Latest / Chief in Tech Deep Dive Shortcast / Chief in Tech S1E19 - Neha Singh, Jeannette Martin, Laura Mosley and Vandana Khanna: Automating Asset Management - Finding Efficiency with Artificial Intelligence

Transcript

Chief in Tech S1E19 - Neha Singh, Jeannette Martin, Laura Mosley and Vandana Khanna: Automating Asset Management - Finding Efficiency with Artificial Intelligence · Chief in Tech Deep Dive Shortcast

• 0:00
  Welcome to this deep dive. We are really glad
• 0:03
  you're here with us today. Yeah, definitely great
• 0:05
  to have you listening If you are looking to just
• 0:08
  bypass all that dense corporate jargon and get
• 0:11
  straight to the actual mechanics of where technology
• 0:14
  is taking high stakes finance, you are in the
• 0:17
  exact right place. It is a totally fascinating
• 0:19
  landscape to explore right now. I mean, we are
• 0:21
  moving way past the theoretical stuff. Right.
• 0:24
  And we're looking at the actual plumbing of how
• 0:26
  these major institutions operate on a daily basis.
• 0:29
  Exactly. So our mission today is to explore how
• 0:31
  artificial intelligence is just completely rewiring
• 0:35
  set management. Down to the studs. Down to the
• 0:37
  studs. We're pulling insights from a really comprehensive
• 0:40
  panel discussion today. It features digital finance
• 0:43
  and security leaders from MetLife, Capgemini
• 0:46
  Invent, PNC Bank, and PGM Global Services. Heavy
• 0:50
  hitters. Yeah, very. And the sheer scale of what
• 0:53
  they are discussing is honestly, it's staggering.
• 0:55
  We are not talking about using a chat bot to
• 0:58
  summarize an article for you. No. We are looking
• 1:00
  at a fundamental shift in how trillions of dollars
• 1:03
  in global wealth are managed, secured, and optimized.
• 1:06
  And that scale, honestly, that is precisely why
• 1:09
  this matters to you, even if you don't manage
• 1:12
  a massive institutional portfolio. Right. Even
• 1:14
  if you're just interested in how enterprise tech
• 1:17
  is evolving. Exactly. What we are really examining
• 1:20
  today is the ultimate modern tightrope walk for
• 1:24
  these massive firms. They're trying to achieve
• 1:27
  this this hyper -efficiency through AI. Which
• 1:31
  everyone wants. Which everyone wants, but while
• 1:33
  maintaining the absolute non -negotiable necessity
• 1:36
  of security. Yeah. One of the panelists from
• 1:38
  PNC Bank framed this perfectly, I think. They
• 1:40
  said the guiding principle for their AI adoption
• 1:43
  is to have foresight, not fear. Foresight, not
• 1:46
  fear. I love that. I think that's a great anchor
• 1:48
  for our discussion, especially when it is just
• 1:50
  so easy to either panic about security or get
• 1:53
  completely swept away by the vendor hype. Oh,
• 1:55
  the hype is everywhere. It is. OK. Let's unpack
• 1:57
  this, starting with the actual productivity leap
• 2:01
  we are seeing on the ground. The panel sources
• 2:04
  cited that generative AI is currently being used
• 2:07
  to reduce manual processes by as much as 80%.
• 2:11
  80%. Yeah, and now, I have to admit, whenever
• 2:14
  I hear a perfectly round number like 80%, I get
• 2:17
  a little skeptical. Sure. It sounds like marketing
• 2:19
  math. What does that actually look like in the
• 2:22
  trenches? That's a healthy skepticism to have.
• 2:24
  But when you look at the specific use cases they
• 2:26
  highlighted, that number starts to make a lot
• 2:29
  of sense. OK. Take the traditional KYC process,
• 2:32
  for example. Know your customer. Right. We all
• 2:35
  know how notoriously bottlenecked that can get.
• 2:39
  Historically, you have analysts just bogged down
• 2:41
  in the mechanics of verification. Pulling disparate
• 2:43
  records. Right. Cross -referencing complex corporate
• 2:46
  hierarchies. and digging through massive unstructured
• 2:49
  PDF documents? just to validate an entity. Oh,
• 2:53
  the dreaded unstructured PDF. Anyone who has
• 2:56
  ever had to manually pull specific data points
• 2:59
  out of a messy 200 -page legal document and put
• 3:02
  it into a spreadsheet knows the sheer physical
• 3:05
  eye strain involved. You're just living with
• 3:08
  dual monitors, using keyboard shortcuts, and
• 3:11
  praying you didn't misread a decimal point. Exactly.
• 3:14
  And what the AI is doing is sweeping that mechanical
• 3:17
  friction away. It effortlessly parses those unstructured
• 3:20
  documents, pulls the relevant entities out, and
• 3:23
  formats the data for the human analysts to review.
• 3:26
  Wow. But here is the critical pivot in the technology.
• 3:30
  We are evolving from simple AI into what the
• 3:34
  industry is calling agentic AI. Agentic AI. Okay,
• 3:37
  that feels like a massive turning point. Break
• 3:39
  down agentic AI for us because that is a term
• 3:41
  a lot of people are hearing but might not fully
• 3:43
  grasp in a corporate context. Sure. Up until
• 3:45
  recently, enterprise AI has mostly been a prompt
• 3:48
  and response tool. You ask a system a question,
• 3:51
  it generates text or code in return. Like a really
• 3:53
  smart search engine. Exactly. Agentic AI is fundamentally
• 3:58
  different. It acts as a proactive assistant that
• 4:01
  has actually granted permission to manage end
• 4:05
  -to -end processes across different software
• 4:07
  platforms. Oh, wow. Yeah. It isn't just waiting
• 4:10
  for a prompt. It is anticipating the next logical
• 4:13
  step. making API calls, and queuing up actions
• 4:17
  for a human to approve. So it shifts the user's
• 4:19
  role from being a fast data gatherer into an
• 4:22
  actual strategic decision maker. Spot on. That
• 4:25
  makes a lot of sense. To make this tangible for
• 4:27
  you, the listener, let's look at a case study
• 4:29
  the Capgemini panelists shared. They applied
• 4:31
  this to the client contact center. A great example.
• 4:34
  Yeah. And instead of a generic retail banking
• 4:36
  example, let's put this in the context of high
• 4:39
  stakes asset management. Imagine you are an institutional
• 4:42
  investor, the market has just taken a massive
• 4:44
  unexpected dip, and you are calling your asset
• 4:47
  manager to figure out your exposure. Right. And
• 4:50
  in a high stress scenario like that, the last
• 4:52
  thing the client wants is to navigate an automated
• 4:55
  phone tree. Press one for panic. Exactly. Or
• 4:57
  sit on hold while an agent frantically tries
• 4:59
  to pull up their specific portfolio. Right. So
• 5:02
  with agentic AI, the system actually listens
• 5:05
  to the voice of the caller. It recognizes the
• 5:08
  client. analyzes the urgency in their voice or
• 5:11
  the keywords they are using, and instantly routes
• 5:14
  them to the appropriate senior desk. Seamlessly.
• 5:16
  Seamlessly. But it's what happens on the agent's
• 5:19
  side that is truly impressive. Before the human
• 5:22
  agent even answers the line, The agentic AI has
• 5:26
  generated a comprehensive pre -call dashboard.
• 5:29
  It connects all the dots before the conversation
• 5:31
  even begins. Exactly. The agent instantly sees
• 5:34
  the client's total exposure to the current market
• 5:36
  volatility, their recent trade history, and even
• 5:39
  automated suggestions for mitigation strategies
• 5:42
  based on the firm's current macroeconomic stance.
• 5:45
  That's huge. The AI has already done the 20 minutes
• 5:48
  of frantic background research and the two seconds
• 5:50
  the phone was ringing. And we shouldn't overlook
• 5:52
  what happens after the client hangs up either.
• 5:54
  Right. The agentic AI performs automated proactive
• 5:57
  post call follow ups. It drafts the summary of
• 5:59
  the agreed upon actions. It queues up the necessary
• 6:02
  trade tickets for the agent to approve and address
• 6:04
  an email to the client outlining the next steps.
• 6:07
  It's just a seamless end to end orchestration.
• 6:10
  Completely. That automated follow up is incredible
• 6:13
  for the client experience. But I'm actually more
• 6:16
  interested in what this means for the person
• 6:18
  sitting on the other side of the desk. How is
• 6:21
  this changing the internal worker's day -to -day
• 6:24
  reality? That's the real question. Because the
• 6:26
  goal here shouldn't just be replacing people.
• 6:29
  It should be giving them better tools. And one
• 6:31
  of those tools the panel discussed is the ability
• 6:34
  to essentially chat with your own proprietary
• 6:36
  data. That capability alone. represents a massive
• 6:40
  paradigm shift in corporate environments. Think
• 6:43
  about the traditional latency involved in getting
• 6:46
  an answer to a complex business question. It
• 6:48
  takes forever. Right. If a portfolio manager
• 6:51
  needed to understand a subtle trend buried in
• 6:54
  three years of transaction data, they couldn't
• 6:57
  just find that themselves. They'd have to submit
• 6:59
  a JIRA ticket to a data engineering team. Oh,
• 7:02
  yeah. Wait a week? and hope the SQL query the
• 7:05
  engineer wrote actually captured the nuance of
• 7:07
  the original request. And half the time it doesn't,
• 7:10
  and you have to start the process all over again.
• 7:12
  But now, that same portfolio manager can use
• 7:14
  natural language to directly ask their data a
• 7:17
  question. Under the hood, the AI translates that
• 7:21
  plain English into a complex query, runs it against
• 7:24
  the database, and instantly generates a custom
• 7:26
  bashboard on the fly. You are completely removing
• 7:29
  the latency between having a question and getting
• 7:32
  an actionable insight. Yeah. And we are seeing
• 7:35
  this acceleration outside of the data science
• 7:37
  realm too. The panel brought up a great example
• 7:40
  regarding sales teams and wholesalers out in
• 7:42
  the field. Yes, the wholesaler example was brilliant.
• 7:44
  If you've ever worked in that side of the industry,
• 7:47
  you know the pain of competitor research. It's
• 7:49
  brutal. You're traveling, meeting with institutional
• 7:52
  clients, and you have to be prepared to explain
• 7:54
  exactly why your firm's product is better than
• 7:57
  a rival's. Historically, that meant digging through
• 8:00
  hundreds, sometimes thousands of pages of incredibly
• 8:03
  dense compa - I always joke that this specific
• 8:06
  kind of tedious manual research used to be a
• 8:09
  rite of passage for junior analysts, fueled entirely
• 8:12
  by bad office coffee and sheer willpower. It
• 8:15
  is incredibly grueling work. You are looking
• 8:18
  for very specific details, subtle differences
• 8:21
  in risk profiles, fee structures, or investment
• 8:25
  restrictions, all buried deep within legal text.
• 8:28
  But armed with this new AI, those teams in the
• 8:31
  field can just query the system directly from
• 8:33
  their phone or tablet. They can ask a highly
• 8:36
  specific complex question about a competitor's
• 8:39
  newly updated policy, and the AI synthesizes
• 8:42
  the answer for them right there, in real time,
• 8:45
  citing the exact page of the prospectus it pulled
• 8:48
  from. It gives them the answer when they actually
• 8:50
  need it, right before they walk into the client
• 8:52
  meeting. And it isn't just data retrieval, it
• 8:54
  is also content creation. The panelists from
• 8:58
  PGM Global Services shared how their marketing
• 9:00
  and comms teams are utilizing this. They're taking
• 9:03
  their past thought leadership materials, white
• 9:05
  papers, event copy, and feeding it into an AI
• 9:08
  so the model learns their highly specific proprietary
• 9:11
  brand voice. They mentioned it took their standard
• 9:13
  campaign creation time from 14 days down to six
• 9:16
  days. Which is a massive jump in efficiency.
• 9:18
  Massive. But my immediate question when I read
• 9:21
  that was, does it actually capture the nuance
• 9:23
  of a premier financial firm, or does it sound
• 9:26
  like a generic robot wrote it? That is the crucial
• 9:29
  distinction. Because they are training it specifically
• 9:33
  on their own curated historical data, the output
• 9:37
  isn't generic. It incorporates their specific
• 9:40
  terminology, their required compliance disclaimers,
• 9:43
  and their house style. OK. The human editors
• 9:46
  are then spending their time refining a highly
• 9:49
  competent first draft rather than staring at
• 9:52
  a blank page. What's fascinating here is how
• 9:54
  this completely reframes the mainstream narrative
• 9:56
  we often hear about AI coming to eliminate everyone's
• 10:00
  jobs. When you look at these enterprise use cases,
• 10:02
  it isn't about human replacement. It is about
• 10:04
  human augmentation. It removed the absolute drudgery
• 10:08
  of the day -to -day workflow. It allows people
• 10:10
  to actually perform the strategic, creative,
• 10:13
  and interpersonal relationship building they
• 10:14
  were originally hired to do. But of course, to
• 10:17
  do any of this effectively, whether we're talking
• 10:19
  about call center dashboards, querying databases,
• 10:21
  or generating thought leadership, we have to
• 10:23
  talk about the underlying engine. And that is
• 10:26
  the data itself. There is a very old, very true
• 10:29
  saying in computer science, garbage in, garbage
• 10:31
  out. Always true. And when you are dealing with
• 10:34
  high stakes finance, you absolutely cannot afford
• 10:36
  for your AI to be making decisions based on unverified,
• 10:41
  noisy public information. It is the single biggest
• 10:44
  risk factor in enterprise AI adoption. An AI
• 10:47
  model is only as accurate and unbiased as the
• 10:50
  data it is trained on. This is why the panelists
• 10:52
  from PGM heavily emphasize a strategy they call
• 10:56
  the walled garden approach. I love that term
• 10:58
  walled garden. Explain the mechanics of that
• 11:01
  for us because it sounds like they are basically
• 11:02
  fencing the AI off from the rest of the world.
• 11:04
  That is essentially what it is. Instead of allowing
• 11:07
  their AI to roam free across the open Internet,
• 11:10
  which is terrifying, which is filled with unverified
• 11:13
  claims, outdated regulations and inherent biases.
• 11:16
  Instead of that, these firms deploy their LLMs
• 11:19
  within a strict internal perimeter. The AI is
• 11:21
  grounded only in curated proprietary data sets.
• 11:25
  It only reads the firm's vetted internal knowledge
• 11:27
  bases, their secure data lakes, and their approved
• 11:29
  policy documents. So it doesn't know what people
• 11:32
  are saying on Reddit. It only knows what the
• 11:35
  firm's chief investment officer has officially
• 11:38
  published. Precisely. And establishing that Waldgarten
• 11:40
  matters immensely. First and foremost, it drastically
• 11:44
  reduces the risk of AI hallucinations. If the
• 11:48
  model doesn't know the answer, It is constrained
• 11:50
  to say, I don't know, rather than confidently
• 11:53
  inventing a false statistic. Which we've all
• 11:56
  seen happen. Exactly. Furthermore, this walled
• 11:59
  garden creates a secure sandbox for the employees.
• 12:02
  It gives them a protected environment to experiment,
• 12:05
  to learn how to properly prompt the system and
• 12:07
  to build their digital literacy without any risk
• 12:10
  of accidentally leaking sensitive internal data
• 12:12
  to a public model. But setting up that walled
• 12:15
  garden isn't just a matter of buying a software
• 12:17
  license. No. The panel really stressed that this
• 12:20
  requires a massive culture shift. There has to
• 12:23
  be an incredibly tight ongoing partnership between
• 12:25
  the business teams and the technology teams.
• 12:28
  If tech operates in a silo, you run into that
• 12:31
  classic pitfall where engineers build a highly
• 12:33
  sophisticated, shiny new tool, they hand it over
• 12:36
  to the portfolio managers, and no one uses it
• 12:39
  because it doesn't actually solve a real business
• 12:41
  problem. We see that happen constantly. The technology
• 12:44
  must be driven by actual business use cases,
• 12:48
  not just the desire to deploy AI for the sake
• 12:51
  of deploying AI. And that requires comprehensive
• 12:54
  enterprise -wide training from the executive
• 12:57
  board all the way down to the junior analysts.
• 13:00
  Everyone needs a baseline understanding of how
• 13:03
  these generative models function, how to evaluate
• 13:05
  their outputs, and most importantly, what their
• 13:08
  limitations are. Which brings us to perhaps the
• 13:10
  most critical part of this entire deep dive,
• 13:13
  the security aspect. Here we go. Because everyone
• 13:15
  in this industry is moving fast. The security
• 13:17
  leader from TNC Bank used a really memorable
• 13:19
  analogy for this. They said that in the current
• 13:22
  market, everyone wants to drive a fast car. Everyone
• 13:24
  wants to adopt AI quickly to stay competitive.
• 13:27
  But to drive a fast car safely, you need really,
• 13:30
  really good brakes. The guardrails have to be
• 13:32
  established before you step on the gas. Here's
• 13:34
  where it gets really interesting. Because the
• 13:37
  consequences of driving this fast car without
• 13:39
  breaks are not theoretical, they are already
• 13:41
  playing out. The sources detailed some severe
• 13:45
  incidents from 2024, where companies, suffering
• 13:48
  from a fear of missing out, allowed employees
• 13:50
  to use publicly accessible Gen. AI chatbots for
• 13:53
  financial analysis without implementing proper
• 13:56
  data loss prevention controls. It is a nightmare
• 13:58
  scenario for any chief information security officer.
• 14:01
  It truly is. You had well -meaning employees
• 14:04
  taking sensitive client portfolios, including
• 14:06
  massive amounts of personally identifiable information,
• 14:09
  and pasting that raw data directly into public
• 14:12
  chatbots just to format a report. Just trying
• 14:15
  to save time. Yeah. They didn't realize that
• 14:17
  by doing so, they were feeding their clients
• 14:19
  private data into the public models training
• 14:21
  set. The result was massive data exposures and
• 14:25
  millions of dollars in regulatory fines, all
• 14:27
  because they bypassed basic internal guardrails.
• 14:30
  And accidental data leakage is only one side
• 14:32
  of the coin. The threat landscape has fundamentally
• 14:35
  changed because malicious actors are also weaponizing
• 14:37
  this technology. Oh, the panel did not hold back
• 14:39
  on this. They highlighted the terrifying reality
• 14:42
  of AI generated phishing and social engineering.
• 14:45
  We are no longer talking about poorly worded
• 14:47
  scam emails full of typos. Right. Attackers are
• 14:51
  using AI to generate hyper realistic deep fakes
• 14:54
  of executives. They will clone a CFO's voice
• 14:57
  and call a junior employee to authorize an urgent
• 14:59
  wire transfer. Or they will execute incredibly
• 15:02
  sophisticated business email compromises that
• 15:05
  perfectly mimic the writing style of internal
• 15:07
  leadership. This is what we refer to as the cybersecurity
• 15:10
  arms race. The adversaries are using AI to scale
• 15:13
  their attacks, automate their vulnerability scanning,
• 15:16
  and generate highly convincing synthetic media.
• 15:19
  So you have to fight fire with fire. Exactly.
• 15:22
  The defending teams have absolutely no choice
• 15:24
  but to utilize AI as well. A security operations
• 15:27
  center at a major asset manager simply cannot
• 15:29
  rely on human analysts to manually sift through
• 15:32
  millions of daily network logs anymore. So how
• 15:35
  are the defenders utilizing AI to actually fight
• 15:37
  back and gain the upper hand? They're using specialized
• 15:40
  internal AI models to process those unimaginable
• 15:45
  volumes of event data at machine speed. The AI
• 15:48
  establishes a baseline of normal network behavior
• 15:50
  and then constantly hunts for anomalies. Like
• 15:53
  finding a needle in a haystack? Faster than any
• 15:55
  human could. Yeah. It helps defenders spot the
• 15:58
  unknown unknowns, the subtle complex attack patterns
• 16:01
  that a human analyst would never catch in time.
• 16:03
  But the panel also offered a stark reminder.
• 16:06
  AI models are, at their core, just software.
• 16:09
  Right. And like any software, they have vulnerabilities.
• 16:12
  The security leaders emphasized that firms must
• 16:14
  continuously stress test these models using established
• 16:17
  cybersecurity frameworks. They specifically mentioned
• 16:19
  leveraging resources like OWAS, which is essentially
• 16:22
  the global standard for application security,
• 16:24
  and OpenAI primers to rigorously hunt for flaws
• 16:27
  in their own walled gardens. It sounds like a
• 16:30
  relentless high stakes battle. But what I found
• 16:32
  really encouraging in the sources is that AI
• 16:34
  isn't just being used to fight off global hackers.
• 16:37
  It is also being aimed at solving one of the
• 16:38
  most massive, tedious pain points in the corporate
• 16:41
  world. I am talking about compliance. Ugh, yes.
• 16:45
  The labyrinth of modern financial regulation.
• 16:47
  The panel called this the dream of killing the
• 16:49
  questionnaire. If you operate anywhere in asset
• 16:52
  management, you know the absolute agony of conflicting
• 16:54
  regulations. You have state privacy laws colliding
• 16:58
  with federal financial regulations, which then
• 17:00
  collide with various industry certifying authorities.
• 17:03
  And it never ends. It doesn't. And typically,
• 17:06
  a firm handles this by deploying armies of compliance
• 17:09
  analysts to do what they call stare and compare
• 17:11
  work. They sit there. manually cross -referencing
• 17:15
  hundreds of spreadsheet cells, trying to figure
• 17:18
  out how to satisfy a 50 -page vendor risk assessment.
• 17:21
  It is a massive drain on resources and it is
• 17:24
  highly prone to human error. But the experts
• 17:26
  on our panel detailed how agentic AI is actively
• 17:29
  being used to harmonize these conflicting frameworks.
• 17:33
  The AI can ingest a massive new federal regulation,
• 17:36
  cross -reference it against the firm's existing
• 17:38
  internal controls, and instantly highlight the
• 17:41
  gaps. It maps the requirements across different
• 17:44
  jurisdictions to create a single rationalized
• 17:46
  set of controls. The ultimate goal these firms
• 17:50
  are driving toward is automating compliance detection
• 17:52
  entirely. They want the AI to continuously monitor
• 17:56
  the firm's systems and output the required reporting
• 17:58
  automatically, effectively cutting out the middleman
• 18:01
  and killing the endless compliance questionnaires
• 18:04
  once and for all. That would change the industry
• 18:06
  overnight. So what does this all mean? We have
• 18:09
  covered incredible ground today. Looking at everything
• 18:12
  these leaders from MetLife, Capgemini, PNC, and
• 18:15
  PGM have laid out, it is abundantly clear that
• 18:18
  AI in asset management is no longer an experimental
• 18:22
  toy. Not at all. It is a top to bottom restructuring
• 18:25
  of the financial plumbing. We are looking at
• 18:27
  a future powered by agentic assistance capable
• 18:30
  of eradicating the bulk of manual friction. But
• 18:33
  it is a future that completely hinges on pristine
• 18:36
  data governance through walled gardens and ironclad,
• 18:39
  thoroughly tested security guardrails. If we
• 18:41
  step back and view this in a broader historical
• 18:44
  context, we are squarely in the middle of the
• 18:46
  Fourth Industrial Revolution. And what this deep
• 18:49
  dive truly underscores is the pivot in the value
• 18:52
  of human capital. For decades, the financial
• 18:55
  industry rewarded the manual gathering, sorting,
• 18:57
  and verifying of data. Right. That era is ending.
• 19:01
  The true value of the human worker is moving
• 19:03
  entirely toward interpreting the profound, synthesized
• 19:06
  insights that these machines deliver. The machine
• 19:10
  does the heavy lifting of data aggregation, while
• 19:13
  the human provides the strategic judgment and
• 19:15
  the ethical oversight. It's a total shift in
• 19:17
  skill sets. It is. But as we look ahead, this
• 19:20
  raises a profound question. Something I want
• 19:23
  you, the listener, to consider long after this
• 19:26
  deep dive ends. Let's fast forward five or ten
• 19:28
  years. If agentic AI successfully automates the
• 19:31
  vast majority of manual analysis, if it flawlessly
• 19:34
  harmonizes global compliance, and if it seamlessly
• 19:37
  manages our client interactions, how will firms
• 19:39
  actually differentiate themselves in the marketplace?
• 19:41
  That's a great point. When every single competitor
• 19:44
  has purchased the exact same hyper -efficient
• 19:46
  AI capabilities, technological speed is no longer
• 19:49
  an advantage. It is just the baseline. In a fully
• 19:53
  automated, flawlessly efficient financial world,
• 19:56
  will deeply held human intuition, empathy, and
• 19:59
  the art of creative relationship building become
• 20:01
  the ultimate scarce premium asset? That is the
• 20:04
  ultimate question we are all going to have to
• 20:06
  answer in the next few years. Thank you so much
• 20:08
  for joining us on this deep dive. We hope this
• 20:10
  exploration helped you cut through the hype and
• 20:12
  grab hold of the real mechanics shaping the future
• 20:14
  of finance. Keep learning, keep questioning the
• 20:17
  information landscape around you, and we will
• 20:19
  catch you on the next one.