Transcriptly

What You Know That AI Doesn’t | Priyanka Vergadia | TED

TED09:19Added Aug 30, 2026

AI is good at seeing patterns, but it’s humans who figure out what to do next, says technologist Priyanka Vergadia. She shares three stories of human excellence sparked by AI insights and offers a pathway to identify and cultivate your irreplaceable qualities, turning the AI revolution from a threat into an opportunity. (Recorded at TEDNext 2025 on November 10, 2025)

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format
  1. 00:03Well, 71 percent of Americans believe
  2. 00:08that AI will cause massive job losses.
  3. 00:13Algorithms are getting smarter, faster,
  4. 00:16more capable every single day.
  5. 00:20My work puts me at the heart of this anxiety,
  6. 00:24where I bring AI applications to market for big tech companies
  7. 00:30and I help customers and businesses
  8. 00:34really take the potential of this technology further
  9. 00:38for their businesses.
  10. 00:40And through it all,
  11. 00:41I have seen brilliant professionals second-guess themselves
  12. 00:46as AI gets smarter.
  13. 00:48But let me tell you this one fundamental truth about AI.
  14. 00:54AI is excelling at identifying patterns.
  15. 00:59It understands data.
  16. 01:01We humans excel at understanding
  17. 01:05what these patterns actually mean
  18. 01:08in this beautifully chaotic world of human behavior.
  19. 01:12And even as these models and algorithms get stronger over time,
  20. 01:17this will stay true.
  21. 01:19Why?
  22. 01:20Because we understand things that cannot be quantified.
  23. 01:27Context, intent, unspoken emotions,
  24. 01:32cultural nuances.
  25. 01:35This depth of understanding comes from lived experiences
  26. 01:40that AI cannot replicate.
  27. 01:43So today I'll share with you three stories from my experience
  28. 01:49to prove this point that AI understands data
  29. 01:53and we understand experiences.
  30. 01:57And the key here is to not compete with AI,
  31. 02:02but to work with it
  32. 02:05while staying irreplaceably human.
  33. 02:09So how do we do that?
  34. 02:12Well, I was recently at a conference
  35. 02:14and met Sarah, a product manager.
  36. 02:18Her team has built an AI-powered analytics dashboard
  37. 02:22that's telling them very clearly
  38. 02:24that 80 percent of their users
  39. 02:28are only using basic features,
  40. 02:30and 20 percent are using advanced features here and there.
  41. 02:37Now Sarah looks at this data
  42. 02:40and she's like, OK, logically it makes sense.
  43. 02:43But she's questioning it.
  44. 02:46And this is the part I really love.
  45. 02:49She didn't just trust the algorithm as-is.
  46. 02:53She picked up the phone
  47. 02:55and called their 20 clients that were their top clients
  48. 02:59and asked them why they're not using these advanced features.
  49. 03:04Not to her surprise,
  50. 03:05she finds that they actually want to use these features,
  51. 03:09but they cannot find them
  52. 03:10because they are buried in some menu options,
  53. 03:13and the documentation isn't clear as well.
  54. 03:16Now, AI identified the pattern:
  55. 03:19that people are not using advanced features,
  56. 03:23but it totally missed the why behind it.
  57. 03:28Sarah's team goes in, rebuilds the entire experience,
  58. 03:31makes these features easier to find,
  59. 03:34and a few months later,
  60. 03:36the advanced feature adoption skyrockets.
  61. 03:41AI saw the symptom.
  62. 03:44Sarah diagnosed the disease.
  63. 03:50Now, the lesson that we take away from this example is clear.
  64. 03:56We've got to question the question.
  65. 03:58When AI recommends something, we need to ask why?
  66. 04:03If we continue to do that, we will be successful.
  67. 04:07On another occasion, I was working with a customer, Marcus,
  68. 04:10who is increasing sales efficiency using AI tools for their sales teams,
  69. 04:16analyzing the data through emails and engagement.
  70. 04:20And their AI tool is telling them
  71. 04:22that one of the biggest deals they have
  72. 04:26has a 95 percent probability to close.
  73. 04:30This was looking amazing.
  74. 04:32The data was saying positive sentiment, lots of engagement,
  75. 04:38but Marcus wanted to dig deeper and make sure that the deal happens.
  76. 04:43When he looks at the human element of this deal,
  77. 04:47he finds that ...
  78. 04:51Not the same people are showing up to these meetings.
  79. 04:54It's different stakeholders every time,
  80. 04:57and the responses in the emails have gotten vague
  81. 05:00and more corporate.
  82. 05:02AI is reading all of this activity as engagement.
  83. 05:07But really, there's something else going on behind the scenes.
  84. 05:11He dug a little further
  85. 05:13and identifies that the customer is going through a restructuring.
  86. 05:18And three teams thought that they owned the decision to make this purchase.
  87. 05:24If Marcus didn't get into this human element of the deal,
  88. 05:29the deal would never happen.
  89. 05:33AI identified the activities.
  90. 05:37Marcus measured meaning in those activities.
  91. 05:42So the lesson to learn from this story
  92. 05:45is you need to read the room,
  93. 05:49not just the dashboard.
  94. 05:54Understand those micro-expressions, the social cues in the room,
  95. 05:59the what are people saying,
  96. 06:01how are they nodding.
  97. 06:03We've all been in meetings where somebody says, "That's interesting."
  98. 06:08Are they politely dismissive or genuinely curious?
  99. 06:14Well, our emotional radar knows that.
  100. 06:17AI doesn't.
  101. 06:20I was with a friend recently,
  102. 06:22her name is Priya, and she works to use social media
  103. 06:28as a platform to help brands grow their revenue.
  104. 06:33Her AI tool is telling her to post fashion-hack videos,
  105. 06:38those videos where you get a lot of fashion tips out,
  106. 06:42for one of the brands.
  107. 06:43And she did that and they saw great engagement,
  108. 06:47lots of follower growth.
  109. 06:49But when talking to the team,
  110. 06:51they identified that none of that follower growth
  111. 06:54and engagement on social media
  112. 06:56was leading to sales or revenue.
  113. 07:01They were building the wrong audience.
  114. 07:03They were attracting bargain hunters,
  115. 07:06that was exactly opposite of the person
  116. 07:10who would pay 200 dollars to buy an ethically made jacket.
  117. 07:14This was what this brand makes.
  118. 07:18Now AI was optimizing for followers and engagement.
  119. 07:23Priya knew they were making the wrong audience,
  120. 07:26so she flips the switch.
  121. 07:28She stops taking AI-recommended content,
  122. 07:32instead, starts building content that is showing sustainable cost
  123. 07:39of building these fashion items.
  124. 07:43She started showing stories of artisans that were making these clothes.
  125. 07:49Now AI in this case was optimizing for activity and engagement.
  126. 07:55Priya optimized for building a community.
  127. 08:01And they started seeing the sales skyrocket.
  128. 08:07So the lesson that we learn here is
  129. 08:11always pause and ask,
  130. 08:13what is the story behind this data?
  131. 08:17And only we can do that.
  132. 08:20So if you see all these examples, there's one thing very common.
  133. 08:26The future doesn't belong to humans or AI.
  134. 08:31It belongs to humans that work closely with AI
  135. 08:35while staying irreplaceably human.
  136. 08:41Our ability to read the room,
  137. 08:44our ability to look at emotions,
  138. 08:49that is irreplaceable.
  139. 08:51Our ability to empathize with people,
  140. 08:54that's irreplaceable.
  141. 08:56So the next time ...
  142. 09:00You're feeling anxious about AI taking your job,
  143. 09:05remember that AI can identify patterns.
  144. 09:09Only we,
  145. 09:11and you can identify the human behind it.
  146. 09:15Thank you.
  147. 09:16(Applause)