The recent MIT research on enterprise AI use has often been incorrectly reported as 95% of AI fails.
The proper headline is 95% of AI approaches fail.
Why?
Last summer we shared concerning research - enterprises spending $5m on $1m problems.
Early this year we published analyses on how such directionless spend and immoderate expectations from AI could soon tank sentiment below the rising potential of AI, leading to missed opportunities.
We see 5 common patterns at firms with underperforming programs from case work and discussion with executives, many of which are alluded to in the study:
1) Setup AI committees and innovation labs with sprawling red tape, low velocity and ivory tower thinking.
2) Appointed bureaucrats to lead the program who thought it would be a wonder drug and lacked the intuition and agility to cope with applying AI.
3) Lacked a strategic plan and let overzealous engineering and data teams run without clarity to develop in-house AI wrappers and tools, often for commodity solutions available in the market.
4) Spent little effort on user experience, workflow integration and managing adoption and behavioral change.
5) Ignored the hard task of designing operating model shifts. Executives looking to shave off weekly 4hrs/FTE had no plan to leverage the 10% productivity gain. More time for cafeteria foosball?
The committees have had great fun, held countless meetings, massaged metrics on slides, churned out platitudes and burned millions of corporate dollars,
Meanwhile, no one else has had fun. Many workers are quietly using shadow AI where they can for personal productivity gains. And AI shows up as cost on the P&L to fund this bureaucracy.
This is the structure of technology revolutions. Years ago similar reports stated that most digital transformations fail yet many companies ultimately found value. An inability to understand novel developments and resulting frame shifts leads to fear, uncertainty and unwise deployment of attention and capital.
Some organizations learn and thrive, new natives emerge and others get bought or fold. Fundamentals of human and organizational behavior and fallibility hasn't evolved much if we study history keenly.
hashtag#ai hashtag#transformation hashtag#mit
The proper headline is 95% of AI approaches fail.
Why?
Last summer we shared concerning research - enterprises spending $5m on $1m problems.
Early this year we published analyses on how such directionless spend and immoderate expectations from AI could soon tank sentiment below the rising potential of AI, leading to missed opportunities.
We see 5 common patterns at firms with underperforming programs from case work and discussion with executives, many of which are alluded to in the study:
1) Setup AI committees and innovation labs with sprawling red tape, low velocity and ivory tower thinking.
2) Appointed bureaucrats to lead the program who thought it would be a wonder drug and lacked the intuition and agility to cope with applying AI.
3) Lacked a strategic plan and let overzealous engineering and data teams run without clarity to develop in-house AI wrappers and tools, often for commodity solutions available in the market.
4) Spent little effort on user experience, workflow integration and managing adoption and behavioral change.
5) Ignored the hard task of designing operating model shifts. Executives looking to shave off weekly 4hrs/FTE had no plan to leverage the 10% productivity gain. More time for cafeteria foosball?
The committees have had great fun, held countless meetings, massaged metrics on slides, churned out platitudes and burned millions of corporate dollars,
Meanwhile, no one else has had fun. Many workers are quietly using shadow AI where they can for personal productivity gains. And AI shows up as cost on the P&L to fund this bureaucracy.
This is the structure of technology revolutions. Years ago similar reports stated that most digital transformations fail yet many companies ultimately found value. An inability to understand novel developments and resulting frame shifts leads to fear, uncertainty and unwise deployment of attention and capital.
Some organizations learn and thrive, new natives emerge and others get bought or fold. Fundamentals of human and organizational behavior and fallibility hasn't evolved much if we study history keenly.
hashtag#ai hashtag#transformation hashtag#mit
The recent MIT research on enterprise AI use has often been incorrectly reported as 95% of AI fails.
The proper headline is 95% of AI approaches fail.
Why?
Last summer we shared concerning research - enterprises spending $5m on $1m problems.
Early this year we published analyses on how such directionless spend and immoderate expectations from AI could soon tank sentiment below the rising potential of AI, leading to missed opportunities.
We see 5 common patterns at firms with underperforming programs from case work and discussion with executives, many of which are alluded to in the study:
1) Setup AI committees and innovation labs with sprawling red tape, low velocity and ivory tower thinking.
2) Appointed bureaucrats to lead the program who thought it would be a wonder drug and lacked the intuition and agility to cope with applying AI.
3) Lacked a strategic plan and let overzealous engineering and data teams run without clarity to develop in-house AI wrappers and tools, often for commodity solutions available in the market.
4) Spent little effort on user experience, workflow integration and managing adoption and behavioral change.
5) Ignored the hard task of designing operating model shifts. Executives looking to shave off weekly 4hrs/FTE had no plan to leverage the 10% productivity gain. More time for cafeteria foosball?
The committees have had great fun, held countless meetings, massaged metrics on slides, churned out platitudes and burned millions of corporate dollars,
Meanwhile, no one else has had fun. Many workers are quietly using shadow AI where they can for personal productivity gains. And AI shows up as cost on the P&L to fund this bureaucracy.
This is the structure of technology revolutions. Years ago similar reports stated that most digital transformations fail yet many companies ultimately found value. An inability to understand novel developments and resulting frame shifts leads to fear, uncertainty and unwise deployment of attention and capital.
Some organizations learn and thrive, new natives emerge and others get bought or fold. Fundamentals of human and organizational behavior and fallibility hasn't evolved much if we study history keenly.
hashtag#ai hashtag#transformation hashtag#mit
0 Comments
0 Shares
740 Views
0 Reviews
Arabic
French
Spanish
Portuguese
Deutsch
Turkish
Dutch
Italiano
Russian
Romaian
Portuguese (Brazil)
Greek
