Latest / Chief in Tech Deep Dive Shortcast / Chief in Tech S1E11 - Linda Yao, Poonam Garg: Visionary Leadership:Shaping the Future of Business with AI & Emerging Technologies

Transcript

Chief in Tech S1E11 - Linda Yao, Poonam Garg: Visionary Leadership:Shaping the Future of Business with AI & Emerging Technologies · Chief in Tech Deep Dive Shortcast

• 0:00
  Welcome back to the Deep Dive. Today, we're really
• 0:03
  on a mission. We want to cut through all the
• 0:05
  noise, the hype around AI adoption. Yeah, there's
• 0:09
  a lot of it. There is. And dive straight into
• 0:11
  the practical stuff. Enterprise readiness. We've
• 0:13
  gathered sources from a major player in tech
• 0:16
  services. A leader. And our goal here is to give
• 0:18
  you a clear roadmap. How do you get from just
• 0:21
  wanting AI adoption to actually being ready and
• 0:25
  scaling it across the whole organization? In
• 0:28
  this Deep Dive, It really benefits from a perspective
• 0:31
  that's, well, deeply rooted in actually doing
• 0:33
  it, getting it done. Execution. Exactly. The
• 0:35
  specific insights, they come from Linda Yow.
• 0:38
  She's the COO for Lenovo's Solutions and Services
• 0:41
  Group. OK. Her background is, frankly, quite
• 0:43
  unique. She was employee number two at Linamar,
• 0:46
  so she has that real rapid growth foundational
• 0:48
  experience. Wow. And crucially, she spent five
• 0:51
  years incubating data science practices at Boeing
• 0:54
  Global Services. Managed teams in the US, in
• 0:57
  India. So real global perspective. Totally. And
• 1:01
  her focus is on services, the delivery, the integration,
• 1:05
  not just the hardware. That's why her views on
• 1:07
  what works are so well grounded, practical. OK,
• 1:11
  let's definitely unpack this. Before we jump
• 1:13
  into the methodology she shared, maybe let's
• 1:15
  frame the moment we're in. What's the single
• 1:17
  biggest trend that explains why technology adoption,
• 1:20
  especially AI, is suddenly a boardroom issue,
• 1:24
  not just, you know, an IT department thing anymore?
• 1:27
  Well, the sources are pretty clear on this. Yeah.
• 1:29
  The role of IT, it's fundamentally changed. How
• 1:33
  so? IT used to be, you know, the back office,
• 1:36
  a cost center, just keeping the lights on, keeping
• 1:38
  systems running. Right. But now, technology is
• 1:40
  moving rapidly out of that back room and directly
• 1:43
  into the hands of the business units themselves.
• 1:44
  OK, wait. So if IT is suddenly everywhere, what
• 1:47
  does that actually mean for the CIO? the person
• 1:50
  in charge. It means the CIO has basically transformed
• 1:53
  into an IT business partner. They're responsible
• 1:56
  not just for keeping servers running, but for
• 1:58
  frontline business outcomes. For P &L. For results.
• 2:02
  Absolutely. And this forces a new mandate. Technologies,
• 2:06
  especially something potentially massive like
• 2:08
  AI, must have demonstrable practical grounding.
• 2:11
  It has to work seamlessly from the headquarters
• 2:15
  boardroom all the way out to, say, a factory
• 2:18
  floor or somewhere else. That sounds like a huge
• 2:20
  shift in accountability, though, if the CIO is
• 2:23
  now on the hook for outcomes in P &L. I mean,
• 2:25
  aren't we just giving them massive new risks
• 2:28
  without maybe giving them more control over everything?
• 2:30
  That's precisely why the conversation has escalated,
• 2:33
  right, to the board level. The stakes are just
• 2:35
  much, much higher now. OK. And the major catalyst,
• 2:37
  the thing that really accelerated this whole
• 2:39
  high -stakes shift, well, It's generative AI.
• 2:43
  Right. GenAI. Of course. It's given AI yet another
• 2:45
  reason to be front and center because GenAI makes
• 2:48
  it truly collaborative, productive, not just
• 2:51
  for the organization as a whole, but for every
• 2:53
  single individual user. And we're seeing real
• 2:55
  tangible effects already, aren't we? Not just
• 2:58
  theory. Exactly. Lindy Au specifically pointed
• 3:01
  out companies using GenAI right now to run, say,
• 3:05
  a much more efficient contact center. or to dramatically
• 3:09
  streamline how software engineering gets done.
• 3:12
  And because gen AIs by its nature sort of democratized
• 3:15
  it, you access it with natural language, right?
• 3:18
  It creates this incredibly powerful multiplicative
• 3:21
  effect. It scales knowledge instantly, broadly
• 3:25
  across the entire enterprise. OK, so that brings
• 3:28
  us right to the big challenge. You've got senior
• 3:30
  executives pushing hard for AI adoption. They
• 3:32
  are. And interestingly, unlike maybe some previous
• 3:36
  tech waves, people aren't really skeptical. They
• 3:38
  seem eager. They're taking courses, watching
• 3:40
  videos. That desire is definitely there. People
• 3:43
  want to embrace it, yes. So if the willingness
• 3:45
  is there, why are so many companies still kind
• 3:47
  of stuck, you know, stuck in pilot project mode,
• 3:50
  failing to actually scale this thing up? Because
• 3:53
  desire and, let's say, Infrastructure readiness
• 3:56
  are two very, very different things. Once organizations
• 3:59
  move past that initial, wow, this is cool excitement,
• 4:03
  they hit these massive, practical roadblocks.
• 4:06
  The willingness is there, sure, but when you
• 4:08
  look at large -scale enterprise adoption, the
• 4:11
  questions just become overwhelming. Like what?
• 4:13
  What are the big ones? Well, first, where exactly
• 4:15
  do we even start this whole process? Then. Is
• 4:19
  our security posture actually strong enough?
• 4:22
  Can we protect our proprietary data while we're
• 4:25
  training these models? Huge concern. Huge. Then
• 4:28
  policies. Are corporate policies mature enough?
• 4:32
  Can they handle responsible AI use? Things like
• 4:34
  bias, ethics. Right, the responsible AI piece.
• 4:37
  And then the immediate stuff. Has our talent
• 4:39
  actually been up skilled for this? Do people
• 4:41
  know how to use these tools effectively? And
• 4:44
  fundamentally. Is the underlying tech stack robust
• 4:47
  enough to even support these capabilities at
• 4:49
  scale? Hmm. This sounds less like a pure technology
• 4:52
  problem and maybe more like a leadership and
• 4:55
  process problem. I think that's fair to say.
• 4:57
  So if readiness is the big hurdle, what's the
• 4:59
  solution? Did Linda Yao and Lenovo share a kind
• 5:02
  of battle -tested methodology for getting over
• 5:04
  it? They did. They have a proven framework. It
• 5:07
  centers on four core pillars. And the guiding
• 5:09
  principle, sort of the operational mantra, is
• 5:12
  security first and people next. Okay, security
• 5:15
  first, people next. Let's break down those four
• 5:18
  pillars. This feels like where the really actionable
• 5:20
  insights are. Pillar one, security. Why start
• 5:26
  there? Why is that the immediate top priority?
• 5:28
  Security is defined, quite simply, as the thing
• 5:32
  that keeps leadership up at night. Huh, okay,
• 5:35
  fair enough. It's not just about warding off
• 5:37
  external hackers, right? It's also about managing
• 5:39
  internal risk, data security, privacy. For AI
• 5:43
  specifically, this means things like mitigating
• 5:45
  the risk of your own intellectual property leaking
• 5:48
  out when models are trained. Right, your secret
• 5:50
  sauce getting into the model. Exactly. and also
• 5:53
  guarding against the very serious threat of data
• 5:55
  poisoning, where someone could deliberately mess
• 5:58
  with your training data to compromise the AI's
• 6:01
  output, its trustworthiness. Okay, so you absolutely
• 6:03
  have to lay down those foundational security
• 6:05
  guardrails before you do anything else. You have
• 6:07
  to. It's non -negotiable. That clarity is crucial.
• 6:10
  So if security is the first really high -stakes
• 6:12
  barrier, the second pillar is people. How do
• 6:16
  they approach the human side of this huge transition?
• 6:20
  Pillar number two, people. This focuses really
• 6:22
  heavily on change management, on adoption, and
• 6:25
  crucially on training. To get buy -in to actually
• 6:29
  make this work, organizations must clearly define
• 6:32
  which tasks GEN .AI is going to augment. Make
• 6:36
  human workers better, faster, versus which tasks
• 6:40
  it might actually replace. That distinction feels
• 6:42
  critical. Augment versus replace. It's absolutely
• 6:44
  critical for trust and adoption. Can you give
• 6:46
  us the sort of concrete example of that? Augment
• 6:48
  versus replace. Sure. So augmentation might look
• 6:50
  like GNI drafting the first version of a complex
• 6:54
  legal brief maybe, or an internal memo. The human
• 6:57
  lawyer or the manager then takes that draft,
• 6:59
  refines it, approves it, it speeds things up.
• 7:01
  Got it. Faster, better work. Right. Replacement,
• 7:04
  on the other hand. Might be GNI handling, say,
• 7:06
  100 % of tier one customer support chats, routing
• 7:08
  issues, classifying common problems, maybe even
• 7:11
  resolving simple ones entirely on its own. Completely
• 7:13
  automated. Exactly. Getting employees to really
• 7:15
  understand which category their specific tasks
• 7:19
  fall into that's essential for building trust
• 7:21
  and making the transition smooth. OK, so if you
• 7:24
  manage security and people, arguably the two
• 7:26
  hardest, most human -centric pillars, What about
• 7:30
  the actual technology needed? Is that the easy
• 7:32
  part? Well, that's pillar number three, technology.
• 7:36
  And interestingly, the expert expressed well.
• 7:40
  quite strong confidence here. Really? Yeah. The
• 7:42
  thinking is if you have your security parameters
• 7:44
  locked down and your people processes are clear
• 7:47
  and working, the technology itself can be iterative.
• 7:51
  Modern big tech players, they're agile, they
• 7:53
  can adapt, they can co -develop solutions, they
• 7:55
  can pivot if needed. The confidence level seems
• 7:58
  high, that if the enterprise knows what it needs
• 8:01
  the AI to do and how to secure it and manage
• 8:03
  the people side, then they can reliably engineer
• 8:06
  the tech solution to do it. That's interesting.
• 8:08
  So the tech isn't the main bottleneck if the
• 8:10
  other pieces are right. It seems to be the perspective.
• 8:12
  OK. And that brings us to the final pillar. Pillar
• 8:15
  four, process. Right. Process. So once security
• 8:20
  and people parameters are set, this is where
• 8:22
  deep industry experience really becomes paramount.
• 8:25
  You define the new workflows. You integrate those
• 8:28
  security guardrails and ethical checks. And then
• 8:31
  you standardize those processes so you can scale
• 8:33
  them across every business unit. So it's not
• 8:35
  just a one -off experiment? Yeah, exactly. This
• 8:38
  structure ensures that a successful pilot project
• 8:40
  doesn't just date a cool little experiment in
• 8:42
  one corner of the company, it becomes a repeatable
• 8:45
  blueprint for everyone. Okay, we have the methodology,
• 8:48
  security, people, technology, process. But, you
• 8:51
  know, in a world full of consultants selling
• 8:54
  frameworks, What did Linda Yao identify as the
• 8:57
  real secret sauce? What separates the companies
• 9:00
  that successfully scale this from those just
• 9:03
  stuck reviewing PowerPoint decks? Yeah, yeah.
• 9:06
  The secret sauce, according to her, isn't some
• 9:08
  magic bullet. It's internal expertise, or as
• 9:11
  they call it, eating its own cooking. Eating
• 9:13
  its own cooking. Meaning Lenovo acts as its own
• 9:16
  internal sandbox, its own guinea pig, basically.
• 9:18
  They implement and they refine AI use cases internally
• 9:22
  first in their own massive global contact centers,
• 9:26
  in their software engineering teams, across individual
• 9:29
  employee productivity tools. They do it themselves
• 9:32
  before they start rolling out similar solutions
• 9:34
  or advice to clients. OK, but how does being
• 9:37
  your own guinea pig actually prevent that common
• 9:39
  problem, where an AI solution looks amazing in
• 9:42
  a controlled lab setting, but then completely
• 9:44
  falls apart in a real messy enterprise environment?
• 9:47
  Because they hit those real -world roadblocks
• 9:50
  first themselves. Ah, they feel the pain. Exactly.
• 9:53
  They learn the hard way how to integrate these
• 9:55
  new tools with messy legacy systems. They figure
• 9:58
  out how to actually train a diverse global workforce
• 10:01
  effectively. They find the policy gaps and the
• 10:04
  practical hurdles. So when they advise clients,
• 10:06
  they're offering a guidebook based on, well...
• 10:08
  failure and iteration and real experience, not
• 10:12
  just abstract theory. That makes a lot of sense.
• 10:14
  Okay, let's shift gears a bit and look at the
• 10:15
  history here. You mentioned AI has been around
• 10:17
  for decades in some form. So why now? Why are
• 10:21
  senior management and boards dedicating so much
• 10:23
  focus, so much capital to it right now? Yeah,
• 10:26
  that's a great question. We can understand the
• 10:28
  current, let's call it, excitement level better
• 10:30
  if we put it in context with the previous two
• 10:33
  sort of waves of AI enthusiasm. The core tech
• 10:37
  might have been there, but the enabling environment
• 10:39
  wasn't quite right before. OK. So what defined
• 10:43
  that first wave? The first wave, I think, roughly
• 10:47
  mid -2000s. It's often called the Flashboys era.
• 10:50
  Ah, high frequency trading. Exactly. HFT in the
• 10:53
  financial sector. Using complex algorithms, super
• 10:56
  high performance computing to execute financial
• 10:59
  trades at, you know, lightning speed. Right.
• 11:01
  This definitely proved AI's raw power and speed,
• 11:04
  but its impact was pretty niche, right? Highly
• 11:07
  specialized. It did, however, spark some of the
• 11:09
  first really serious conversations around needing
• 11:12
  new policies, new processes to govern this stuff.
• 11:15
  OK, so it showed power in finance, but maybe
• 11:17
  not broad applicability for the average company.
• 11:19
  What changed in the second wave? That was more
• 11:21
  the 2010s, and it was defined primarily by the
• 11:25
  rise of accessible cloud computing. Cloud. Cloud
• 11:28
  made computing power relatively cheap. scalable,
• 11:31
  easy to access for almost anyone. And this just
• 11:35
  fundamentally boosted awareness of the value
• 11:37
  of data. It spurred the adoption of sophisticated
• 11:40
  data visualization tools. And it led to, well,
• 11:44
  almost every industry... hiring data science
• 11:47
  departments. That's when DI really took off.
• 11:49
  Absolutely. Business intelligence, foundational
• 11:51
  data governance, that all matured significantly
• 11:54
  during that second wave. OK, so we went from
• 11:56
  niche finance innovation with HFT to widespread
• 11:59
  data awareness thanks to the cloud. And now we
• 12:01
  hit this third wave, gen AI democratization in,
• 12:05
  say, 2024. If gen AI is supposed to be the ultimate
• 12:08
  goal, the big prize, does that mean all that
• 12:10
  foundational BI work, the data governance from
• 12:12
  the last wave is that now obsolete, should companies
• 12:15
  just leapfrog that? Oh, quite the opposite, actually.
• 12:18
  Linda Yao suggested that GNI is like the tide
• 12:21
  that's going to raise all boats. OK, explain
• 12:23
  that. How does the fancy new stuff help the older
• 12:25
  foundational stuff? Well, here's the key takeaway
• 12:28
  she offered. While loads of organizations are
• 12:31
  racing toward full GNI adoption, many quickly
• 12:35
  discover they aren't technically ready for it
• 12:37
  yet. They lack the data quality, the governance.
• 12:40
  Right, the stuff from Wave 2. Exactly. But the
• 12:43
  sheer push for GEN .AI from the top forces these
• 12:46
  companies to move up the data and analytics value
• 12:49
  chain anyway, even if they can't implement full
• 12:51
  GEN .AI tomorrow. So the ambition drives foundational
• 12:54
  improvement. Precisely. It means investment,
• 12:57
  education, executive attention. It's now massively
• 13:01
  boosted for those fundamental building blocks,
• 13:04
  data quality, data governance, even traditional
• 13:06
  BI. The fundamentals end up benefiting enormously
• 13:09
  from all the excitement around the advanced stuff.
• 13:11
  That connects the dots really well between organizational
• 13:14
  priorities. Okay, let's circle back to security
• 13:16
  for a moment, specifically thinking about the
• 13:18
  modern, often remote workforce. How on earth
• 13:21
  do you manage AI risks when employees are using
• 13:23
  these tools outside the company's physical walls,
• 13:27
  potentially using apps you haven't even approved?
• 13:29
  Yeah, that's a huge challenge. Yeah. And it sits
• 13:31
  right at that critical intersection of the security
• 13:34
  pillar and the people pillar. OK. The expert
• 13:38
  view here was that managing AI outside the company
• 13:41
  walls isn't about trying to build some giant,
• 13:44
  impenetrable digital fence. That's probably futile.
• 13:46
  Right. Instead, it requires identifying the user
• 13:49
  persona and the specific use case they're working
• 13:52
  on. That allows you to define appropriate tailored
• 13:55
  guardrails. Because everyone uses AI differently,
• 13:58
  right? Yeah. And their need for data access varies
• 14:00
  hugely. Can you elaborate on that? How different
• 14:03
  can those personas be within one company? Oh,
• 14:05
  absolutely. Think about three distinct users
• 14:08
  that were mentioned in the source material. First,
• 14:10
  you might have your typical office worker, maybe
• 14:12
  dialing in from an airport lounge, using AI for
• 14:15
  quick collaboration, maybe drafting emails, prepping
• 14:18
  for a big presentation. Pretty standard productivity
• 14:20
  stuff. OK, persona one. Second, picture someone
• 14:22
  on a manufacturing floor. They might be using
• 14:24
  computer vision. powered by AI to inspect the
• 14:27
  quality of components coming down the line in
• 14:29
  real time. Very different context, different
• 14:31
  data. Totally different. Okay, third. Third,
• 14:34
  maybe someone in the ESG office. environmental
• 14:37
  social governance. They could be using AI to
• 14:40
  review vast amounts of documents, maybe legal
• 14:42
  transcripts, regulatory filings, to monitor environmental
• 14:45
  compliance. Again, highly sensitive data, specific
• 14:49
  task. And the guardrails, the security controls
• 14:51
  for each of those three must be fundamentally
• 14:53
  different. Precisely. The manufacturing worker
• 14:55
  needs guardrails around the integrity and accuracy
• 14:58
  of that vision model's output. Is it spotting
• 15:00
  defects correctly? The ESG analyst needs strict
• 15:03
  guardrails around data privacy, handling potentially
• 15:06
  sensitive of regulatory or legal documents. The
• 15:09
  office worker needs controls around maybe not
• 15:11
  pasting confidential company strategy into a
• 15:13
  public AI tool. The solution has to be tailored
• 15:16
  to the objective, the data sensitivity, and the
• 15:19
  user's actual need to know, which loops right
• 15:22
  back to why that people pillar understanding
• 15:24
  users, training them is so critical alongside
• 15:27
  security. Okay, finally then, let's tackle the
• 15:30
  broader, more general risks we always hear about
• 15:33
  with AI. Things like bias creeping into models,
• 15:36
  AI hallucinating or making things up, ethical
• 15:38
  concerns, privacy violations. How do we manage
• 15:41
  those? Well, the perspective shared was that
• 15:43
  managing these kinds of ethical and quality risks,
• 15:46
  it always points back to those four foundational
• 15:48
  pillars again. Security, people, technology,
• 15:52
  process. Exactly. Risk mitigation here requires
• 15:55
  careful, safe experimentation. It really comes
• 15:58
  down to fundamentally understanding the user,
• 16:00
  their objective, what they're trying to achieve
• 16:02
  with the AI. and then ensuring you hire or maybe
• 16:05
  retain the right internal or external experts
• 16:08
  who can advise on how to draw those specific
• 16:10
  parameters for safe use and testing. So it's
• 16:13
  about structured management, not just hoping
• 16:15
  for the best. Absolutely. In the experience shared
• 16:18
  in the sources, almost every single risk management
• 16:20
  concern you can think of, whether it's bias,
• 16:23
  hallucination, privacy, ethics. It can usually
• 16:26
  be traced back to a weakness in one or more of
• 16:29
  those four boxes. Maybe poor security protocols
• 16:32
  allowed bad data in. Maybe inadequate training
• 16:35
  meant people used it wrong. Perhaps an unstable
• 16:37
  tech architecture led to unreliable outputs.
• 16:40
  Or maybe there was just a lack of defined processes
• 16:43
  for checking errors or handling exceptions. That
• 16:45
  four pillar structure seems incredibly robust
• 16:48
  for diagnosing problems then. Okay, we've covered
• 16:51
  readiness, the history of the risks. One last
• 16:53
  big question. Does adopting these emerging technologies
• 16:56
  like AI, does it fundamentally change a company's
• 16:59
  business model? Or does it mostly just make the
• 17:01
  existing model run more efficiently? That's a
• 17:04
  challenging question. And the insight here was,
• 17:05
  well, quite pointed. It doesn't always change
• 17:07
  the business model. Yeah. But the strong consensus
• 17:10
  seemed to be that it should. It should. Why?
• 17:13
  Think about it. If a company puts this enormous
• 17:16
  effort, massive investment in time, money, focus
• 17:19
  into refining AI, making it secure, ethical,
• 17:23
  responsible, unbiased, adaptable, all that work,
• 17:26
  and yet their fundamental business model remains
• 17:29
  static, unchanged, that arguably that enormous
• 17:33
  effort was ultimately wasted, or at least under
• 17:36
  -leveraged. The potential for true business model
• 17:39
  transformation, that's seen as the real payoff.
• 17:42
  That's what justifies the immense effort involved.
• 17:45
  That really brings us to a compelling final thought
• 17:47
  for you, the listener. History kind of shows
• 17:49
  us, doesn't it, that whenever a powerful new
• 17:51
  technology emerges, I think the internet, maybe
• 17:53
  mobile phones, now AI, there's always this corresponding
• 17:56
  wave of concern about relevance. People worry,
• 17:59
  right? Will this replace my job? And it's understandable.
• 18:02
  But historical data, again and again, consistently
• 18:05
  shows that technology generally helps augment
• 18:07
  most workers. It tends to take over the tedious,
• 18:10
  the repetitive tasks. Freeing people up. Exactly.
• 18:12
  Making our jobs hopefully more productive, maybe
• 18:15
  even more enjoyable, and allowing us as individuals
• 18:17
  to focus our unique human expertise on higher
• 18:20
  value, more creative, more strategic tasks. That's
• 18:23
  a great perspective to maybe end on. So considering
• 18:27
  those three ways of AI, we talked about the niche
• 18:29
  HRT wave, the cloud and data awareness wave,
• 18:32
  and now this gen AI democratization wave. Here's
• 18:35
  a final thought for you, the listener. to ponder,
• 18:38
  where do you see the next big policy debate emerging?
• 18:41
  Will the political and corporate focus be mainly
• 18:43
  on security, on preventing IP leakage and misuse?
• 18:46
  Or will it shift more towards defining those
• 18:48
  tricky societal boundaries between automation
• 18:50
  and argumentation? And maybe more personally,
• 18:53
  what skill will you prioritize learning or honing
• 18:56
  next now that you know those four pillars of
• 18:58
  readiness? Some interesting questions to think
• 19:00
  about. Thank you for joining us on this deep
• 19:02
  dive into strategic AI readiness and leadership.
• 19:04
  Until next time.