Side-by-Side Comparison
You are in Sample Mode. Try Live Mode
1{
2 "output": [
3 [
4 {
5 "page": 1,
6 "lines": [
7 "QuantumBlack AI by McKinsey.",
8 "The state of AI in early 2024: Gen AI adoption spikes and starts to generate value",
9 "May 2024"
10 ]
11 },
12 {
13 "page": 2,
14 "lines": [
15 "As generative Al adoption accelerates, survey respondents report measurable benefits and increased mitigation of the risk of inaccuracy.",
16 "A small group of high performers lead the way. This article is a collaborative effort by Alex Singla, Alexander Sukharevsky, Lareina Yee, and Michael Chui, with Bryce Hall, representing views from QuantumBlack, Al by McKinsey and McKinsey Digital.",
17 "If 2023 was the year the world discovered generative Al (gen AI), 2024 is the year organizations truly began using-and deriving business value from—this new technology.",
18 "In the latest McKinsey Global Survey on Al, 65 percent of respondents report that their organizations are regularly using gen Al, nearly double the percentage from our previous survey just ten months ago.",
19 "Respondents' expectations for gen Al's impact remain as high as",
20 "they were last year, with three-quarters predicting that gen Al will",
21 "lead to significant or disruptive change in their industries in the",
22 "years ahead.",
23 "Organizations are already seeing material benefits from gen Al",
24 "use, reporting both cost decreases and revenue jumps in the",
25 "business units deploying the technology. The survey also provides",
26 "insights into the kinds of risks presented by gen Al-most notably,",
27 "inaccuracy-as well as the emerging practices of top performers to",
28 "mitigate those challenges and capture value.",
29 "The state of Al in early 2024: Gen Al adoption spikes and starts to generate value",
30 "1"
31 ]
32 },
33 {
34 "page": 3,
35 "lines": [
36 "AI adoption surges",
37 "Interest in generative Al has also brightened the spotlight on a broader set of Al capabilities. For the past six years, Al adoption by respondents' organizations has hovered at about 50 percent.",
38 "This year, the survey finds that adoption has jumped to 72 percent (Exhibit 1). And the interest is truly global in scope. Our 2023 survey found that Al adoption did not reach 66 percent in any",
39 "region; however, this year more than two-thirds of respondents in nearly every region say their organizations are using Al.¹ Looking by industry, the biggest increase in adoption can be found in professional services.²",
40 "Exhibit 1",
41 "Al adoption worldwide has increased dramatically in the past year, after years of little meaningful change.",
42 "Organizations that have adopted Al in at least 1 business function,¹ % of respondents",
43 "100",
44 "80",
45 "100",
46 "Adoption of Al",
47 "80",
48 "72",
49 "58",
50 "60",
51 "56",
52 "60",
53 "65",
54 "50",
55 "50",
56 "47",
57 "55",
58 "40",
59 "40",
60 "20",
61 "20",
62 "0",
63 "T",
64 "2017",
65 "2018",
66 "2019",
67 "2020",
68 "2021",
69 "Use of generative Al",
70 "33",
71 "20",
72 "2022",
73 "2023",
74 "0",
75 "2024",
76 "'In 2017, the definition for Al adoption was using Al in a core part of the organization's business or at scale. In 2018 and 2019, the definition was embedding at",
77 "least 1 Al capability in business processes or products. Since 2020, the definition has been that the organization has adopted Al in at least 1 function.",
78 "Source: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024",
79 "McKinsey & Company",
80 "1 Organizations based in Central and South America are the exception, with 58 percent of respondents working for organizations",
81 "based in Central and South America reporting Al adoption.",
82 "2Includes respondents working for organizations focused on human resources, legal services, management consulting, market",
83 "research, R&D, tax preparation, and training.",
84 "The state of Al in early 2024: Gen Al adoption spikes and starts to generate value",
85 "2"
86 ]
87 }
88 ],
89 [
90 {
91 "page": 1,
92 "content": "Also, responses suggest that companies are now using Al in more parts of the business. Half of\nrespondents say their organizations have adopted Al in two or more business functions, up from\nless than a third of respondents in 2023 (Exhibit 2).\n\nExhibit 2\nSurvey findings suggest that organizations are using Al in more business\nfunctions now than in previous years.\n\nBusiness functions at respondents' organizations that have adopted Al,1% of respondents\n\n1 or more functions\n2 or more functions\n3 or more functions\n4 or more functions\n72\n5 or more functions\n100\n90\n80\n70\n60\n50\n50\n40\n30\n20\n10\n27\n15\n8\n0\n2021\n2022\n2023\n2024\n'In 2021, n = 1,843; in 2022, n = 1,492; in 2023, n = 1,684; in early 2024, n = 1,363.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\n\nMcKinsey & Company\n\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n3"
93 },
94 {
95 "page": 2,
96 "content": "Gen Al adoption is most common in the functions where it can create the most value\nMost respondents now report that their organizations-and they as individuals are using gen Al.\nSixty-five percent of respondents say their organizations are regularly using gen Al in at least one\nbusiness function, up from one-third last year. The average organization using gen Al is doing so\nin two functions, most often in marketing and sales and in product and service development-\ntwo functions in which previous research determined that gen Al adoption could generate the\nmost value3-as well as in IT (Exhibit 3). The biggest increase from 2023 is found in marketing\nand sales, where reported adoption has more than doubled. Yet across functions, only two use\ncases, both within marketing and sales, are reported by 15 percent or more of respondents.\n\n3\"The economic potential of generative Al: The next productivity frontier,\" McKinsey, June 14, 2023.\n\nExhibit 3\nRespondents most often report generative Al adoption in their marketing-\nand-sales, product- and service-development, and IT functions.\n\nRespondents' organizations regularly using generative Al (gen Al), by function, % of respondents\n34\n23\n17\n16\n16\n13\n12\n8\n7\n6\n4\n----------\n..........\n................... ______\n------\n------\n----\nMarketing\nIT\nand sales\nService\noperations\nHuman\nresources\nStrategy and\ncorporate finance\nManufacturing\nProduct and/or\nservice development\nOther corporate\nfunctions\nSoftware\nengineering\nRisk\nSupply chain/\ninventory management\n\nMost commonly reported gen Al use cases within function, % of respondents\n\nMarketing and sales\n16\nProduct and/or service development\nIT\n10\n7\nContent support for marketing strategy\nDesign development\n15\nPersonalized marketing\n6\nIT help desk chatbot\n7\nScientific literature and research review\nData management\n8\nSales lead identification and prioritization\n6\nAccelerated early simulation/testing\n6\nIT help desk Al assistant¹\n'Eg, providing real-time assistance and script suggestions to help desk employees during human-to-human conversations.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\n\nMcKinsey & Company\n\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n4"
97 },
98 {
99 "page": 3,
100 "content": "Gen Al also is weaving its way into respondents' personal lives. Compared with 2023,\nrespondents are much more likely to be using gen Al at work and even more likely to be using\ngen Al both at work and in their personal lives (Exhibit 4). The survey finds upticks in gen Al use\nacross all regions, with the largest increases in Asia-Pacific and Greater China. Respondents at\nthe highest seniority levels, meanwhile, show larger jumps in the use of gen Al tools for work and\noutside of work compared with their midlevel-management peers. Looking at specific industries,\nrespondents working in energy and materials and in professional services report the largest\nincrease in gen Al use.\n\nExhibit 4\nRespondents are much more likely now than in 2023 to say they are using generative Al.\n\nPersonal experience with generative Al tools, by job title, and age, 2023-24,1% of respondents\n\nRegularly use\nRegularly use for work\nRegularly use\nHave tried\nNo exposure\nfor work\nand outside of work\noutside of work\nat least once\nDon't know\n2023 2024\n4\n2\n2\n23\n4\n3\n4\n5\n4\n3 3\n4\n9\n8\n8\n18\n10\n7\n18\n15\n12\n11\n17\n19\n18\n32\n35\n33\n39\n40\n31\n42\n42\n15\n28\n31\n36\n37\n35\n30\n17\n16\n16\n15\n20\n23\n24\n13 26\n16\n20\n21\n18\n16\n31\n28\n22\n34\n26\n24\n16\n14\n14\n16\n17\n18\n22\n15\n15\n15\n13\n8\n8\n10\n12\n7\n8\n6\n7\n9\n5\nOverall\nC-level\nSenior\nMidlevel\nBorn in 1964\nBorn\nBorn\naverage¹\nexecutives2\nmanagers2\nmanagers2\nor earlier³\n1965-803\n1981-963\n\nNote: Figures may not sum to 100%, because of rounding.\n¹In 2023, n = 1,684; in 2024, n = 1,363.\n2In 2023, C-suite respondents, n = 541; senior managers, n = 437; and middle managers, n = 339. In 2024, C-suite respondents, n = 474; senior managers, n = 406; and middle\nmanagers, n = 206.\n3In 2023, for respondents born in 1964 or earlier, n = 143; for respondents born between 1965 and 1980, n = 268; and for respondents born between 1981 and 1996, n = 80. In 2024,\nfor respondents born in 1964 and earlier, n = 158; for respondents born between 1965 and 1980, n = 331; and for respondents born between 1981 and 1996, n = 184. Age details were\nnot available for all respondents.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\n\nMcKinsey & Company\n\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n5"
101 }
102 ],
103 {
104 "page1": {
105 "data": [
106 {
107 "2023": "4",
108 "2024": "5",
109 "industry": "Media and telecom",
110 "Don't know": "3",
111 "No exposure": "1",
112 "Have tried at least once": "6",
113 "Regularly use outside of work": "3",
114 "Regularly use for work and outside of work": "12"
115 },
116 {
117 "2023": "2",
118 "2024": "3",
119 "industry": "Technology",
120 "Don't know": "4",
121 "No exposure": "2",
122 "Have tried at least once": "4",
123 "Regularly use outside of work": "5",
124 "Regularly use for work and outside of work": "3"
125 },
126 {
127 "2023": "3",
128 "2024": "5",
129 "industry": "Business, legal, and professional services",
130 "Don't know": "8",
131 "No exposure": "10",
132 "Have tried at least once": "11",
133 "Regularly use outside of work": "8",
134 "Regularly use for work and outside of work": "7"
135 },
136 {
137 "2023": "8",
138 "2024": "17",
139 "industry": "Energy and materials",
140 "Don't know": "26",
141 "No exposure": "26",
142 "Have tried at least once": "14",
143 "Regularly use outside of work": "13",
144 "Regularly use for work and outside of work": "21"
145 },
146 {
147 "2023": "15",
148 "2024": "24",
149 "industry": "Advanced industries",
150 "Don't know": "31",
151 "No exposure": "39",
152 "Have tried at least once": "29",
153 "Regularly use outside of work": "28",
154 "Regularly use for work and outside of work": "32"
155 },
156 {
157 "2023": "23",
158 "2024": "14",
159 "industry": "Consumer goods and retail",
160 "Don't know": "26",
161 "No exposure": "44",
162 "Have tried at least once": "41",
163 "Regularly use outside of work": "9",
164 "Regularly use for work and outside of work": "6"
165 },
166 {
167 "2023": "34",
168 "2024": "42",
169 "industry": "Financial services",
170 "Don't know": "51",
171 "No exposure": "41",
172 "Have tried at least once": "43",
173 "Regularly use outside of work": "38",
174 "Regularly use for work and outside of work": "40"
175 },
176 {
177 "2023": "44",
178 "2024": "26",
179 "industry": "Healthcare, pharma, and medical products",
180 "Don't know": "30",
181 "No exposure": "13",
182 "Have tried at least once": "39",
183 "Regularly use outside of work": "27",
184 "Regularly use for work and outside of work": "38"
185 }
186 ],
187 "note": "Note: Figures may not sum to 100%, because of rounding.\n'In 2023, media, entertainment, and telecommunications, n = 69; technology, n = 175; business, legal, and professional services, n = 215; energy and materials,\nn = 152; advanced industries (includes automotive and assembly, aerospace and defense, advanced electronics, and semiconductors), n = 112; consumer goods and retail, n =\n128; financial services, n = 248; healthcare, pharmaceuticals, and medical products, n = 130. In 2024, media, entertainment, and telecommunications, n = 70; technology, n =\n184; business, legal, and professional services, n = 166; energy and materials, n = 113; advanced industries, n = 86; consumer goods and retail, n = 100; financial services, n =\n201; healthcare, pharmaceuticals, and medical products, n = 109. Analyses for 2023 were updated to include additional industries within advanced industries and energy and\nmaterials.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024",
188 "note2": "Note: Figures may not sum to 100%, because of rounding.\n¹In 2023, Asia-Pacific, n = 164; Europe, n = 515; North America, n = 392; Greater China (includes Hong Kong and Taiwan), n = 337; and developing markets (includes India,\nLatin America, and Middle East and North Africa), n = 276. In 2024, Asia-Pacific, n = 116; Europe, n = 457; North America, n = 401; Greater China (includes Hong Kong and\nTaiwan), n = 153; and developing markets (includes India, Latin America, and Middle East and North Africa), n = 234.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024",
189 "title": "Exhibit 4 (continued)",
190 "footer": "McKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value 6",
191 "columns": [
192 "Regularly use for work",
193 "Regularly use for work and outside of work",
194 "Regularly use outside of work",
195 "Have tried at least once",
196 "No exposure",
197 "Don't know"
198 ],
199 "heading": "Respondents are much more likely now than in 2023 to say they are using generative Al.",
200 "locations": [
201 {
202 "2023": "3",
203 "2024": "1",
204 "location": "Asia-Pacific",
205 "Don't know": "14",
206 "No exposure": "19",
207 "Have tried at least once": "6",
208 "Regularly use outside of work": "1",
209 "Regularly use for work and outside of work": "3"
210 },
211 {
212 "2023": "23",
213 "2024": "30",
214 "location": "Greater China",
215 "Don't know": "34",
216 "No exposure": "45",
217 "Have tried at least once": "46",
218 "Regularly use outside of work": "36",
219 "Regularly use for work and outside of work": "34"
220 },
221 {
222 "2023": "14",
223 "2024": "30",
224 "location": "Developing markets",
225 "Don't know": "27",
226 "No exposure": "19",
227 "Have tried at least once": "13",
228 "Regularly use outside of work": "14",
229 "Regularly use for work and outside of work": "12"
230 },
231 {
232 "2023": "20",
233 "2024": "18",
234 "location": "North America",
235 "Don't know": "17",
236 "No exposure": "15",
237 "Have tried at least once": "18",
238 "Regularly use outside of work": "11",
239 "Regularly use for work and outside of work": "27"
240 },
241 {
242 "2023": "15",
243 "2024": "7",
244 "location": "Europe",
245 "Don't know": "11",
246 "No exposure": "10",
247 "Have tried at least once": "6",
248 "Regularly use outside of work": "9",
249 "Regularly use for work and outside of work": "14"
250 }
251 ],
252 "subheading": "Personal experience with generative Al tools, by industry, 2023–24,1% of respondents",
253 "subheading2": "Personal experience with generative Al tools, by location, 2023-24,1% of respondents"
254 },
255 "page2": {
256 "title": "Investments in gen Al and analytical Al are beginning to create value",
257 "footer": "The state of Al in early 2024: Gen Al adoption spikes and starts to generate value 7",
258 "content": "The latest survey also shows how different industries are budgeting for gen Al. Responses\nsuggest that, in many industries, organizations are about equally as likely to be investing more\nthan 5 percent of their digital budgets in gen Al as they are in nongenerative, analytical-Al\nsolutions (Exhibit 5). Yet in most industries, larger shares of respondents report that their\norganizations spend more than 20 percent on analytical Al than on gen Al. Looking ahead,\nmost respondents-67 percent-expect their organizations to invest more in Al over the next\nthree years.\nWhere are those investments paying off? For the first time, our latest survey explored the\nvalue created by gen Al use by business function. The function in which the largest share of\nrespondents report seeing cost decreases is human resources. Respondents most commonly\nreport meaningful revenue increases (of more than 5 percent) in supply chain and inventory\nmanagement (Exhibit 6). For analytical Al, respondents most often report seeing cost benefits\nin service operations-in line with what we found last year-as well as meaningful revenue\nincreases from Al use in marketing and sales.",
259 "highlight": "Looking ahead, most respondents—\n67 percent—expect their\norganizations to invest more in\nAI over the next three years."
260 },
261 "page3": {
262 "note": "Note: Figures may not sum to 100%, because of rounding.\n'Question was asked only of respondents who said their organizations have adopted Al in at least 1 business function. For technology, n = 128; for energy and materials, n = 63;\nfor financial services, n = 107; for media, entertainment, and telecommunications, n = 50;\nfor consumer goods and retail, n = 67; for advanced industries, n = 50; for business, legal, and professional services, n = 101; and for healthcare,\npharmaceuticals, and medical products, n = 58.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024",
263 "title": "Exhibit 5",
264 "footer": "McKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value 8",
265 "heading": "In most industries, organizations are about equally likely to invest more than 5 percent of their digital budgets in generative Al and analytical Al.",
266 "industries": [
267 {
268 ">20%": "11",
269 "6-10%": "16",
270 "≤5%": "40",
271 "11-15%": "16",
272 "16-20%": "10",
273 "industry": "Technology",
274 "Don't know": "7"
275 },
276 {
277 ">20%": "17",
278 "6-10%": "8",
279 "≤5%": "46",
280 "11-15%": "9",
281 "16-20%": "7",
282 "industry": "Energy and materials",
283 "Don't know": "13"
284 },
285 {
286 ">20%": "11",
287 "6-10%": "47",
288 "≤5%": "12",
289 "11-15%": "26",
290 "16-20%": "4",
291 "industry": "Financial services"
292 },
293 {
294 ">20%": "7",
295 "6-10%": "7",
296 "≤5%": "63",
297 "11-15%": "6",
298 "16-20%": "4",
299 "industry": "Media and telecommunications",
300 "Don't know": "16"
301 },
302 {
303 ">20%": "5",
304 "≤5%": "64",
305 "11-15%": "12",
306 "16-20%": "6",
307 "industry": "Consumer goods and retail",
308 "Don't know": "12"
309 },
310 {
311 ">20%": "5",
312 "≤5%": "70",
313 "16-20%": "18",
314 "industry": "Advanced industries",
315 "Don't know": "6"
316 },
317 {
318 ">20%": "5",
319 "6-10%": "7",
320 "≤5%": "63",
321 "11-15%": "3",
322 "16-20%": "4",
323 "industry": "Business, legal, and professional services",
324 "Don't know": "19"
325 },
326 {
327 ">20%": "3",
328 "6-10%": "5",
329 "≤5%": "61",
330 "11-15%": "2",
331 "16-20%": "8",
332 "industry": "Healthcare, pharmaceuticals, and medical products",
333 "Don't know": "21"
334 },
335 {
336 ">20%": "6",
337 "6-10%": "13",
338 "≤5%": "55",
339 "11-15%": "7",
340 "16-20%": "7",
341 "industry": "Overall",
342 "Don't know": "19"
343 }
344 ],
345 "subheading": "Share of organization's digital budget spent on generative Al,¹ % of respondents",
346 "industries2": [
347 {
348 ">20%": "18",
349 "6-10%": "23",
350 "≤5%": "28",
351 "11-15%": "13",
352 "16-20%": "5",
353 "industry": "Technology",
354 "Don't know": "13"
355 },
356 {
357 ">20%": "3",
358 "6-10%": "11",
359 "≤5%": "66",
360 "11-15%": "9",
361 "16-20%": "5",
362 "industry": "Energy and materials",
363 "Don't know": "7"
364 },
365 {
366 ">20%": "12",
367 "6-10%": "13",
368 "≤5%": "44",
369 "11-15%": "6",
370 "16-20%": "11",
371 "industry": "Financial services",
372 "Don't know": "14"
373 },
374 {
375 ">20%": "11",
376 "6-10%": "10",
377 "≤5%": "48",
378 "11-15%": "7",
379 "16-20%": "3",
380 "industry": "Media and telecommunications",
381 "Don't know": "21"
382 },
383 {
384 ">20%": "16",
385 "6-10%": "11",
386 "≤5%": "48",
387 "11-15%": "8",
388 "16-20%": "3",
389 "industry": "Consumer goods and retail",
390 "Don't know": "13"
391 },
392 {
393 ">20%": "6",
394 "≤5%": "67",
395 "11-15%": "17",
396 "16-20%": "1",
397 "industry": "Advanced industries",
398 "Don't know": "7"
399 },
400 {
401 ">20%": "11",
402 "6-10%": "3",
403 "≤5%": "47",
404 "11-15%": "3",
405 "16-20%": "2",
406 "industry": "Business, legal, and professional services",
407 "Don't know": "34"
408 },
409 {
410 ">20%": "15",
411 "6-10%": "18",
412 "≤5%": "37",
413 "11-15%": "6",
414 "16-20%": "2",
415 "industry": "Healthcare, pharmaceuticals, and medical products",
416 "Don't know": "22"
417 },
418 {
419 ">20%": "11",
420 "6-10%": "11",
421 "≤5%": "45",
422 "11-15%": "8",
423 "16-20%": "6",
424 "industry": "Overall",
425 "Don't know": "16"
426 }
427 ],
428 "subheading2": "Share of organization's digital budget spent on analytical Al technology,¹ % of respondents"
429 }
430 },
431 [
432 {
433 "Exhibit 6": "Organizations most often see meaningful cost reductions from generative Al use in HR\nand revenue increases in supply chain management."
434 },
435 {
436 "Cost decrease and revenue increase from generative Al adoption in 2023, by function,¹% of respondents": null
437 },
438 {
439 "Decrease by <10%": "Decrease by 10-19%",
440 "Decrease by ≥20%": "Increase by >10% Increase by 6-10% Increase by ≤5%"
441 },
442 {
443 "Marketing and sales": "37\n22\n11\n4\n7\n12\n34\n53"
444 },
445 {
446 "Risk, legal, and compliance": "33\n15\n3\n15\n3 13\n46\n62"
447 },
448 {
449 "Human resources": "50\n19\n16\n15\n66\n21\n33"
450 },
451 {
452 "Product or service development": "37\n23\n68\n48\n23\n35"
453 },
454 {
455 "Supply chain and inventory management": "46\n31\n11\n4\n5\n18\n30\n53"
456 },
457 {
458 "Service operations": "45\n34\n74\n3 13\n29\n45"
459 },
460 {
461 "IT": "42\n26\n9\n7\n4 10\n42\n56"
462 },
463 {
464 "Software engineering": "42\n21\n16\n5\n7\n9\n30\n46"
465 },
466 {
467 "Other corporate functions": "34 13\n12\n9\n9\n13\n10 32"
468 },
469 {
470 "Average across all functions": "39\n24\n9\n6\n5 10\n29\n44"
471 },
472 {
473 "Use of analytical Al most often yields cost reductions in service operations and revenue\nincreases in marketing and sales.": null
474 },
475 {
476 "Cost decrease and revenue increase from analytical Al adoption in 2023, by function,¹ % of respondents": null
477 },
478 {
479 "Marketing and sales": "34\n28\n42\n4\n23\n34\n71"
480 },
481 {
482 "Risk, legal, and compliance": "34\n20\n59\n7\n11\n35\n53"
483 },
484 {
485 "Human resources": "37\n21\n8 8\n58\n52\n65"
486 },
487 {
488 "Product or service development": "23\n11 4 8\n10\n16\n30\n56"
489 },
490 {
491 "Supply chain and inventory management": "43\n29\n11 3\n10\n14\n39\n63"
492 },
493 {
494 "Service operations": "49\n28\n16\n5\n12 10\n35\n57"
495 },
496 {
497 "IT": "37\n22\n11\n4\n11\n12\n27\n50"
498 },
499 {
500 "Software engineering": "41\n20\n17\n4\n11\n14\n19\n44"
501 },
502 {
503 "Other corporate functions": "25 15\n9\n8 8\n24\n40\n1"
504 },
505 {
506 "Average across all functions": "35\n23\n75\n6\n17\n35\n58"
507 },
508 {
509 "Questions were asked only of respondents who said their organizations have adopted Al in a given function. Respondents who said \"cost increase,\" \"no change,\"\n\"not applicable,\" or \"don't know\" for the effects of analytical Al on costs are not shown, and respondents who said \"revenue decrease,\" \"no change,\" \"not applicable,\" or \"don't\nknow\" for the effects of analytical Al on revenues are not shown. Data for manufacturing and strategy and corporate finance are not shown, because the base sizes were too\nsmall to meet the reporting threshold.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-March 5, 2024": null
510 },
511 {
512 "McKinsey & Company": "The state of Al in early 2024: Gen Al adoption spikes and starts to generate value 9"
513 },
514 {
515 "McKinsey commentary": "Alex Singla"
516 },
517 {
518 "Senior partner and global coleader of QuantumBlack, Al by McKinsey": null
519 },
520 {
521 "In 2024, generative Al (gen Al) is no longer a novelty. Nearly two-thirds of respondents to\nour survey report that their organizations are regularly using gen Al, nearly double what our\nprevious survey found just ten months ago, and four in ten are using gen Al in more than\ntwo business functions. The technology's potential is no longer in question. And while most\norganizations are still in the early stages of their journeys with gen Al, we are beginning to get\na picture of what works and what doesn't in implementing-and generating actual value with-\nthe technology.": null
522 },
523 {
524 "One thing we've learned: the business goal must be paramount. In our work with clients, we\nask them to identify their most promising business opportunities and strategies and then\nwork backward to potential gen Al applications. Leaders must avoid the trap of pursuing tech\nfor tech's sake. The greatest rewards also will go to those who are not afraid to think big. As\nwe've observed, the leading companies are the ones that are focusing on reimagining entire\nworkflows with gen Al and analytical Al rather than simply seeking to embed these tools into\ntheir current ways of working.": null
525 },
526 {
527 "For that to be effective, leaders must be ready to manage change at every step along the\nway. And they should expect that change to be constant: enterprises will need to design a\ngen Al stack that is robust, cost-efficient, and scalable for years to come. They'll also need\nto draw on leaders from throughout the organization. Realizing profit-and-loss impact from\ngen Al requires close partnership with HR, finance, legal, and risk to constantly readjust the\nresourcing strategies and productivity expectations.": null
528 },
529 {
530 "Inaccuracy: The most recognized and experienced risk of gen AI use": null
531 },
532 {
533 "As businesses begin to see the benefits of gen Al, they're also recognizing the diverse risks\nassociated with the technology. These can range from data management risks such as data\nprivacy, bias, or intellectual property (IP) infringement to model management risks, which tend\nto focus on inaccurate output or lack of explainability. A third big risk category is security and\nincorrect use. Respondents to the latest survey are more likely than they were last year to say\ntheir organizations consider inaccuracy and IP infringement to be relevant to their use of gen Al,\nand about half continue to view cybersecurity as a risk (Exhibit 7).": null
534 },
535 {
536 "Conversely, respondents are less likely than they were last year to say their organizations\nconsider workforce and labor displacement to be relevant risks and are not increasing efforts\nto mitigate them. In fact, inaccuracy-which can affect use cases across the gen Al value chain,\nranging from customer journeys and summarization to coding and creative content-is the only\nrisk that respondents are significantly more likely than last year to say their organizations are\nactively working to mitigate.": null
537 },
538 {
539 "The state of Al in early 2024: Gen Al adoption spikes and starts to generate value 10": null
540 },
541 {
542 "Exhibit 7": "Inaccuracy and intellectual property infringement are increasingly considered relevant\nrisks to organizations' generative Al use."
543 },
544 {
545 "Gen Al risks that organizations consider relevant,¹ % of respondents": "2023\n2024"
546 },
547 {
548 "63\n56": "53\n52\n51\n46\n45\n43\n42\n39\n40\n39\n34\n31\n30\n27\n29\n24"
549 },
550 {
551 "Inaccuracy": "Cybersecurity\nRegulatory\ncompliance\nEquity and\nfairness\nOrganizational\nreputation\nEnvironmental\nimpact\nPhysical\nsafety\nNational\nsecurity\nPolitical\nstability\nNone\nof the\nabove"
552 },
553 {
554 "Intellectual\nproperty\ninfringement": "Personal/\nindividual\nprivacy\nExplainability\nWorkforce\nlabor\ndisplacement"
555 },
556 {
557 "Gen Al risks that organizations are working to mitigate,1% of respondents": null
558 },
559 {
560 "38\n38": "33\n32,\n28\n25 25\n23\n24\n20\n18\n17\n16\n16\n13\n12\n12\n11\n9\n8\n55\n6\n4 4\n3\n3\n2"
561 },
562 {
563 "Inaccuracy": "Cybersecurity\nRegulatory\ncompliance\nEquity and\nfairness\nOrganizational\nreputation\nEnvironmental\nimpact\nPhysical\nsafety\nNational\nsecurity\nPolitical\nstability\nNone\nof the\nabove"
564 },
565 {
566 "Intellectual\nproperty\ninfringement": "Personal/\nindividual\nprivacy\nExplainability\nWorkforce\nlabor\ndisplacement"
567 },
568 {
569 "Question was asked only of respondents whose organizations have adopted Al in at least 1 function. Respondents who said \"don't know/not applicable\" are\nnot shown. In 2023, n = 913; in 2024, n = 1,052.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024": null
570 },
571 {
572 "McKinsey & Company": "The state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n11"
573 }
574 ],
575 {
576 "page1": {
577 "text": "In fact, some organizations have already experienced negative consequences from the use of\ngen Al, with 44 percent of respondents saying their organizations have experienced at least one\nconsequence (Exhibit 8). Respondents most often report inaccuracy as a risk that has affected\ntheir organizations, followed by cybersecurity and explainability.\nExhibit 8\nNearly one-quarter of respondents say their organizations have\nexperienced negative consequences from generative Al's inaccuracy.\nGenerative-Al-related risks that caused negative consequences for organizations,¹% of respondents\n39\n23\n16\n12\n11\n10\n9\n8\n7\n7\n4\n4\n4\n4\nInaccuracy\nCybersecurity\nExplainability Regulatory Organizational\nEquity and\nNational\nEnvironmental\ncompliance\nreputation\nfairness\nsecurity\nimpact\nIntellectual\nPersonal/\nWorkforce\nproperty\nindividual\nlabor\nPhysical\nPolitical\nNone\nsafety\nstability\nof the\ninfringement\nprivacy\ndisplacement\nabove\n'Question was asked only of respondents whose organizations have adopted generative Al in at least 1 function, n = 876. The 17 percent of respondents who\nsaid \"don't know/not applicable\" are not shown.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\nMcKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n12"
578 },
579 "page2": {
580 "text": "Our previous research has found that there are several elements of governance that can help in\nscaling gen Al use responsibly, yet few respondents report having these risk-related practices\nin place. For example, just 18 percent say their organizations have an enterprise-wide council or\nboard with the authority to make decisions involving responsible Al governance, and only\none-third say gen Al risk awareness and risk mitigation controls are required skill sets for\ntechnical talent.\n4\"Implementing generative Al with speed and safety,\" McKinsey Quarterly, March 13, 2024.\nMcKinsey commentary\nLareina Yee\nSenior partner, McKinsey; chair, McKinsey Technology Council\nResponsible Al needs to start on day one, and there is still much work to be done in terms\nof education and action. It begins with a company's values-organizations must establish\nclear principles for how they apply generative Al (gen Al) and set up guardrails to ensure its\nsafe implementation. For example, recognizing the importance of data security means that\ncompany-level data and prompts remain within the enterprise walls. For that to happen, the\nenterprise must have secure contracts with large language model and application providers,\nas well as robust training, to make sure employees understand the difference between\nenterprise tools and public tools so that code or proprietary data are not inadvertently shared\nin public models.\nResponsible Al also starts upstream of compliance and monitoring. Leading companies in\ndeploying gen Al incorporate risk practices in the development of their Al applications. This\nincludes ensuring that technical teams understand risk and mitigation practices. Gen Al\nsolutions are probabilistic models that can make mistakes or inadvertently amplify biases\nin training data, so testing models before they are deployed is essential. Without a robust\ntesting approach, it is hard to deliver on responsible Al.\nFinally, companies must develop a clear governance model to help ensure that gen Al\napplications conform to governing principles. What we see in the survey results and in\nour conversations with clients is a growing awareness of responsible Al and an urgency to\nget it right. Still, even with increasing understanding, a little less than one-quarter of the\nrespondents in our survey report having a clear process to embed risk mitigation in their\nsolutions. Moving from awareness to action will be critical.\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n13"
581 },
582 "page3": {
583 "text": "Bringing gen AI capabilities to bear\nThe latest survey also sought to understand how, and how quickly, organizations are deploying\nthese new gen Al tools. We have found three archetypes for implementing gen Al solutions:\ntakers use off-the-shelf, publicly available solutions; shapers customize those tools with\nproprietary data and systems; and makers develop their own foundation models from scratch.5\nAcross most industries, the survey results suggest that organizations are finding off-the-shelf\nofferings applicable to their business needs-though many are pursuing opportunities to\ncustomize models or even develop their own (Exhibit 9). About half of reported gen Al uses within\nrespondents' business functions are utilizing off-the-shelf, publicly available models or tools,\nwith little or no customization. Respondents in energy and materials, technology, and media\nand telecommunications are more likely to report significant customization or tuning of publicly\navailable models or developing their own proprietary models to address specific business needs.\n5 \"Technology's generational moment with generative AI: A CIO and CTO guide,\" McKinsey, July 11, 2023.\nExhibit 9\nOrganizations are pursuing a mix of off-the-shelf generative Al capabilities\nand also significantly customizing models or developing their own.\nStrategy for developing generative Al (gen Al) capabilities, % of reported instances of gen Al use¹\nEnergy and materials\n60\nTechnology\n56\n40\nSignificant\ncustomization\nor developed\nown model\n44\nPrimarily off\nMedia and telecommunications\n54\n46\nthe shelf, with\nlittle or no\ncustomization\nConsumer goods and retail\n50\n50\nFinancial services\n47\n53\nHealthcare, pharmaceuticals,\n47\n53\nand medical products\nAdvanced industries\n42\n58\nBusiness, legal, and\n37\n63\nprofessional services\nOverall\n47\n53\n'Question was asked only of respondents who said their organizations regularly use generative Al in at least 1 business function. Figures were calculated after\nremoving respondents who said \"don't know.\"\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\nMcKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n14"
584 }
585 },
586 [
587 {
588 "page": 1,
589 "text": "Gen AI high performers are much\nmore likely than others to use\ngen AI solutions in risk, legal,\nand compliance; in strategy and\ncorporate finance; and in supply\nchain and inventory management.\nWhat else are these high performers doing differently? For one thing, they are paying more\nattention to gen-Al-related risks. Perhaps because they are further along on their journeys,\nthey are more likely than others to say their organizations have experienced every negative\nconsequence from gen Al we asked about, from cybersecurity and personal privacy to\nexplainability and IP infringement. Given that, they are more likely than others to report that their\norganizations consider those risks, as well as regulatory compliance, environmental impacts, and\npolitical stability, to be relevant to their gen Al use, and they say they take steps to mitigate more\nrisks than others do.\nGen Al high performers are also much more likely to say their organizations follow a set of risk-\nrelated best practices (Exhibit 11). For example, they are nearly twice as likely as others to involve\nthe legal function and embed risk reviews early on in the development of gen Al solutions-that\nis, to \"shift left.\" They're also much more likely than others to employ a wide range of other best\npractices, from strategy-related practices to those related to scaling.\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n18"
590 },
591 {
592 "page": 2,
593 "text": "Exhibit 11\nOrganizations seeing the largest returns from generative Al are more likely than others\nto follow a range of best practices.\nOrganizations engaging in each practice,¹ % of respondents\nRisk\nStrategy\nAll other respondents\nGen Al high performers2\nGen Al risk awareness and mitigation\nare required skills for technical talent\n34\n68\nHave clear processes to embed risk mitigation in\ngen Al solutions (eg, involving the legal function)\n23\n44\nGen Al models are designed to allow audits,\nbias checks, and risk assessment\n18\n43\nHave an enterprise-wide council or board to make\ndecisions on responsible Al governance\n18\n24\n|\n|\n1\n1\n0\n20\n40\n60\n1\n80\n100\nSenior leaders understand how gen Al\ncan create value for the business\n39\nHave an enterprise-wide road map for gen Al,\nprioritized based on value, feasibility, and risk\n25\nHave appointed a credible, empowered\nleader of gen Al initiatives\n21\n32\n59\n64\nTalent\n0\n20\n40\n60\n1\n80\n|\n100\nHave curated learning journeys, tailored by role,\nto build critical gen Al skills for technical talent\n18\n43\nHave clearly defined the talent (ie, both roles and skills)\nneeded to execute the gen Al strategy\n15\n32\nHave a talent strategy that allows effective recruitment,\nonboarding, and integration of gen-Al-related talent\n16\n31\n0\n20\n40\n60\n80\n100\nOperating model\nHave a centralized team that coordinates\nand links gen Al efforts across the organization\n35\n49\nDeliver gen Al solutions following well-defined\nagile team processes and standards\n19\n43\nHave funding and budgeting processes that\nsupport agile delivery of gen Al solutions\n14\n27\n1\n0\n20\n40\n60\n80\n100\n'Asked only of respondents who said their organizations are regularly using generative Al in at least 1 business function.\n2Respondents who said that at least 11% of their organizations' EBIT in 2023 was attributable to their use of gen Al. For gen Al high performers, n = 46; for all other\nrespondents, n = 830.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\nMcKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n19"
594 },
595 {
596 "page": 3,
597 "text": "Exhibit 11 (continued)\nOrganizations seeing the largest returns from generative Al are more likely than others\nto follow a range of best practices.\nOrganizations engaging in each practice,¹ % of respondents\nTechnology and data\nAll other respondents\nGen Al high performers2\nTesting and validation are embedded\nin release process for each model\n17\n58\nClear processes are in place to iteratively\nimprove model outputs\nProcesses are defined to determine when models\nneed human validation (eg, human in the loop)\nGen Al foundations are built with a\nstrategy to enable reuse across solutions\nThere is a defined, comprehensive data strategy\nto enable the gen Al road map\n15\n46\n19\n43\n15\n43\nLive monitoring of entire system is set up,\nenabling rapid issue resolution\n7\nModular components are developed that\ncan be reused across solutions\n11\nAdoption and scaling\nNontechnical personnel understand the potential value\nand risks of using gen Al in their day-to-day work\nData are used consistently to create\ninsights that affect bottom-line performance\nThere is a clear performance management infrastructure\n(eg, KPIs) to measure and track value of gen Al\n17\n42\n41\n31\n|\n1\n0\n20\n40\n60\n80\n100\n13\n21\n24\n44\n37\n52\n0\n20\n40\n60\n80\n100\n'Asked only of respondents who said their organizations are regularly using generative Al in at least 1 business function.\n2Respondents who said that at least 11% of their organizations' EBIT in 2023 was attributable to their use of gen Al. For gen Al high performers, n = 46; for all other\nrespondents, n = 830.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024\nMcKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value\n20"
598 }
599 ],
600 [
601 {
602 "page": 21,
603 "text": "In addition to experiencing the risks of gen Al adoption, high performers have encountered other\nchallenges that can serve as warnings to others (Exhibit 12). Seventy percent say they have\nexperienced difficulties with data, including defining processes for data governance, developing\nthe ability to quickly integrate data into Al models, and an insufficient amount of training data,\nhighlighting the essential role that data play in capturing value. High performers are also more\nlikely than others to report experiencing challenges with their operating models, such as\nimplementing agile ways of working and effective sprint performance management.\nExhibit 12\nGenerative Al high performers report experiencing a range of challenges in\ncapturing value from the technology.\nElements that have posed challenges in capturing value from generative Al (gen Al), % of respondents"
604 },
605 {
606 "page": 21,
607 "text": "Data\nRisk and responsible Al\nOperating model\nTechnology\nStrategy\nTalent\nAdoption and scaling\nGen Al high performers¹\n70\n48\n47\n43\n42\n37\n33\nAll other respondents\n36\n34\n28\n30\n39\n35\n38\nNote: Figures do not sum to 100%, because respondents could choose multiple answer options.\n'Respondents who said that at least 11% of their organizations' EBIT in 2023 was attributable to their use of generative Al. For respondents at Al high perform-\ners, n = 46; for all other respondents, n = 830. Respondents who said \"don't know/not applicable\" are not shown.\nSource: McKinsey Global Survey on Al, 1,363 participants at all levels of the organization, Feb 22-Mar 5, 2024"
608 },
609 {
610 "page": 21,
611 "text": "McKinsey & Company\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value 21"
612 },
613 {
614 "page": 22,
615 "text": "Find more content like this on the\nMcKinsey Insights App\nMcKinsey\n& Company\nScan Download Personalize"
616 },
617 {
618 "page": 22,
619 "text": "McKinsey commentary\nBryce Hall\nAssociate partner\nWe've been conducting research on Al for seven years now, and the pace of innovation,\nthe evolution of new companies and capabilities, and the wave of investment have been\nremarkable. And now we're seeing how leading companies are capturing business value from\nthese often-dazzling Al and generative Al (gen Al) capabilities.\nOne of the most interesting findings in this year's survey is that among the high performers\ncapturing the most value from gen Al, most solutions are highly customized or bespoke (what\nwe refer to as \"shaper\" or \"maker\" archetypes of gen Al solutions). While many companies\nare finding value from off-the-shelf gen Al solutions (or the \"taker\" archetype), capturing the\nfull value of this technology often requires significant customization-for example, training\nmodels on proprietary company and customer data or tuning models to improve performance\nwithin a specific industry or business context.\nThe survey also sheds new light on high performers' practices. High performers, for example,\nare significantly more likely than others to embed testing and validation in the release process\nfor models, as well as to develop clear processes to iteratively improve model outputs. Over\ntime, these kinds of practices will become even more important, as highly customized and\nbespoke solutions are the ones that will truly be differentiating for companies. Off-the-shelf\nsolutions, by contrast, are likely to become table stakes. Collectively, these data on practices\nare consistent with our ongoing work and research on digital and Al transformations, which\nshows that competitive advantage comes from building organizational and technological\ncapabilities to broadly innovate, deploy, and improve solutions at scale-in effect, rewiring the\nbusiness for distributed digital and Al innovation."
620 },
621 {
622 "page": 22,
623 "text": "About the research\nThe online survey was in the field from February 22 to March 5, 2024, and garnered responses\nfrom 1,363 participants representing the full range of regions, industries, company sizes,\nfunctional specialties, and tenures. Of those respondents, 981 said their organizations had\nadopted Al in at least one business function, and 878 said their organizations were regularly\nusing gen Al in at least one function. To adjust for differences in response rates, the data are\nweighted by the contribution of each respondent's nation to global GDP."
624 },
625 {
626 "page": 22,
627 "text": "Alex Singla and Alexander Sukharevsky are global coleaders of Quantum Black, Al by McKinsey, and senior\npartners in McKinsey's Chicago and London offices, respectively; Lareina Yee is a senior partner in the Bay Area\noffice, where Michael Chui, a McKinsey Global Institute partner, is a partner; and Bryce Hall is an associate partner\nin the Washington, DC, office.\nThey wish to thank Kaitlin Noe, Larry Kanter, Mallika Jhamb, and Shinjini Srivastava for their contributions to this work.\nDesigned by McKinsey Global Publishing\nCopyright © 2024 McKinsey & Company. All rights reserved.\nThe state of Al in early 2024: Gen Al adoption spikes and starts to generate value 22"
628 }
629 ]
630 ]
631}