BCG
工作场所中的
人工智能
战略比工具更重要
第四版 | 2 0 2 6年6月
波士顿咨询公司（BCG）人工智能与工作调研报告
2
调查参数
核心数据 关键市场
11,749
名受访者
澳大利亚 1,014
英国 1,013
美国 1,008
西班牙 1,005
印度 1,005
日本 1,002
巴西 1,000
德国 917
法国 913
意大利 890
北欧地区 825
南非 503
中东 355
比荷卢地区 299
角色分布 年龄分布 公司收入 行业分布
32% 一线员工
34% 管理者
34% 领导者
2% 65岁以上
12% 55-65岁
14% 45-54岁
36% 35-44岁
25% 25-34岁
10% 18-24岁
23% >100亿美元
19% 50-100亿美元
17% 20-50亿美元
14% 10-20亿美元
18% 5-10亿美元
6% 1-5亿美元
3% 1亿美元以下
21% TMT（科技、媒体、
电信）
22% 金融服务
14% 消费品/零售
10% 工业制造
6% 能源/公用事业
5% 保险
5% 医疗保健
6% 公共部门
15% 其他
来源：AI at Work, 2026 (n=11,749)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。TMT = 科技、媒体和电信。北欧地区包括瑞典、丹麦、挪威和芬兰。中东包括阿联酋、
沙特阿拉伯、科威特和卡塔尔。比荷卢地区包括比利时和荷兰。
3
新发现强化了2025年观察到的趋势
个人节省的时间并不能自动转
化为价值
42%的定期使用AI的一线员工每周
节省一整天或更多时间。但66%
的人仍然获得有限或没有关于如何
利用节省时间的指导，超过一半的
人没有将其重新投入到战略性工作
中。
端到端重新思考工作是创造价
值的前提
越来越多的组织正在使用AI进
行"创新"，构建新的商业模式。端
到端重新设计工作流程的公司在价
值捕获和员工满意度方面表现优于
仅部署工具的公司。差距不断扩
大，驱动力来自更清晰的路线图和
对人才的更深入投资。
适当的培训和领导支持仍然是
最大的未兑现承诺和释放AI潜
力的强大杠杆
72%的受访者表示对技能的期望
已经转变，但只有36%感到接受
了充分培训，与2025年持平。只
有33%的一线员工表示领导层就AI
进行了清晰沟通，28%认为领导
层所说的与组织实际所做的之间存
在高度一致性。
五大关键要点
1
不再存在"硅天花板"：一线
员工已将AI融入日常工作
74%的一线员工现在是定期AI用
户，比2025年增加了23个百分
点。印度和中东领先采用，而美
国、法国和意大利落后。
2
真正的挑战现在是组织性和管
理性的
72%的受访者报告技能期望发生
了变化，近一半表示角色已转向
管理和指导AI，而非亲自执行工
作。
3
商业价值和员工满意度不是权
衡取舍，而是由相同的力量驱
动
67%的定期AI用户更享受工作。
捕获最多商业价值的组织也是员
工最享受工作的地方。
4
AI"蜜月期"不会持续，除非领导者带来战略清晰
度
每个人都谈论节省的时间，但真正的影响在于整合。
AI的新颖性和认知拉伸在早期带来享受。但持续的满
意度来自战略清晰度。当方向真实且信息通过强有力
的CEO参与传达给员工时，员工才能蓬勃发展。
5
AI智能体已从概念走向现实，但运营模式尚未跟
上
自2025年以来，将智能体融入工作流程的组织增加
了一倍以上，61%的受访者相信智能体能在三年内完
成他们一半的工作。然而治理（监督、问责）仍然远
远落后于技术。
领导者获得最大的AI回报，部分职能比其他职能走得
4
更远
各层级/职能每周节省至少8小时的受访者比例
来源：AI at Work, 2026 (n=8,989，包括定期AI用户，排除异常值)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
差距很大：管理者和领导者感觉指导更清晰，感知到的影响比一线员工更大
5
来源：AI at Work, 2026 (n=8,989，包括定期AI用户，与节省时间分析样本一致)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
一线员工报告获得的指导最少
关于如何花费AI节省的时间，66%的一线员工获得有限或没有指导，超过一半的人没有将其重新
投入到战略性工作中。
组织如何实施AI工具
部署（Deploy）：支持GenAI工具的采用并促进生产力提升
重塑（Reshape）：端到端重新设计工作流程和流程，重新构想职能
创新（Invent）：构建和创新新的商业模式和产品以推动增长
来源：AI at Work, 2026 (n=11,749)；BCG分析。
6
追求"重塑"或"创新"举措的公司创造更多价值……其员工也更蓬勃发展
来源：AI at Work, 2026 (n=10,990，包括2,919名来自"部署"公司的受访者和8,071名来自"重塑"或"创新"公司的受访者；总计排除了不
确定其组织如何实施AI的受访者)；BCG分析。
追求"重塑"或"创新"举措的公司建立更清晰的路线图……并投资于背后的人
才
来源：AI at Work, 2026 (n=10,990，包括2,919名来自"部署"公司的受访者和8,071名来自"重塑"或"创新"公司的受访者；总计排除了不
确定其组织如何实施AI的受访者)；BCG分析。
7
大多数受访者预计未来五年需要技能提升，但只有
36%感到接受了充分培训
来源：AI at Work, 2026 (n=11,749)；BCG分析。
注：两项指标的受访者比例与2025年相比没有变化。
技能提升需求强烈且持续存在，但回应仍显不足
不再存在"硅天花板"：一线员工已将AI融入日常工作
8
各层级员工的定期AI使用情况
来源：AI at Work, 2026 (n=11,749)；BCG分析。
注：定期AI用户 = 每天或每周多次使用AI的人。一线员工 = 无管理职责的白领个人员工。
9
印度、中东和澳大利亚领先一线员工采用率，法国、意大利和美国低于平均
水平
来源：AI at Work, 2026 (n=4,040名一线员工)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。中东包括阿联酋、沙特阿拉伯、科威特和卡塔尔。比荷卢地区包括比利时和荷兰。北欧
地区包括瑞典、丹麦、挪威和芬兰。
10
支持职能在一线员工中领先采用，而销售和运营落后
来源：AI at Work, 2026 (n=4,040名一线员工)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
真正的挑战现在是组织性和管理性的
11
受访者表示AI已改变的工作方面
来源：AI at Work, 2026 (n=11,749)；BCG分析。
注：表示AI将显著或适度改变对岗位所需技能期望的受访者比例。
AI正在重塑工作，影响工作的本质和管理方式
从技能期望的转变到角色向管理AI的演变，AI正在深刻改变组织的运作方式。"足够好"的标准已
经提高，员工需要花费更多时间审查和纠正AI输出。
商业价值和员工满意度不是权衡取舍，而是由相同的
力量驱动
12
超过三分之二的定期AI用户报告工作满意度提升……
来源：AI at Work, 2026 (n=9,923名定期AI用户)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
AI的"满意度悖论"：它让工作变得更好，也更困难
按员工满意度和可衡量影响提升排名的前五大组织杠杆
来源：AI at Work, 2026 (n=9,923名定期AI用户)；BCG分析。
注：可衡量影响 = 归因于AI的关键业务指标的改善。
13
希望AI"蜜月期"持续？战略清晰度胜过工具，在驱动
持续影响方面更为重要
基于AI战略清晰度和AI工具可用性，报告可衡量影响的受访者
来源：AI at Work, 2026 (n=9,923名定期AI用户)；BCG分析。
注：强战略清晰度定义为将AI作为优先事项，拥有清晰的AI战略和关于节省时间的指导方针。强工具获取定义为可以使用AI工具或进
行AI工具测试和实验。
14
使用AI时工作满意度的驱动因素随时间演变……
……导致辛劳的因素并未随时间改变
来源：AI at Work, 2026 (n=6,998名定期使用AI的受访者，使用不到6个月 vs 超过一年)；BCG分析。
注：+/-pp = 当驱动因素存在时享受工作的员工额外比例（+），或当阻碍因素存在时失去的比例（-）。
15
AI智能体已从概念走向现实，但运营模式尚未跟上
来源：AI at Work, 2026 (n=11,749)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
缺失的一环：运营模式未能跟上AI智能体的部署步伐
更多人了解AI智能体并认识到其重要性，但很少有人了解它们是什么：这是领导者需要更好地说
明它们能做什么的信号。
16
AI智能体按类型使用情况
自2025年以来，将AI智能体融入工作流程的组织增加了一倍以上，管理者和领导者层面预期产生有意义的影响。
17
大多数人相信智能体未来能完成至少一半的工作，管理者和领导者预期最大
转变
来源：AI at Work, 2026 (n=9,923名定期AI用户)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
一半员工缺乏管理人机×AI团队的明确治理：一线员工和管理者感受最深
来源：AI at Work, 2026 (n=8,849名听说过AI智能体的受访者和定期AI用户)；BCG分析。
注：一线员工 = 无管理职责的白领个人员工。
18
问责制是一个普遍关注的问题，在各层级之间平均分布
五大CEO要务
1
将战略清晰度作为首要任务，并亲自掌控
战略清晰度不是沟通任务，而是一种领导姿态。将AI设定为明确的最高优先事项，清晰说明公
司的发展方向，并确保每个人都理解，包括一线员工。亲自掌控转型的CEO在每个维度上都表
现更优：价值捕获、员工满意度和信任。
2
改变计分板：衡量价值，而非采用率
采用率只告诉你人们在用AI，不说明是否有回报。个人节省的时间除非被追踪并有目的地重新
投资，否则会从组织中流失。因此，关注业务成果而非使用量。
3
投资于端到端重新设计工作，而非更多工具
大多数公司仍将AI视为个人生产力工具，但更重要的变化是集体性的：AI正在重塑团队协作方
式和任务在组织中的流转方式。捕获这种价值意味着端到端地重新设计一些核心流程。
4
将人置于重新设计的核心
重新设计只有在人参与的情况下才有效。前瞻性地看待未来几年的角色转变，为最重要的技能
提供培训，并让人们参与塑造变革，而非仅仅向他们展示变革。让人们保持参与的是知道AI如
何帮助他们成长，而非他们工作得多快。
5
将其视为移动目标来治理，而非一次性项目
技术发展比任何公司都快。将AI视为需要持续引导的事物，而非有终点线的项目。建立一个轻
量级的常设治理机制，重新检查什么有效，重新衡量价值，并随着模型和智能体的演变进行调
整。
FOURTH EDITION | JUNE 2026
BCG AI AT WORK
Strategy Matters
More Than Tools
Survey parameters
Sources: AI at Work, 2026 (n=11,749); BCG analysis.
Notes: Frontline employees = individual white-collar employees with no managerial responsibilities. TMT = technology, media, and telecommunications. Nordics include Sweden, Denmark, Norway, and Finland. Middle East includes
UAE, Saudi Arabia, Kuwait, and Qatar. Benelux includes Belgium and the Netherlands.
11,749
respondents
Managers
Frontline employees
Leaders
Key Markets
917
913
890
825
503
355
299
Australia
UK
US
Spain
India
Japan
Brazil
Germany
France
Italy
Nordics
South Africa
Middle East
Benelux
1,014
1,013
1,008
1,005
1,005
1,002
1,000
34%
34%
32% Roles
25–34
55–65
65+
45–54
35–44
18–24
Age
24%
36%
22%
12%
2%
4%
$500M–$1B
$5B–$10B
>$10B
$2B–$5B
$1B–$2B
$100M–$500M
Company
revenue
23%
18%
6%
14%
14%
25%
TMT
Financial services
Consumer/retail
Health care/medical
Industrial goods
Insurance
Public sector
Others
3%
Energy/utilities
15%
6%
5%
6%
10%
17%
19%
21%
Industry
New findings reinforce trends observed in 2025
The time individuals save
doesn’t automatically translate
into value
42% of frontline employees who
are regular AI users save a full day
or more per week. But 66% still
receive limited or no guidance on
what to do with time they save,
and more than half don’t redirect
it to strategic work.
Rethinking work end-to-end
is a prerequisite for creating
value
More organizations are using AI to
“invent,” building new business
models. Companies that redesign
workflows end-to-end outperform
those that only deploy tools on
value captured and employee joy.
The gap keeps widening, driven by
a clearer roadmap and deeper
investment in people.
Proper training and leadership
support remain the biggest unmet
promises and strong levers to
unlock AI’s potential
72% of respondents say expectations
about skills have shifted, yet only
36% feel they received sufficient
training, stable vs 2025. Only 33% of
frontline employees say leadership
communicates clearly about AI, and
28% see strong alignment between
what leaders say and what the
organization actually does.
The real challenge is
now organizational
and managerial
Everyone talks about
time saved, but the
real shift is deeper
and structural. 72% of
respondents report
skill expectations
have changed, and
nearly half say their
roles have shifted
toward managing and
directing AI instead of
doing the work itself.
1 2 3
No more “silicon
ceiling”: frontline
employees have
integrated AI into
their daily work
74% of frontline
employees are now
regular AI users, an
increase of 23
percentage points
from 2025. India and
the Middle East lead
adoption, while the
US, France, and Italy
trail behind.
Five key takeaways
Business value and
employee joy aren’t
tradeoffs, they’re
driven by the same
forces
67% of regular AI
users enjoy work
more. The
organizations that
capture the most
business value are
also the places where
employees enjoy
working the most.
4
The AI “honeymoon”
won’t last unless
leaders bring
strategic clarity
driving sustained
impact
AI’s novelty and
cognitive stretch fuel
enjoyment early on. But
sustained joy comes
from strategic clarity.
Employees thrive when
the direction is real and
the message reaches
them with strong CEO
involvement.
5
AI agents went from
concept to reality,
but operating models
haven’t caught up
Integration into
workflows more than
doubled since 2025,
and 61% of respondents
believe agents could do
half their job within
three years. Yet
governance (oversight,
accountability) still lags
far behind the tech.
New findings reinforce
trends observed in 2025
Sources: AI at Work, 2026 (n=8,989, includes regular AI users, excludes outliers); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities.
Among frontline
employees who
are regular AI
users, 42% report
saving at least a
workday per week
Leaders have the biggest AI payoff, with
60% saving at least a workday per week
Some job functions are further
ahead than others
52%
Overall
respondents
42%
Frontline
employees
52%Managers
60%Leaders
Finance49%
Human resources50%
Data science
and analytics48%
IT53%
Marketing60%
Top job functions where frontline employees
are saving at least 8 hours per week
Respondents saving at least 8 hours per week
Frontline employees
report getting the
least guidance on
how to spend the
time AI saves, and
more than half of
them don’t redirect
it into strategic work
The gap is wide: managers and leaders feel guidance is clearer and
perceive more impact than frontline employees
Sources: AI at Work, 2026 (n=8,989, includes regular AI users, consistent sample with time-saved analysis); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities.
Respondents who say their organizations
give limited or no guidance on what to do
with the time saved
Respondents who declare not reinvesting
time saved into more strategic work
Frontline employees LeadersManagersOverall respondents
61%
66% 66%
52%
45%
58%
43%
36%
+14pp +22pp
Organizations are
moving past
individual use case
deployments; Invent
initiatives have
nearly doubled
How organizations are implementing AI tools
Sources: AI at Work, 2026 (n=11,749); BCG analysis.
78%
57%
42%
72%
50%
22%
Deploy Reshape Invent
2026 2025
Supporting adoption of
GenAI tools and
fostering productivity
Redesigning end-to-end
workflows and processes to
reimagine functions
Building and innovating
new business models and
products to drive growth
+20pp
Using AI for
Reshape or Invent
initiatives pays off
by delivering more
value and providing
a better employee
experience
Sources: AI at Work, 2026 (n=10,990, including 2,919 respondents at Deploy companies and 8,071 respondents at Reshape or Invent companies; total excludes
respondents who were not sure how their organizations were implementing AI); BCG analysis.
Companies pursuing Reshape or
Invent initiatives deliver more value…
… and their employees thrive
Respondents at companies that
only focus on Deploy initiatives
Respondents at companies that focus
on Reshape or Invent initiatives
31%
53%
48%
68%
Save more time
Employees who
save at least a day
per week
Enjoy work more
Employees who
report increased
job satisfaction
+22pp +20pp
43%
67%
56%
79%
Prove the impact
Employees who see
measurable business
improvement
Show their value
Employees who find it
easier to demonstrate
unique value
+24pp +23pp
26%
44%
55%
66%
Earn trust
Employees who fully
trust leadership’s AI
communications
Feel more confident
Employees who feel
confident working
with AI
+18pp +11pp
What makes
Reshape or Invent
initiatives more
successful: a clearer
roadmap and
deeper investment
in people
Sources: AI at Work, 2026 (n=10,990, including 2,919 respondents at Deploy companies and 8,071 respondents at Reshape or Invent companies; total excludes
respondents who were not sure how their organizations were implementing AI); BCG analysis.
Companies pursuing Reshape or Invent
initiatives build a clearer roadmap…
… and invest in the people
behind it
Respondents at companies that
only focus on Deploy initiatives
Respondents at companies that focus
on Reshape or Invent initiatives
Sets the direction
Employees who say
AI strategy is clear
Brings people in
Employees who
participate in
process redesign
Defines the rules
Employees who see
adequate guardrails
in place
Measures what
matters
Employees who see
AI value creation
properly tracked
Puts agents to work
Employees who see
agents integrated
into workflows
Invests in skills
Employees who
experience a major
reskilling initiative
12%
43%
+31pp
31%
52%
+21pp
41%
59%
+18pp
13%
30%
+17pp
47%
65%
+18pp
27%
52%
+25pp
Demand for
upskilling is loud
and persistent.
The response still
falls short
Sources: AI at Work, 2026 (n=11,749); BCG analysis.
Note: Share of respondents for both metrics is unchanged from 2025.
Most respondents expect a need for upskilling in the next five
years, yet only 36% feel properly trained
Of respondents believe
they need major upskilling
in the next five years
Of respondents feel
properly trained
88% 36%
No more “silicon ceiling”: frontline
employees have integrated AI into
their daily work
74% of frontline
employees now
describe themselves
as regular AI users,
driving overall
adoption
Sources: AI at Work, 2026 (n=11,749); BCG analysis.
Notes: Regular AI users = people who use AI daily or several times a week. Frontline employees = individual white-collar employees with no managerial responsibilities.
Regular AI use across worker levels
Frontline
employees
Managers
Leaders
+23pp
+10pp
74%
88%
93%
51%
78%
85%
52%
64%
88%
20%
46%
80%
2025
2023
2024
2026
India, the Middle East, and Australia lead adoption for frontline employees,
while France, Italy, and US trail the average
Sources: AI at Work, 2026 (n=4,040 frontline employees); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities. Middle East includes UAE, Saudi Arabia, Kuwait, and Qatar. Benelux includes Belgium and the Netherlands. Nordics include Sweden, Denmark,
Norway, and Finland.
Frontline employees who use AI at least several times a week, by market
Brazil South
Africa
Germany UK Spain BeneluxIndia Japan US France Italy
95% 93%
86% 82% 79% 75% 75% 72% 69% 68% 66% 62% 62% 62%
Average
74%
AustraliaMiddle
East
Nordics
Sources: AI at Work, 2026 (n=4,040 frontline employees); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities.
Support functions lead adoption among frontline employees, while sales
and operations lag behind
Frontline employees who use AI at least several times a week, by function
88% 85% 83% 83%
76% 75% 73% 72% 72% 68%
61%
IT Marketing Finance Data and
analytics
Compliance
and risk
management
Human
resources
Purchasing
and
procurement
Logistics
and
planning
Administrative
and
clerical
Sales and
business
development
Operations
and
production
Average
74%
The real challenge is now
organizational and managerial
AI is reshaping
work, impacting the
very nature of jobs
and management
Sources: AI at Work, 2026 (n=11,749); BCG analysis.
1Share of respondents who report AI will significantly or moderately change the expectation for the skills needed for the role.
Aspect of work that respondents say have been changed by AI
Agree DisagreeNeutral
72%
67%
60%
52%
47%
41%
16%
17%
22%
23%
23%
24%
13%
16%
17%
25%
30%
35%
Changed skills expectations
for the role1
Took over simpler tasks, leaving
complex high-stakes work
The bar for what is considered
“good enough” is higher
Increased time spent reviewing
and correcting AI output
Roles shifted toward
managing and directing AI
Increased decision making
Business value and employee
joy aren’t tradeoffs, they’re
driven by the same forces
The “joy paradox”
of AI: it makes
work both better
and harder
Sources: AI at Work, 2026 (n=9,923 regular AI users); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities.
More than two-thirds of regular AI users report an increase in job satisfaction…
Change in day-to-day work enjoyment and satisfaction since adopting AI
… but 41% report increased mental strain associated with it
Change in cognitive load since adopting AI
67%All respondents
57%Frontline employees
65%Managers
77%Leaders
41%All respondents
37%Frontline employees
37%Managers
48%Leaders
The actions that
drive business
impact are the same
ones that make
employees thrive
Sources: AI at Work, 2026 (n=9,923 regular AI users); BCG analysis.
Note: Measurable impact = Improvement in key business metrics attributed to AI.
Top 5 organizational levers ranked by uplift across employee joy and
measurable impact
Enjoyment Measurable impact
Align AI actions with messaging
Track value from AI
Engage employees in AI ideation
Reward and recognize AI adoption
+14pp
+16pp
+20pp
+22pp
+13pp
+14pp
+28pp
+19pp
+24pp
+22pp
Redesign the organization around AI
Want the AI “honeymoon” to last?
Strategic clarity beats tools in
driving sustained impact
Strategic clarity
beats tools:
employees with clear
strategy but limited
access to tools
outperform those
with strong access
but no direction
Sources: AI at Work, 2026 (n=9,923 regular AI users); BCG analysis.
Notes: Strong strategic clarity is defined as making AI a priority and having a clear AI strategy and guidelines for time saved. Strong access to tools is
defined as having access to AI tools or to AI tools testing and experimentation.
Respondents reporting measurable impact based on AI strategy clarity
and availability of AI tools
55%
60%
80%
83%
Limited strategic clarity and
limited access to tools
Limited strategic clarity and
strong access to tools
Strong strategic clarity and
limited access to tools
Strong strategic clarity and
strong access to tools
+5pp
+25pp
The drivers of joy at work when using AI evolve over time...
…the drivers of toil do not
Regular AI users for less than 6 months Regular AI users for more than a year
Regular AI users for more than a yearRegular AI users for less than 6 months
The AI “honeymoon”?
At first employees
enjoy the cognitive
load, but strategic
clarity is the unlock
to sustaining joy
over time
Sources: AI at Work, 2026 (n=6,998 respondents who use AI regularly, less than 6 months vs more than a year); BCG analysis.
Note: +/−pp = the extra share of employees who enjoy work when a driver is present (+) or the share lost when a blocker is present (−).
Cognitive load
Clear AI strategy
Clear guidance
on saved time
Clear AI strategy
Clear guidance
on saved time
Cognitive load
Difficulty showing
unique value
Inadequate
training
Difficulty showing
unique value
Inadequate
training
+43pp
+24pp
+23pp
+30pp
+25pp
+21pp
–29pp
–15pp
–26pp
–16pp
AI agents have evolved from
concept to reality, but operating
models haven’t caught up
More people are
aware of AI agents
and see their
importance, but few
understand what
they are: a sign for
leaders to make a
better case for what
they can do
Sources: AI at Work, 2026 (n=11,749); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities.
Frontline employees LeadersManagersOverall respondents
Respondents who have
heard about AI agents
84% 87%
94%
72%
+12pp
vs 2025 overall
Respondents who have
a limited understanding
of what AI agents are
52%
61%
58%
39%
–15pp
vs 2025 overall
Respondents who think AI
agents will be important
in the next 2 to 3 years
72% 75%
86%
57%
–5pp
vs 2025 overall
More than double
the organizations
from 2025 have
integrated AI agents
into workflows, with
meaningful impacts
expected at the
managerial level
Sources: AI at Work, 2026 (n=9,923 regular AI users); BCG analysis.
Note: Frontline employees = individual white-collar employees with no managerial responsibilities.
Most people believe agents could do at least half their job in the future,
and managers and leaders expect the biggest shift
AI agent use by type
Respondents who think that in the next three years AI agents could perform at least half of their job
13%
30%
56%
50%
31%
20%
2025
2026
61%
52%
65%
All respondents
Frontline employees
Managers and leaders
Integrated into workflows Not yet deployedUsed experimentally, as in pilots
Respondents reporting measurable impact based on AI strategy clarity
and availability of AI tools
The missing piece:
operating models
haven’t caught up
with AI agents’
deployment
Sources: AI at Work, 2026 (n=8,849 respondents who have heard of AI agents and regular AI users); BCG analysis.
Notes: Frontline employees = individual white-collar employees with no managerial responsibilities.
Half of employees lack clear governance for managing human × AI teams:
frontline employees and managers feel it the most
Accountability is a universal concern, equally shared across all levels
Respondents who say their company has not put in place clear guidance on managing
human × AI teams
Respondents who rank AI-driven accountability as a top 3 concern in the next 2 to 3 years
50%Overall respondents
54%Frontline employees
55%Managers
41%Leaders
47%Overall respondents
46%Frontline employees
47%Managers
46%Leaders
Five CEO imperatives
Change the scoreboard:
measure value, not
adoption
Adoption tells you that people use AI, not whether it pays off. The time that individuals save leaks
out of the organization unless it is tracked and deliberately reinvested. So, watch business
outcomes rather than usage.
1
2
3
Make strategic clarity
your top priority, and
own it personally
Strategic clarity is not a communications task, it’s a leadership posture. Set AI as an explicit top
priority, be clear about where the company is heading, and make sure everyone gets it, the
frontline included. CEOs who personally own the transformation outperform on every dimension:
value captured, employee joy, and trust.
Invest in redesigning
work end-to-end, not in
more tools
Most companies still treat AI as a tool for individual productivity, but the more important change
is a collective one: AI is reshaping how teams work together and how tasks flow across the
organization. Capturing that value means redesigning a few core processes from end-to-end.
4 Put people at the heart of that redesign
The redesign only works if people are in it. Look ahead at how roles will shift over the next few
years, train for the skills that matter most, and bring people into shaping the change rather than
presenting it to them. What keeps people engaged is knowing how AI helps them grow, not how
much faster they work.
5
Govern it as a
moving target, not a
one-off program
Technology moves faster than any company can. Treat AI as something you keep steering, not a
program with a finish line. Put a light, standing governance in place that rechecks what works,
remeasures the value, and adjusts as the models and agents evolve.
