Latest / In The Loop / EP19 - Mary Meeker AI Trends 2025: Three Reasons Why AI Is Different Than Any Other Tech In History

Transcript

EP19 - Mary Meeker AI Trends 2025: Three Reasons Why AI Is Different Than Any Other Tech In History · In The Loop

• 0:00
  After a six year break, the very famous analyst Mary Mika is back
• 0:03
  with her first. Trends.
• 0:04
  Report since 2:00. 1019 and this time.
• 0:08
  It's entirely focused on guess what?
• 0:10
  AI, all 365 pages of it. For those that are unfamiliar
• 0:15
  with Mary. Mika.
• 0:16
  She was one of the. OG tech influencers.
• 0:19
  In the 90s. Who was at Wall?
• 0:21
  Street. And an analyst?
• 0:22
  Who produced extensive. Reports on everything.
• 0:25
  About the Internet. It became.
• 0:26
  The must read document. Anybody that's trying to
• 0:29
  understand where technology? Is heading.
• 0:31
  When she. Released her.
• 0:31
  Reports. They weren't just documents or
• 0:33
  slides. They were cultural moments that
• 0:35
  made everybody in the industry. Stop now after a long.
• 0:38
  Six years of. Silence.
• 0:39
  She's back. In today's episode, I'm going
• 0:41
  to. Break down three of the main.
• 0:43
  Takeaways that explain. Why AI is?
• 0:45
  Different than. Any of the.
• 0:46
  Technology trend that's. Ever come before it taken?
• 0:49
  From her 365 page report and. I'm going to give you an honest
• 0:52
  take about. What I think?
• 0:53
  This shows about the state of AI.
• 0:55
  Understanding today. This is in the loop.
• 0:58
  With Jack Horton. Hope you enjoy the show.
• 1:11
  OK, so let's. Dig into.
• 1:12
  Today's episode as. I said there's 300 and. 65 pages
• 1:16
  released. By Mary Meeker and her.
• 1:17
  Team and I think there's three or four key things I want to.
• 1:21
  Discuss. That really tells us about why
• 1:23
  AI is just. So different than any other.
• 1:25
  Technology trend that's. Come before it.
• 1:27
  I think the first reason and the first theme let's.
• 1:30
  Say is speed. So that's come through really
• 1:34
  strongly in her. Report and it's one of the.
• 1:36
  Key ways that AI. Differs from any past.
• 1:39
  Technology trend in. History she's.
• 1:41
  Explicitly drawn a lot of. Comparisons, of course.
• 1:43
  Between previous. Platform shifts so.
• 1:45
  She talked about. Vincent serves. 1999.
• 1:48
  Observation that a year in the Internet business is.
• 1:51
  Like a whole. Dog Year.
• 1:53
  And then noted that a is. Pace is really.
• 1:55
  Just. 10 times more. Extreme technology compounds and
• 1:58
  AI is building on everything that came before it now.
• 2:02
  In today's episode, we. Will go through a number of.
• 2:03
  Charts and stats that are in a report there's. 360. 5 pages of
• 2:07
  it and I. Picked.
• 2:07
  Out what I think is. The most interesting, but.
• 2:10
  If you're just. Listening on the audio version
• 2:11
  and you're wondering? What some of these?
• 2:13
  Stats and charts pertain to if. You go to the Spotify.
• 2:16
  Video or the YouTube video? Or even the.
• 2:19
  Article that's. Produced for every single
• 2:20
  episode. Every single chart that's
• 2:22
  discussed. Will be included in that.
• 2:25
  So since. 2010 one of the. Key starts you've actually
• 2:28
  highlighted was that there's been a. 260%.
• 2:31
  Annual growth rate. In the.
• 2:32
  Use of data. To train new AI models and a
• 2:35
  three. 160. Percent.
• 2:36
  Increase in computing power. Which has led to an over. 167%.
• 2:41
  Growth. Over the last four years alone.
• 2:44
  In the number of AI models that have been produced, which as you
• 2:46
  can imagine, is driving just. Unprecedented uptake across
• 2:50
  consumers and business. And the.
• 2:51
  Speed of this. Adoption across different
• 2:53
  technological eras. Tells us.
• 2:55
  A very powerful story here because the PC took 20.
• 2:58
  Years to reach. 50% adoption in. US households.
• 3:02
  The desktop Internet. ERA took 12 years to reach. 50%
• 3:05
  of. People.
• 3:06
  The mobile took 6. But when it comes to.
• 3:09
  The AI era. It's looking.
• 3:10
  Like it's been about three years to reach. 50% of all.
• 3:14
  Households using it. So each.
• 3:15
  Cycle of new technology is roughly.
• 3:17
  Half the cycle of the previous one.
• 3:19
  And you can really understand this when we think.
• 3:20
  About ChatGPT which? Just recently passed 300 and. 65
• 3:24
  billion annual searches, which is 5 1/2.
• 3:27
  Times, by the way. Faster than Google?
• 3:29
  Reached 300 and. 65 billion searches because.
• 3:32
  If you're not using it yet. The search feature.
• 3:34
  Inside chat. GBT is just.
• 3:35
  Brilliant and. Another theme that's consistent
• 3:39
  throughout the. Mary Mika report is that ChatGPT
• 3:41
  is just. Utterly dominant.
• 3:43
  The growth of that. Product has been unprecedented.
• 3:46
  It's currently the AI product to beat by a mile.
• 3:49
  You know the. Report shows chat.
• 3:50
  GBT subscriber and revenue. Growth being just exponential.
• 3:53
  It's reached 800. Million active.
• 3:55
  Users in the report and 20 million paid subscribers and
• 3:58
  they're nearly hitting 4 billion revenue and they've done that in
• 4:01
  just a few. Years and.
• 4:02
  What's really interesting about this growth?
• 4:03
  Rate and. Adoption in households.
• 4:05
  And amongst ChatGPT and others. Is just the.
• 4:07
  Demographic usage, actually. That's why most.
• 4:10
  Users are aged. 18 to 4920% of people, according to this
• 4:14
  report. Are age 65 and.
• 4:15
  Older and have claimed to. Have used church.
• 4:17
  GBT at least once. So this isn't just a young
• 4:19
  person technology. You know all the generations are
• 4:23
  using this as a Google replacement.
• 4:24
  And younger people are almost treating this.
• 4:26
  As a life advisor. We discussed this.
• 4:28
  In a recent. Episode just a couple of weeks.
• 4:30
  Back how chat. GBT's most.
• 4:32
  Popular use case now. And a IS.
• 4:34
  Property use case. Has actually become.
• 4:36
  Therapy I for. One use it as.
• 4:37
  A life coach. Almost every single day.
• 4:39
  Whether at work or at home. And when we.
• 4:41
  Think about the theme of speed. It's not just.
• 4:42
  The speed of. Adoption.
• 4:44
  It's also retention. Because daily usage.
• 4:47
  On Just ChatGPT alone has grown from less than 10 minutes a day
• 4:50
  to 20 minutes a day. In just.
• 4:52
  The last 18 months. So what's happening is a
• 4:54
  fundamental change in the. Way we interact with technology.
• 4:56
  We're moving from. Cooking and typing to a.
• 4:59
  Conversational interface. And this shift could even drive.
• 5:02
  People away from. Traditional.
• 5:03
  Platforms so social. Media or Netflix?
• 5:05
  Or YouTube. And as I've.
• 5:07
  Said before every. One of these.
• 5:08
  Companies wants to. Be.
• 5:09
  Essentially the chatbot interface that rules them all.
• 5:12
  Because every single. Interface, as it becomes
• 5:15
  conversational will connect to other.
• 5:17
  Conversational interfaces. So a person only has.
• 5:19
  To use. One platform to connect with
• 5:21
  many other platforms and go further with this theme of speed
• 5:24
  in the report by Mary Meeker when we talk.
• 5:27
  About developers and as a. Proxy to show.
• 5:29
  Developer involvement and the number of people.
• 5:31
  Building in this AI world now. In the.
• 5:35
  Google ecosystem developer. Ecosystem, that is.
• 5:38
  In the last 12 months has grown from 1.4 million developers to 7
• 5:41
  million. Developers in just as I said, a
• 5:44
  year. That's a 5X increase.
• 5:46
  Another area that the. Theme of speed.
• 5:47
  Applies to. Is also AI.
• 5:49
  Performance. Which is just.
• 5:50
  Accelerating at breakneck. Speed.
• 5:53
  AI models are becoming. So sophisticated that as of.
• 5:56
  February this year. ChatGPT. 4.5.
• 5:59
  Are consistently fueling. Humans into thinking.
• 6:02
  They're chatting with another human.
• 6:05
  Previously the win rate for that.
• 6:07
  Test was over. 70% so 70% of the time people would identify this
• 6:12
  as an AI system and in parallel the report.
• 6:14
  Also. Discusses the speed of
• 6:15
  investment as well. So the Big 6.
• 6:17
  Tech companies alone, so the. Googles the Facebooks.
• 6:19
  The Amazons. Has spent 212 billion in 2024
• 6:23
  alone, which is up. 63%. Year on year.
• 6:26
  Which is far. Above.
• 6:27
  Any historical norm of spending? This brings on.
• 6:30
  To a very interesting point about AI that the.
• 6:32
  Pod does a great job of. Showcasing which is.
• 6:35
  The reduction inference cost and the.
• 6:36
  Speed of that reduction as well. So yes.
• 6:39
  People are spending more time on AI.
• 6:41
  Companies are spending. More on.
• 6:42
  Building AI. Whilst at the.
• 6:44
  Same time. Inference cost so.
• 6:46
  The cost to run a language model on servers.
• 6:49
  Has dropped by 99. Percent over the last two years.
• 6:52
  So what used? To cost pounds or dollars now.
• 6:54
  Cost maybe pennies. And what cost pennies today
• 6:57
  might even cost. A fraction of a penny.
• 6:59
  In the future. To understand really the speed
• 7:01
  of this reduction it took. 80 years for the light bulb to.
• 7:04
  Become that cheap to use. Versus a year.
• 7:07
  For some AI models to. Achieve a similar cost
• 7:09
  reduction. This cost?
• 7:10
  Reduction is. Only going to.
• 7:11
  Democratize AI access. Further.
• 7:13
  Which, as you can imagine, is only going to.
• 7:15
  Increase its usage and adoption. Which will only.
• 7:18
  Increase the amount. People will then spend on it.
• 7:20
  And another. Area of speed.
• 7:21
  That the report. Highlighted was.
• 7:23
  The adoption of. This technology and enterprise.
• 7:26
  Level. Because this isn't just a
• 7:27
  consumer. Technology trend that's.
• 7:29
  Trickled all the way. Up to enterprise.
• 7:31
  Everyone's been jumping in at once.
• 7:33
  The proportion. Of S&P 500 companies mentioning
• 7:37
  AI during their earning. Calls grew from. 10% when
• 7:40
  chapter. BT was first released.
• 7:42
  To now 50% and a 2024 Morgan Stanley survey found that 75.
• 7:46
  Percent of global CMOS. Were already.
• 7:48
  Running AI. Tests with virtually.
• 7:51
  Everyone else planning to start? Within 12 months.
• 7:53
  And that was 2024. We're now a.
• 7:55
  Year ahead of that. Again, there's an element.
• 7:57
  Of this report because it's 3. 160. 5 pages long and it's so
• 8:00
  extensive. It feels almost outdated.
• 8:03
  If you're listening to this podcast, every single.
• 8:05
  Week you'll. Know about the growth of
• 8:08
  adoption? And usage in companies.
• 8:11
  You understand? How fast?
• 8:12
  Things are really changing. OK, so I'm.
• 8:25
  Now going to move on from that big.
• 8:26
  Theme of speed and all the data and the statistics around speed.
• 8:29
  As well. To something else that was.
• 8:31
  Becoming very apparent. Inside the report which is that?
• 8:34
  AI has been global from day one and this is another real.
• 8:39
  Big way that. It differs.
• 8:40
  From any immediate. Or previous technology.
• 8:43
  Because previous. Technology trends started maybe
• 8:45
  in the US, and slowly. Spread worldwide.
• 8:48
  You know the Internet took 20. Three years.
• 8:50
  For 90. Percent of users.
• 8:52
  To be outside of North America and with chat.
• 8:54
  GBT that just took three years. The top.
• 8:57
  Countries in ChatGPT now isn't the US, it's.
• 8:59
  India, which represents. 13.5% of total.
• 9:02
  Usage compared to the US. Is 8. .9.
• 9:05
  And crazily after that. Comes.
• 9:06
  Indonesia and Brazil. Both with over 5%.
• 9:10
  Followed by Egypt. With nearly 4.
• 9:11
  Percent of total usage. And as a.
• 9:13
  Result it's been. Geopolitical.
• 9:14
  From day one too. This isn't a technology that
• 9:17
  developed in a vacuum and. Then became geopolitically
• 9:20
  relevant. The US and China have been in
• 9:22
  this AI competition from day one, and the report highlights
• 9:25
  the fact that there is an. Arms race going on.
• 9:27
  And the reality? I quote the reality is AI
• 9:30
  leadership. Could be get geopolitical
• 9:32
  leadership and not vice versa essentially.
• 9:34
  The nation or company that leads AI.
• 9:37
  Is going to. Have the strategic.
• 9:39
  Edge on the geopolitical stage. And the report?
• 9:41
  Backs us. Up by comparing.
• 9:42
  Progress across countries. For example, China.
• 9:45
  Is portrayed as. Surging ahead in certain areas
• 9:48
  by 2025, China's. Released three major.
• 9:51
  Large. Scale open source models, which
• 9:53
  is more than any other country. The.
• 9:54
  Chinese language model. DeepSeek, which we've.
• 9:56
  Discussed on this episode went. From zero to. 21% of global LLM.
• 10:00
  Users in literally just a few months.
• 10:03
  Highlighting how? Quickly, China is.
• 10:05
  Mobilizing that massive. User base.
• 10:07
  And she also noted that the open source model.
• 10:10
  Produced by Alibaba. Called.
• 10:11
  Quen 2.5 is outperforming many Western models.
• 10:15
  And this battle really comes through.
• 10:17
  In the data because the total number of large.
• 10:19
  Scale AI systems in both the. US and China.
• 10:22
  Literally dwarfs the rest. Of the world.
• 10:25
  There is more. Large scale AI systems deployed.
• 10:27
  In both. The US and China versus the
• 10:30
  entire rest. Of the world combined, but for.
• 10:32
  China, they've got a huge. Advantage when it comes.
• 10:34
  To embodied AI. So this is.
• 10:36
  AI inside of. Robots.
• 10:37
  So China has more. Robots in industry installed
• 10:41
  than the. Rest.
• 10:41
  Of the world combined. With the US.
• 10:43
  Lagging you could. Even say.
• 10:44
  Embarrassingly far behind. The third major.
• 10:58
  Theme. I think that shows that this
• 10:59
  technology is so different than any other technology.
• 11:02
  Before it is. Just the breadth of.
• 11:03
  Disruption. Mary Mika.
• 11:04
  Isn't just saying? Things are changing faster,
• 11:06
  she's saying that. Everything is changing.
• 11:08
  In the. Report I tried to predict.
• 11:10
  Loads of different areas that AI will make an impact and to be
• 11:12
  honest it. Talked about coding.
• 11:14
  And. Research.
• 11:15
  And logistics which? To me just felt outdated.
• 11:18
  Because I'm already seeing these systems.
• 11:20
  Deployed. We're also seeing.
• 11:21
  New AI first. Companies just.
• 11:23
  Grow at unprecedented. Rates like livable.
• 11:26
  He was featured. In the report as.
• 11:28
  Growing their recurring revenue from pretty much 0.
• 11:30
  In 2020. 4 to 50 million. Six months later, these aren't.
• 11:34
  Small improvements these are. Exponential.
• 11:37
  Leaps. And throughout the report.
• 11:38
  Really, Mika? Highlights the.
• 11:40
  Fact that the. Competitive dynamics have never.
• 11:42
  Been like this. Before for example.
• 11:43
  One chart showed. Autonomous vehicles taking over
• 11:46
  Lyft. And even.
• 11:47
  Uber in certain. Areas in the US this isn't.
• 11:50
  Just this sci-fi. Experiment.
• 11:51
  It's the. Reality in San Francisco.
• 11:55
  In a real. Major city where?
• 11:56
  Driverless cars are. Becoming the.
• 11:58
  Go to. Form of transport.
• 12:00
  So yeah, the breadth of disruption across.
• 12:01
  Every single industry is. Unprecedented and when?
• 12:04
  Combined with the. Speed.
• 12:06
  And the fact that this. Is so global.
• 12:08
  You can. See why people are both?
• 12:10
  Excited and scared. But I thought I'd I'd.
• 12:12
  Give an honest take on all of this.
• 12:15
  Looking at. Mika's very.
• 12:16
  Comprehensive report. I think she nails this macro 30.
• 12:19
  1000 foot story. But I think there's a big gap
• 12:22
  between this 30,000 foot. Analysis and.
• 12:25
  What's actually happening? On the ground.
• 12:27
  When it. Comes to trying.
• 12:28
  To. Extract.
• 12:29
  Value from AI. So I'll try and walk you through
• 12:31
  some. Of the.
• 12:32
  Things that I. See as disconnects.
• 12:33
  The capital. Deployment story for example.
• 12:35
  Is quite fascinating. Mika documents big technology
• 12:39
  companies. Spending, you know, 200 and. 12
• 12:41
  billion on AI. In 2024 alone, which is.
• 12:43
  Just staggering by quite. Literally any measure.
• 12:46
  But when you actually. Follow the money.
• 12:47
  Trail so. Where people are actually making
• 12:49
  money from AI. All the way.
• 12:51
  Back to the. PNL.
• 12:52
  In part the. Results are very.
• 12:53
  Uneven the winds are. Very real, but they're very
• 12:56
  specific rather than very broad I.
• 12:58
  Did an entire episode recently on the types of AI features.
• 13:01
  Being released. So take strikes.
• 13:03
  Fraud detection system. They have this.
• 13:05
  AI system that analysis. Payment patterns and has.
• 13:08
  Helped jump their. Catch rates the number of times
• 13:11
  they catch fraud. From 57.
• 13:12
  Percent to 97%. Pretty much overnight.
• 13:16
  That's not a. Small improvement?
• 13:17
  That's a. Huge change and that.
• 13:19
  Directly impacts. The bottom line?
• 13:20
  And. What's very noticeable?
• 13:22
  About all. The stories I've covered.
• 13:23
  When it comes. To AI features being released,
• 13:25
  is that B to B focused? And they're typically
• 13:27
  integrated. Into a very.
• 13:28
  Specific industry or customer workflow and they're solving
• 13:32
  very. Particular.
• 13:33
  Problems with a very clear, measurable outcome.
• 13:35
  They aren't these general purpose consumer AI.
• 13:38
  Assistants trying to. Be everything to.
• 13:40
  Everyone. That purpose, Bill.
• 13:42
  They're designed to slide into existing processes and platforms
• 13:45
  where the. Value proposition.
• 13:47
  Is just obvious. And quantifiable and.
• 13:49
  If we compare this to say. Consumer AI.
• 13:51
  Two years after ChatGPT or three years?
• 13:53
  After ChatGPT came out. What are we still left with?
• 13:57
  It's ChatGPT still. And ChatGPT.
• 13:59
  Plus, for those that want to pay $200 a month.
• 14:02
  Despite thousands of. Startups and billions being
• 14:05
  invested in these companies. That said, they're going to be
• 14:08
  the. Consumer application.
• 14:09
  For everything. Nobody.
• 14:11
  'S been that. Successful.
• 14:13
  They've really. Struggled to find the.
• 14:15
  Sustainable footing. Perplexity is.
• 14:17
  Probably one of the few that. Have done it and I think this
• 14:19
  really explains I guess a. Broader issue with how people.
• 14:22
  Interpret AI capabilities. You know every.
• 14:25
  Quarter, there's this new. Big headline that.
• 14:28
  This new language model achieves human level performance.
• 14:30
  X or Y task. And the benchmark?
• 14:33
  Improvements. Are very impressive and they're
• 14:35
  genuine. They're real.
• 14:37
  And the models are getting so much better.
• 14:38
  So quickly, but there's this. Huge gap between.
• 14:42
  This benchmark. Result.
• 14:44
  And real world. Business value or.
• 14:46
  Personal value that solves a real problem.
• 14:49
  Benchmarks are. Essentially tests with.
• 14:52
  Very, very clean data and very. Clear objectives, but.
• 14:55
  Businesses have very messy systems and legacy.
• 14:58
  Constraints and operational. Friction with annoying.
• 15:01
  And frustrated people. In the process that have to.
• 15:03
  Implement things so. If I make this a little.
• 15:06
  Bit clearer for. Programmers.
• 15:07
  If you're. Using this AI.
• 15:09
  Programming tool for. A brand new.
• 15:10
  Product. Or that this perfect code?
• 15:12
  Base. That's very.
• 15:13
  Small with very little errors. These AI tools.
• 15:16
  Will do. A great job and increase.
• 15:17
  Productivity you know about? 10 times probably.
• 15:20
  But use this. Tool on a multi million line
• 15:23
  enterprise. Code base it.
• 15:25
  Will struggle. The magic won't be as magical
• 15:29
  because that. Raw model power.
• 15:31
  Hits a bit. Of a wall.
• 15:33
  It will make. Mistakes.
• 15:34
  It can't. Identify all the dependencies.
• 15:36
  And you can. Repeat this example.
• 15:38
  Across almost all. Work domains.
• 15:40
  Feed. One of these Incredible.
• 15:41
  Language models a perfectly formatted PDF and it will.
• 15:44
  Probably extract. Almost all the.
• 15:46
  Information perfectly but. Hand her a.
• 15:48
  Folder of. Chaotic, messy documents.
• 15:50
  With random. File types and.
• 15:52
  Really poorly scanned images and it will break down quite
• 15:55
  quickly. And the obsession with AI agents
• 15:57
  is another. Way this is illustrated.
• 15:59
  Perfectly. You know Google.
• 16:01
  Searches for AI. Agents, as highlighted in the
• 16:03
  report, has increased by 1000. Percent in early 2025.
• 16:06
  So obviously there's a clear. Vision of the future that people
• 16:10
  are having and they see agents as key to that.
• 16:12
  And I do too. But this?
• 16:13
  Obsession with autonomous systems that can.
• 16:15
  Plan. And reason and execute.
• 16:17
  Multi step workflows without. Any oversight or involvement
• 16:20
  from a human? Is.
• 16:22
  Well into the future point one of these models.
• 16:25
  At a business. Process and.
• 16:26
  Tell it to autonomously orchestrate and.
• 16:28
  Decide everything. It will often not work and.
• 16:33
  It'll only really work in a very constrained and.
• 16:35
  Specific scenario, but for. Many use.
• 16:39
  Cases I'd argue the most successful.
• 16:40
  Deployments will. Share a few critical
• 16:42
  characteristics. So the scope is quite tight.
• 16:46
  Second, the inputs are going. To be quite predictable.
• 16:48
  So it might be a. Recording from.
• 16:51
  Conversations or. It could be.
• 16:53
  A customer support. Conversation.
• 16:55
  And 3rd, the risk of any mistakes by the agent or the AI
• 16:57
  is very well managed. So there's some.
• 16:59
  Form of human. Oversight or ability?
• 17:01
  To. Order and quickly change.
• 17:02
  Things the moment you step outside.
• 17:04
  Any of this carefully. Constricted I guess.
• 17:06
  Comfort zones. Agents are going to become
• 17:08
  exponentially. More fragile because you're
• 17:11
  going to have. To have.
• 17:11
  Complex memory, that is. Remembering the correct.
• 17:15
  Things across many different interactions.
• 17:17
  You have to manage multi tool orchestration and despite.
• 17:20
  Lots of cool looking videos a. Production agent system still
• 17:23
  requires human operators. We call them AI managers.
• 17:27
  Who can look at errors or? Look at the.
• 17:29
  Data. And make improvements sometimes
• 17:31
  at. 345. AM in the morning to fix
• 17:33
  something that's. Just not working.
• 17:34
  So I guess what I'm saying is. The data doesn't.
• 17:38
  Tell the whole story. Because it's not.
• 17:40
  Just about. Technology curves or.
• 17:42
  Benchmarks. It's about the.
• 17:44
  Implementation. Specifics that makes a
• 17:46
  difference between. Cool looking videos and
• 17:48
  something that's reliable. And used every single day.
• 17:51
  And for those. That are.
• 17:51
  Already at technology. Companies or a builders or
• 17:54
  thinkers and innovators or business leaders.
• 17:57
  All I'd say is when AI. Capabilities are just.
• 17:59
  Becoming increasingly commoditized and competitive.
• 18:02
  Features are going. To be replicated in maybe weeks,
• 18:04
  not even months or years. The only competitive.
• 18:08
  Advantage you've got is to move. Faster have direct.
• 18:10
  Relationships. With budget holders and decision
• 18:12
  makers. Structure your.
• 18:14
  Contract renewal process. Really carefully make it.
• 18:17
  Really. Difficult to switch.
• 18:19
  From you, develop a brand that people love.
• 18:21
  And as you'll notice, these aren't.
• 18:23
  Magical AI Futuristic. New things that have come into
• 18:27
  play. These are quite.
• 18:28
  Traditional business strategies. Anyway, I hope that was helpful.
• 18:31
  That was a bit of a chunky episode of lots of.
• 18:33
  Information and data. Please do.
• 18:34
  Share This. To a friend in a Slack channel
• 18:37
  and Microsoft Teams channel. Give it a.
• 18:39
  Share online. It's growing quite.
• 18:41
  Fast and it's all through. Word of mouth, so I'll see you.
• 18:44
  Next week.