarna, February 2024).

63. Resolution time fell from 11 minutes to under 2 minutes, with a 25% drop in repeat inquiries (Klarna, February 2024).

64. By May 2025 Klarna confirmed it was re-hiring human customer-service staff, with the chatbot still handling two-thirds of inquiries and response times improved 82% (reported by CX Dive from Klarna’s own statements and a Bloomberg interview with its CEO, May 2025; Klarna published no release of its own on the reversal).

The 700-agents figure matched the number of staff Klarna had laid off in 2022, which several journalists noted at the time. Klarna didn’t switch the bot off. It stopped claiming humans were optional. The $40 million was a forward-looking projection and no audited result has ever been published against it.

What Gartner Expects To Fail

65. More than 40% of agentic AI projects will be canceled by the end of 2027, on escalating costs, unclear value and inadequate risk controls (Gartner, 25 June 2025).

66. Of the thousands of vendors describing themselves as agentic AI companies, Gartner estimates roughly 130 are real (Gartner, 25 June 2025).

67. In a January 2025 Gartner webinar poll of 3,412 attendees, 19% of organizations had made significant agentic AI investments, 42% conservative investments and 8% none, with the remaining 31% taking a wait-and-see approach or unsure (Gartner, 25 June 2025).

That poll is a self-selected sample of people who registered for a webinar about agentic AI, so it overweights the enthusiastic by construction. The cancellation figure is analyst judgment, not a measurement, and it should be quoted that way every single time.

What The Automation Vendors Actually Disclose

Vendor growth figures are where this category goes wrong most often, because third-party ARR estimates get repeated until they read like filings.

The Numbers That Are Filed

68. Salesforce reported $800 million in Agentforce annual recurring revenue in Q4 FY26, up 169% year over year (Salesforce, February 2026).

69. Salesforce closed 29,000 Agentforce deals in the quarter, up 50% quarter over quarter (Salesforce, February 2026).

70. Data 360 ingested 112 trillion records across FY26, up 114% year over year, and Salesforce processed more than 19 trillion tokens, converting them into more than 2.4 billion agentic work units (Salesforce, February 2026).

71. A year earlier, Salesforce had closed 5,000 Agentforce deals since October 2024, of which more than 3,000 were paid, on $900 million of Data Cloud and AI annual recurring revenue, up 120% year over year (Salesforce, February 2025).

72. On help.salesforce.com, Agentforce handled 380,000 support conversations with an 84% resolution rate (Salesforce, February 2025).

Salesforce counts free pilots inside “Agentforce deals” and says so itself, which is exactly why the filing is worth more than an analyst’s estimate of the same thing.

The Private Company Milestones

73. Lovable passed $100 million ARR in eight months (Lovable, July 2025), and raised a $200 million Series A led by Accel at a $1.8 billion valuation the same month (Lovable, July 2025).

74. Replit crossed $100 million ARR by June 2025, up from about $10 million at the end of 2024, per CEO Amjad Masad (Replit, September 2025).

75. Vercel raised a $300 million Series F at a $9.3 billion post-money valuation, co-led by Accel and GIC (Vercel, 30 September 2025), having doubled its user base over the prior year with 82% top-line growth year over year (GIC, September 2025).

76. More than 4 million people have used v0 to turn ideas into apps, up from 3.5 million at the Series F (Vercel, February 2026).

77. n8n closed a €55 million Series B led by Highland Europe with more than 200,000 users worldwide (n8n, March 2025); its lead investor put the count at 230,000 active users, ARR grown 5x, and LLM integrations at more than 75% of workflows (Highland Europe, March 2025).

78. Zapier’s platform processes more than 2 billion AI tasks per month across 3.7 million total companies (Zapier, late 2025).

79. At Zapier Central’s launch, more than 50 million tasks per month were being delegated to AI products and 388,000 customers were using Zapier’s AI features (Zapier press release, March 2024).

80. Workato’s last publicly disclosed valuation is the $200 million Series E it raised at $5.7 billion (Workato, November 2021, archived).

Every ARR figure you see for Zapier, Workato or Bolt in a statistics roundup traces back to Sacra, Latka or a similar estimator. Those are informed guesses. They’re not disclosures, and roundups presenting them as disclosures are why this category has a credibility problem. Zapier hasn’t confirmed an ARR figure since a roughly $35 million disclosure in 2018, and Workato has published nothing since 2021.

If you’re choosing between these platforms rather than counting them, the reviews of n8n and Zapier go into pricing and where each one breaks.

Frequently Asked Questions

How Many Companies Actually Use AI In Their Workflows?

88% of organizations regularly use AI in at least one business function, up from 78% a year earlier, per McKinsey’s November 2025 State of AI survey of 1,993 respondents across 105 nations. Its August 2026 follow-up, with 1,719 respondents, found 37% attributing any EBIT impact to that use and 6% qualifying as high performers, both essentially unchanged on the previous wave. Microsoft and LinkedIn’s 2024 Work Trend Index, with 31,000 respondents, found 75% of knowledge workers using AI at work. All of these count “uses AI somewhere” rather than “runs automated workflows.”

What Is The Realistic Productivity Gain From AI Workflow Automation?

15% on average for support work and around 26% for software development, from the two most credible studies available. Brynjolfsson, Li and Raymond measured 5,172 customer-support agents and published 15% in the Quarterly Journal of Economics in 2025. Cui and colleagues pooled three randomized trials across 4,867 developers at Microsoft, Accenture and a Fortune 100 company and found a 26.08% increase in completed tasks. The 55.8% figure everybody quotes comes from a 95-person vendor-funded lab trial on a greenfield coding task.

Are AI Agents Really In Production At Most Companies?

Depends entirely who you ask, and the spread is enormous. LangChain’s December 2025 survey of its own developer community says 57.3%. McKinsey’s August 2026 cross-industry sample says 40% of respondents at organizations above $1 billion in revenue are scaling agents, up from 27%. WRITER’s April 2026 survey found 97% of executives claiming deployment against 52% of employees reporting use. A minority of companies run agents in production and almost every company says it has.

Why Doesn’t AI Productivity Show Up In Company Results?

Because the gain attenuates at every handoff. The clearest measurement of this is an NBER study of more than 500,000 GitHub developers: autonomous coding agents raise commits by 240%, but that falls to 80% for projects started and 30% for actual releases, with no increase in total app usage across four marketplaces. The human review steps around the automated ones didn’t get faster. McKinsey’s survey shows the same shape from a different angle, with 80% of AI users reporting personal productivity gains against 37% of respondents reporting any EBIT impact.

How Much Are Companies Spending On AI Automation?

WRITER’s April 2026 survey of 2,400 respondents, split evenly between C-suite and employees, found 59% of companies investing more than $1 million a year in AI, while 29% report significant ROI from generative AI. McKinsey found 20% of respondents saying AI operating costs have already constrained their use of the technology. On the vendor side, Salesforce booked $800 million in Agentforce ARR in Q4 FY26.

Why Does Everyone Quote Klarna’s AI Story?

Klarna announced in February 2024 that its OpenAI-powered assistant was doing the equivalent work of 700 full-time agents, a number that matched the staff it had laid off in 2022. By May 2025 the company confirmed it was re-hiring humans for customer service. The bot still handles two-thirds of inquiries and response times improved 82%, so this was a correction rather than a collapse. If you cite the 700 figure, cite the walk-back with it.

What Share Of Agentic AI Projects Will Fail?

Gartner forecasts that more than 40% of agentic AI projects will be canceled by the end of 2027, citing escalating costs, unclear value and inadequate risk controls. The firm also estimates that only around 130 of the thousands of self-described agentic AI vendors are real. Both are analyst forecasts published in June 2025, not measurements, and quoting them as data is a mistake.

Why Isn’t The Workflow Automation Market Size In This Article?

Because no primary source for it’s publicly readable. The market-size figures circulating for workflow automation trace back to paywalled Gartner and IDC forecasts, or to vendor press releases that cite no source at all. A number that can’t be checked has no place on a page whose entire claim is that every number was checked. The Sources and Methodology section lists the rest of the exclusions.

Sources And Methodology

Every figure on this page was traced to the organization that produced it and checked against the primary document in September 2026: survey PDFs and press releases, arXiv preprints, peer-reviewed journal articles, SEC-filed earnings releases, and company funding announcements. Aggregator blogs and news write-ups weren’t accepted as sources, except where the primary record is itself a journalist interview, which is stated on the line.

A statistic was included only when three conditions were met. The originating organization is named and dated. Sample size or methodology is disclosed. The figure survived a check against the primary source rather than a citation chain.

Where a study has been revised or superseded, both readings appear. The Brynjolfsson, Li and Raymond working paper’s 14% average and 34% novice gain became 15% on a slightly different sample in the peer-reviewed version, and both are stated above. McKinsey’s November 2025 wave and its August 2026 successor are each labeled by date, because several widely quoted breakdowns exist only in the earlier wave.

Vendor-published research is labeled inline. McKinsey, Deloitte, Microsoft, HubSpot, WRITER, LangChain and the Marketing AI Institute all sell into the market they’re measuring, and several of their samples are drawn from audiences already committed to AI. Those figures appear because they’re the largest cross-industry trackers with disclosed methodology, and they should be read as directional rather than as population estimates.

Analyst forecasts are labeled as forecasts. Gartner’s predictions for 2027 are analyst judgment supported by client surveys and proprietary models, and the underlying webinar poll is a self-selected convenience sample of attendees.

The following were excluded. Workflow automation and iPaaS market-size figures, because the underlying Gartner and IDC reports are paywalled and the widely repeated numbers carry no primary citation. ARR estimates for Zapier, Workato, n8n and Bolt from third-party trackers such as Sacra and Latka, because they’re estimates presented as disclosures. Generic “hours saved per week” figures, because no cross-vendor study defines the unit consistently. GitHub star counts and self-hosted instance counts, because they’re unreviewed scrapes rather than company disclosures. A widely repeated Barclays estimate of a 40% traffic decline at one vendor, because no quotable primary version of that note could be located. Klarna’s $40 million profit improvement is presented as the forward-looking projection it was, since no audited result has been published against it.

Related reading on this site: content marketing automation, human in the loop content marketing, AI hallucination statistics and AI content detection statistics.