The Data Mix - S3 Ep2 - Rosanne Werner

The Data Mix - Season 3 Episode 2 - Rosanne Werner

And just like that, we’re back with Season 3, Episode 2 of #thedatamix.

This time, we’re joined by Rosanne Werner, CEO at XcelerateIQ and former Data and Digital Strategy Lead at Coca-Cola, for a candid discussion on the intersection of data, AI, and human behaviour. Rosanne shares insights on:

- Why companies struggle with adoption despite investments in AI and data.
- How neuroscience and behavioural science can improve data literacy.
- The role of storytelling in making data memorable and actionable.
- Bridging the gap between technical teams and people-focused strategies.

Plus, don’t miss The Data Fix, our new segment with sharp analysis on the latest tech developments and their data implications.

Join us on Tuesday, December 17th, 2024, at 3:00 PM GMT / 10:00 AM EST for a thoughtful look at the way habits, behaviours, mindset and culture affect our Data and AI Journey.

Chapters

• 0:00 Introduction
• 2:38 The Data Fix - SORA and AI Video Generation
• 10:15 Embracing Incremental Change
• 11:00 Rosanne Introduction and Background
• 14:05 Coca-Cola's Christmas Culture
• 20:30 Coca-Cola’s Data Literacy Journey
• 25:00 Common Shortcomings in Data Literacy Programs
• 30:25 Data Literacy Programs and Their Effectiveness
• 34:33 Cross-Functional Learning and Collaboration
• 40:56 AI Training Integration
• 45:41 Strategies for Implementing Change
• 47:07 The Role of Soft Skills in AI
• 51:40 Next Episode - Mitchell Tan

Transcript

The Data Mix - Season 3 Episode 2 - Rosanne Werner
Dec 19, 2024 • 10:16 AM - 11:10 AM • Brian Booden • Audio Upload
Brian Booden 00:00
Your regular fix for the best guests from the world of data and analytics. This is the data mix with Brian Boone and George. Good
afternoon, George.
Speaker 1 00:21
Good afternoon, Brian.
Brian Booden 00:22
How are you, my good friend? You all right?
Speaker 1 00:24
I'm really good. Back after just two weeks, just like we said we would be.
Brian Booden 00:28
I know. I know. Almost like this is a thing, right? So I hear that your son told you off for not looking at the camera enough, is what
you told me before the episode. So I'm going to be watching you watching yourself now to see…
Speaker 1 00:43
What I'm trying to do, I'm going to try and put my window where I'm looking at you and our guest as close to my camera as possible.
Yeah. That's what I tend to do. I look on screen at whoever I'm speaking to.
Brian Booden 00:56
That's what I do. I'm going straight down the barrel. But let's work out our presenting issues as we go here. It's amazing to be back
on with you, mate, every two weeks. Here we are now with the data mix. And, again, if you've been following our promos, you know
we've got an amazing guest on there to start with.
Brian Booden 01:11
But let's get a few of the formalities out of the way, George. It's been a couple of weeks, and we had an amazing response from our
last episode with Brendan Grady. And because we had such an amazing response, it's only fair for us to speak to you guys, our
wonderful audience, again. If you like what we do, if you watch our show, whether you download us on YouTube, Spotify, Apple,
Brian Booden 01:36
you watch us on LinkedIn Live, please, fingers crossed, give us a wee review on one of those platforms. I think our Spotify reviews
went pretty crazy after the last episode, George. I think we got up to like 10 or 11, so please keep at it, guys. We really love your
support. Any comments, George?
Speaker 1 01:53
Yeah, it means we can get the best guests on. The more popular we are, the more people want to come and talk about what they do,
which is exactly what we want to share with our audience.
Brian Booden 02:06
Super stuff, super stuff. So, we've now bought enough time for you to check out those QR codes. Hopefully, you gave them a little bit
of a scan. But look, George, I think it's time to like get into the show for real. So, another few comments that we had on the show last
week
Brian Booden 02:23
is the datafix really went down as a cool segment, when that's good, because it's part of our show moving forward. So, it's great to
hear that. So, shall we move into the datafix part of the show, mate? Let's do the datafix, yeah. Well, let's do it then, right.
www.sembly.ai Page 1 of 16
Brian Booden 02:38
Okay, let's get this coordination going. So, we've got the datafix and here's your little windup here. So, the datafix is our segment
where we talk about some of the big things that are happening in tech right now, George. So, why don't you take it away for us, mate?
Brian Booden 02:55
What's been happening in the world of tech over the last few weeks?
Speaker 1 02:58
Well, Brian, I don't know if you'd be a bit like me, but in the corporate world, I've been in London for the last two weeks and it's just
been wall-to-wall Christmas party. Every pub you go into in the evening has got a corned off area and there's a Christmas party going
on.
Speaker 1 03:14
And we both had our little data mix Christmas party, at least we had some Christmas macaroni and cheese, mac and cheese in your
local last week, which was very festive. But we're not the only ones. I think it's worthwhile talking about OpenAI because over the last
week and a bit,
Speaker 1 03:35
OpenAI has had their 12 days of shipmas, which I thought was quite fun. And during these 12 days, they've been dropping quite a
few goodies. A lot of these goodies have been teased over the whole year, but a couple of things that we got were, for example,
Speaker 1 03:53
they've got their O1 full version of their reasoning model is out. Which I haven't been able to play with yet, because I'm not a
subscriber. I subscribe to Claude, which is Anthropic's version, which is also pretty cool.
Brian Booden 04:11
How much is a sub, mate? Do you know?
Speaker 1 04:13
£18 a month in UK money, I think. What, about $20, $22, that. So, I mean, it's worth it, it's worth it. But a couple of interesting things
that OpenAI dropped, their O1 reasoning model, they have a £200 a month now subscription. Sorry, I think that's $200 a month
subscription.
Speaker 1 04:35
That was something that was teased earlier on in the year. I think it was met with amazement and derision. Nobody would pay that.
Yeah, I can imagine. But people are, people are, which is interesting. And they also dropped a Santa voice. So you can have OpenAI in
Santa mode now, which is quite fun.
Speaker 1 04:58
So that's open to anybody, even if you're not paying. And it is quite cool. But one of the biggest things that they dropped was Sora.
So Sora was something that they teased way back in February this year. This is their video generating tool. So you can use a text
prompt to generate a video. And that is cool.
Speaker 1 05:21
Now, again, I haven't played with it, but I've seen some of the outputs from it. And it's amazing. I think it's coming out with mixed
responses, partly because I think if you are in the graphics creating industry, You're wondering, is this going to impinge on my work
now? But something that's been doing the rounds on the Internet is Coca-Cola short.
Speaker 1 05:52
That's several little videos stitched together, made entirely from AI and it's pretty good. Yeah. I believe, Brian, we've got that.
www.sembly.ai Page 2 of 16
Brian Booden 06:04
We actually do, yeah, we actually do have that and we've got one of the one of the three shorts here and It's called unexpected Santa
and I thought it may be worth just playing it to the channel actually just so we can have a look At it and maybe have a quick chat about
where the the great parts and where the parts that need work are Yeah, we do it. I think we should. All right, man. Let's do it. Right.
Thank you so much for that intro mate Right, so let's play this and we'll we'll just let everyone enjoy this All right It's unexpected Santa
created by real magic AI
Brian Booden 07:38
So like I'll have a quick starting point here George So I there's one big thing I noticed on this to start with and that's when the
coca-cola trucks were coming along Like two of them had the coca-cola signage that was really quite good quality, but the middle
truck Had it all like blurry like and I don't know why but it really just totally stuck out to me But overall the other thing I noticed was
that the clips are very short so the camera cuts are very short almost like it can't create the
Brian Booden 08:09
Sort of five to ten seconds in a row. So those were a couple of things I mentioned, but overall Pretty decent. I would say it is pretty
good
Speaker 1 08:17
I mean, I was also impressed you're right to pick out the very short clips So what one and a half seconds two seconds was was about
as long as as each clip was And on the tech side of things. Well, I think AI still has problems with text and images And but but overall,
I mean we don't know how long it took them to produce this The continuity is not bad. If you think each each one and a half second
clip, they're gonna have to reprompt And they're not going to get the continuity from the previous Segment either so, you know, it's not
bad. Do I think Hollywood needs to worry right now? Not quite
Speaker 1 09:00
but It's kind of moving in that direction, isn't it?
Brian Booden 09:06
Yeah, look, I think it's a pretty impressive goal. And I'm pretty sure we'll get our guests definitely to have a comment on Unexpected
Santa as well. I think that's a must. But not a bad first effort, I would say. Really quite enjoyed watching it. And it felt quite natural.
Brian Booden 09:25
But it'll be interesting to hear from our audience what you thought about that as an AI-generated ad and whether it cut the mustard
for you. If your M&S ads came up on the TV, sorry, but I should, for our global audience, Marks & Spencers is a very British brand that
traditionally gives an advert out every Christmas.
Brian Booden 09:42
But I'd be interested to hear your thoughts. Pop them into the chat if you want to discuss about AI as well. Anything else, George, to
cover on this topic? Are we in good shape?
Speaker 1 09:51
I think while we're talking about video shorts, we should probably say that the data mix is now starting to do two or three 20 to
30-minute shorts. We're just following the trend. We're pushing them out on LinkedIn, YouTube, and Instagram. So if you haven't
seen these yet,
Speaker 1 10:12
then hopefully they will be coming to your feed in the next few days.
Brian Booden 10:17
All right, well, that was awesome and that was the data fix for this week's episode. Thank you very much, George. That was
fascinating. I look forward to seeing how that evolves as we start to move on with our tech as well. So, if this was a Christmas dinner,
that would be your, like, I don't know, what would be your starter for Christmas dinner? Prawn cocktail.
www.sembly.ai Page 3 of 16
Speaker 1 10:43
That's our prawn cocktail.
Brian Booden 10:46
Prawn cocktail. Well, now we're moving on to the full turkey roast, aren't we? And it's time to get our main guest on for the day. And
our guest today is the wonderful Roseanne Werner. So, let's see if we can switch her on. Look at that, first time.
Speaker 1 11:00
Hi, Roseanne. Hi.
Speaker 2 11:01
Hi, George. Hi, Brian. Thanks for having me.
Brian Booden 11:05
Oh, you're very welcome. Rosanne, before we get started anyway, what did you think of Unexpected Santa? How did you, like, what
did you think about it?
Speaker 2 11:13
I thought it was really a really good go at AI. I mean, it's emerging and the fact that they've been able to create that straight from AI,
it's their way of being innovative about the, you know, the emerging technologies. So it was a really good go and people are talking
about it. So that's even better.
Brian Booden 11:34
Yeah, and that's good. And I think there's a couple of accompanying ones as well. So it's an interesting watch if you want to go and
dig them out on YouTube. But, Rosanne, it's wonderful to have you here. We're very excited for you to be here for our Christmas
episode. But we are very aware that we've been talking to you a lot, but maybe the whole world doesn't know who you are.
Brian Booden 11:55
So why don't you give us a little bit of an intro of who you are, what you can do?
Speaker 2 11:59
Sure. So my name's Rosanne Werner and I'm the CEO of Accelerate IQ. So we help companies with the data and digital enablement.
So it's all about data and AI fluency, habits and behaviours, which is the foundation of mindset and culture. So we do this through a
people-first approach. So how the brain changes and learns with information,
Speaker 2 12:26
but then also behavioural principles such as like social learning and habit formation. So, I had spent my last 10 years prior to setting
up this company at Coca-Cola, so it's a very big passion of mine, being at Coca-Cola for the last 10 years, and I've been doing finance
transformation prior to that, so I'm a child of accountants by trade, and through finance transformation, leading into data and digital
transformation.
Brian Booden 12:56
Well, that's a great introduction, and a great place to start here, and I think the part that's really ... We'll dig into a lot of that, but the
part that's really interesting for me is the behavioral science part, right? Because George, it's not something that we've attacked in
great amount of detail so far, right? The behavioral science part of- Not at all.
Brian Booden 13:16
Not at all. Just learning about AI, but the culture within organizations and how that all plays a part, so interesting stuff, right?
Speaker 1 13:21
Yeah, and maybe we'll get some answers as to why whenever I go into an organization and try and teach data literacy, I feel like I'm
banging my head off a brick wall.
www.sembly.ai Page 4 of 16
Brian Booden 13:31
It seems to be that way sometimes, depending on which organization you go into. But look, following on from the sort of look, look,
we make no bones about it, right? We picked the Coca-Cola theme and we picked the video and we knew you were there and it's
Christmas and it all works together, right? So I think this is a good place to start, right? I think, Roseanne, you said by your own
admission that you spent quite a bit of time with Coca-Cola. So do you have any anecdotes from Christmas that you might have
regarding Coca-Cola and
Brian Booden 14:05
what happens?
Speaker 2 14:06
Yes, so Coca-Cola at Christmastime is a really buzzing time, they deck out the whole office with, there's like Christmas Santa
everywhere. One of the things that is really good that they do in the Coca-Cola office is that they call out office-based staff to help our
field sales guys on the ground. So you can imagine Christmastime, it is absolutely manic in there, Coca-Cola products are flying out
the door, and so we would go to the local Sainsbury's or Tesco's, Morris and Asta and help our field sales teams in making sure that
the products are well-stocked, they're
Speaker 2 14:52
readily available for consumers on their shelf. So those weeks leading up to Christmas, not only are we being able to bond with the
field sales guys, you meet people from outside of your functions, it's a really great way to learn cross-functionally, as well as we're
lifting a lot of Coca-Cola bottles, so you get a little bit of workout before your Christmas dinner, and you also hear Mariah Carey on the
Christmas door going on repeat all the time, but it really does give a really good chance for teams across Coca-Cola to learn about the
challenges faced by the field sales
Speaker 2 15:31
teams, and then also reflect on what their own processes in the office are like, and how they can support the guys on the front lines.
So it's a really good way, you know, leading up to Christmas, if you really get into the spirit, we all wear our Coca-Cola Christmas
jumpers, and it's a really nice way to just kind of spend time with people across different functions.
Brian Booden 15:52
So my question on that, does anyone try and dress up as the Coca-Cola version of Santa Claus? That's what we really want to know,
right? How many Santa suits are kicking around with all that red and white?
Speaker 2 16:05
Yeah, that's really hard to stack shelves with, if you've got these big suits on.
Brian Booden 16:10
What about you, George? Any requirements for you to know about what's going on in Coca-Cola's culture at Christmas time?
Speaker 1 16:18
There is the urban legend that the Red Santa was a marketing thing that came from Coca-Cola and before that he wore a green suit
or a green jacket. Do you know anything about that, Roseanne? We know that Coca-Cola have a very strong marketing department.
Speaker 2 16:39
Yeah, so Santa came about from Sitnik, which is I think 280 AD, but it's in the 19th century when Santa became red, and that was
through a cartoonish, I think his name was Thomas Nast. So that's when he became, turned up on US confectionaries and the
magazines as Red Santa. But in the 1830s was when Coca-Cola started advertising with their version of Red Santa, and that's when
that came through. So Coke didn't create Red Santa, but they started doing the advertising in about 1830s.
Speaker 2 17:20
Interesting.
Speaker 1 17:21
Okay, well, let's put that one to bed for me.
www.sembly.ai Page 5 of 16
Brian Booden 17:24
Well, I think so, and it's, you know what, it's always nice to get something personal from our guests that we would never have known
otherwise, so I'm glad that we've now put that myth to bed. I bet it's a super fun place to be at that time of year, so we really
appreciate you talking about that as well. Just while we've been chatting,
Brian Booden 17:45
a few things have come in, Roseanne, so we'll just pop these up on the screen to mention them here. We've got, hi, Adrian, thanks for
joining the stream here, but again, going back to, I've forgotten the name of the clip already, Unexpected Santa, again, existing data
visualizations
Brian Booden 18:02
as images and modifying them and finding the correct data points. I actually see this one all the time whenever you see one of these
big billboards that's generated by AI, and George, I'm pretty sure we went through actually a couple of examples
Brian Booden 18:16
where we were looking at AI posters and just nonsense text on there and pie charts that don't add up and all kinds of things. So. It
depends what your purpose is, I suppose, right? It's a, are you just wanting a quick hit
Brian Booden 18:31
of just an infographic that no one's gonna pay any attention to, or do you really, if you really need that detail, I think that's where the
AI is struggling. And I saw another one this morning of Sora trying to replicate a gymnast, a floor gymnast, and that did not go well.
Brian Booden 18:47
Apparently that's one of the most complex things to be able to generate, so thanks for chipping in with that, Adrian. And hopefully,
hopefully, Roseanne knows who Ange from AdMaple is, because surely everyone knows Roseanne. It sounds like a Christmas
comedy show.
Brian Booden 19:04
Surely everyone knows Roseanne.
Speaker 2 19:07
Bless Ange, thanks Ange.
Brian Booden 19:10
So thank you so much, folks, and keep your comments coming in. We'll continue to reference them through the show and have a bit
of a chat with the Q&A with Roseanne at the end. But directing a bit more towards matters at hand, it's lovely to talk about Coca-Cola,
Brian Booden 19:26
but some of the things that you mentioned, Roseanne, in terms of technology change and the way that technology is changing so
fast and that the companies have to change at that pace as well, right? And there's a lot of challenge there, and I'm sure, I see you
write about these things
Brian Booden 19:44
every day on LinkedIn, so what are your opening thoughts on that situation in terms of the rate of technology change?
Speaker 2 19:51
Yeah, I think there's two key parts in this, in terms of how fast the amount of data is growing to, I think, from the latest stats, it's like
90% of the world's data has been generated in the last two years, and that rate is accelerating. So, but at the same time, people are
saying, oh, I want more data, you know, I want more data to answer my questions. So, there's that bit of, you know, how much data is
being generated and sifting through the noise.
www.sembly.ai Page 6 of 16
Speaker 2 20:23
And then you've got the rate of technology, which is accelerating at a higher pace than what organisations can really absorb it. And
that's what they call, there's a, it's called a Martex law. So. The way, you know, the way technology is changing so rapidly, the pace is
accelerating, but then organisations, it takes time to change the way of thinking, and then you're talking about people's behaviours
and habits as well, which don't change overnight, and let alone
Speaker 2 20:54
talk about existing structures and processes and the cultural momentum behind it, so it's a, you know, that gap is getting bigger and
bigger.
Speaker 1 21:04
So, Martek's law, I haven't heard of that. Is that a little bit like Moore's law then?
Speaker 2 21:12
Yeah, so Moore's law is this part, like so Moore's law I think is like the transistors like improving itself and being more powerful and
less costly. So, yeah, it's like the technology is like accelerating really quickly and then that you've got that gap with a lot of
accelerating and then like organizations
Speaker 2 21:31
not being able to absorb that change. So, you need to be like really strategic in terms of what technology tools you're absorbing and
wanting to adopt. So, it's having that strategy as a choice in terms of what technology you're adopting and then more importantly,
what you're not absorbing.
Speaker 1 21:52
Yeah, because you're not suggesting that we should slow down the rate of data generation more. It's just how we manage it.
Speaker 2 22:00
Is that right? Exactly, it's choosing what to adopt and making sure that the technology you're adopting solves the real problems
you're trying to solve. If you're trying to adopt all the AI tools out there, you're gonna end up diluting your efforts too much that you're
not actually getting any benefits from it.
Brian Booden 22:16
Yeah, I think that's one thing I see a lot of, Roseanne, is like technology starts splintering and especially with the pick the newest
thing and then start it, don't finish it and actually leaving behind current relevant projects to dabble in new stuff. And I suppose this
is a very general message, right?
Brian Booden 22:37
But just ignoring the baseline and in lieu of like catching the spike, if you like. And it's quite a difficult one to manage nowadays. And
I think it's applicable from both like a software and like a literacy point of view as well. It's the physical stuff that happens in your
organization, but also there's so much to learn, right?
Brian Booden 22:58
And if you don't consolidate that literacy and whether that's data or AI literacy, we'll have another topic on this in a future episodes
about what the differences between those are. But it's a bit of a minefield, right? So what tips do you have to sort of help to absorb
that change effectively?
Brian Booden 23:17
Because part of absorbing is concentrating on core topics, right? So what are some tips you can follow there?
Speaker 2 23:25
I would say pick out the tools that align with your organization's overall strategy. What are the high-impact initiatives that drive the
real value, and then go with those ones. And identify which ones to pursue, which ones to ignore, and which ones to park. And
identify those ones that would really solve the real value and the real problems in your businesses. And then I would say then break
that down.
www.sembly.ai Page 7 of 16
Speaker 2 23:56
So you can't adopt everything overnight, and it overwhelms people as well. So how do you break that down in terms of that
incremental adoption? So breaking down the change into manageable trunks so that they can incorporate that into the daily
operations. So it's like little incremental changes so that they can build their confidence over time, and then learn, and then feedback,
and then build their confidence again, and learn a bit more.
Speaker 2 24:24
And I would say like taking that people-first approach in terms of getting diverse and cross-functional collaborations together, so the
tools that are being adopted and integrated into processes are fit for purpose for all, so you don't want it to be done for the marketing
guys but then the supply chain guys can't use it, so how do you embrace that change for all in terms of bringing that collaboration
together? And then you mentioned in terms of the employee readiness and the capabilities building side of things, so building the
knowledge on what those tools are but then turning those knowledge into action, so what are they actually doing
Speaker 2 25:12
with it, with the tools and how are they measuring the benefits from the tools?
Speaker 1 25:19
That's interesting and quite relatable for certainly the line of work that I'm in and I know a lot of our listeners, where digital
transformation sometimes can be entirely organisational and encompassing, so in a sort of... private equity arena where an
organization has just been acquired, sometimes they will use technology to look for improvements to things like efficiency. But there
then you've got an entire organization that are going through a change.
Speaker 1 26:03
And there could be a PMO, there could be a project management team doing that. But from your experience, Roseanne, what for
these sort of changes that will take an entire department, a function or even an organization, what sort of timescales might we be
talking about
Speaker 1 26:20
and who should be leading this, who's sort of responsible for it? And again, what are some tips for success?
Brian Booden 26:27
Great question, mate, great question.
Speaker 2 26:29
Yeah, so you will have a project lead, but it's important that the leadership alignment is there as well. So the message coming from
the top in terms of, this is why we're adopting the change. This is how it aligns with our strategic goals. But I also feel like you have
your champions
Speaker 2 26:50
or influencers that are on the ground that influences their teams to challenge status quo. Why are they doing things that they're
doing? Are they embedding the new routines and embedding the new technology? And they're the ones that can influence the
people's day-to-day routines
Speaker 2 27:11
and just kind of go tap into that. We talked about that incremental change. So how do you tap into the incremental change of, okay,
we're having a team meeting. How do we embrace the new technology in the first five minutes of the team meeting? How do we
share the success stories that we've experienced?
Speaker 2 27:30
Kind of generate that pool within the team about. what they've been using the tools for, what are the success stories, what are the
learnings, to kind of pull everyone else on that journey as well. So that social learning is really important in terms of people learn
better in a group. They feel the brains are wired to learn more in a group.
Speaker 2 27:55
So when you're in a group setting, that really helps people with that change, not just by reading about the change or having a
learning materials on the changes, actually, and okay, well, I can see this change in action. I've seen the success stories. How do I be
part of this change?
www.sembly.ai Page 8 of 16
Speaker 2 28:14
So you've got the message from the top in terms of why we're doing it on a big level, and then the educational materials, but then
that influencer champions bit, that's a bit that then drives the behaviors into the teams.
Brian Booden 28:30
I think with that behaviors part, and you actually mentioned this in a recent post as well, was just that chunking methodology of
instead of trying to attack everything as a big, huge problem, the same goes for learning, right? And especially when learning is
applied to literacy, it's very difficult for a lot of people
Brian Booden 28:50
to sit down and consume a three-day course of how to be data literate. It's almost like self-affecting, or it can be a bit negative in a
way, because you're just so overloaded that you come out of it. And it's such a theory-based proposition that it really lends itself to
be broken down
Brian Booden 29:11
into smaller parts.
Speaker 2 29:13
Exactly. And there's studies to say that we learn 10% of what we read. So that's only a little, small chunk. But then we actually learn
like 80% of what we experience and 95% of what we teach others. So there's different types of learning and you know, you can't, I'm a
big rider, I'm a big cyclist and my little one always goes, I want to ride a bike and he needs to go out and ride the bike.
Speaker 2 29:42
You know, I can't just read him a book and expect him to hop on a bike and ride around the park. It's like really getting on, trying it. So
you learn by doing a bit of reading, but then how do you then turn that into doing? How do you turn the theory into practice?
Brian Booden 29:58
And that's an important part, right? And especially for data literacy as well. And we talked about in terms of data literacy, a program
can mean a lot of things, right? But that has a lot of advantages, but a lot of shortcomings as well. So what are some of the, like,
because I'm guessing from a culture, a mindset and a kind of point of view, you probably see a lot of those shortcomings of common
data literacy strategies that are out there in companies.
Brian Booden 30:25
So what are some of those that you've noticed and how can they be improved?
Speaker 2 30:29
I've been speaking to a lot of data leaders in Europe, in the US, and in Asia, and what I'm seeing is that they've gone through the data
literacy programs, and it might be a big e-learning program, but then they've spent the money, but then they're not seeing the change
in the behaviors. And this is where you've got that 10% of learning in terms of the reading, but are they actually turning that into
day-to-day changes in their behaviors on how they interact with data, how they understand data, how they interpret data, and use the
data for the decisions as
Speaker 2 31:12
opposed to relying back on gut feel. So, I led the data and AI mindset and culture program at Coca-Cola, and we were recognized as
a finalist in the Data IQ Awards for the best data skills program, and we were also listed in the CIO 100 Awards. So, what really
worked in that program was that there was the learning piece, but then it was the reflection piece in terms of, okay, this is what you've
learned, how are you going to apply this into your day-to-day in the next week, and then share back in terms of
Speaker 2 31:51
how you've done it, what are the learnings. And then we also had global P2P workshops, so we had about 200 across the world data
catalysts turning up in the last workshop on storytelling, and it was about... we learned about the data storytelling module, but these
catalyst workshops was, okay, how do we tackle these missions together? And you had people from cross functions coming in,
www.sembly.ai Page 9 of 16
Speaker 2 32:20
so marketing, talking to supply chain, and you get that real data lineage exposure in terms of, oh, I've got access to this information,
this data, or I enter this information in, and actually that's got a effect down the line. So it's kind of that awareness of that data
lineage, and that social learning piece of discussing
Speaker 2 32:41
about how they interpret what good quality data is, what practices do they take into account to make sure that their data quality is
correct, but then also how they use the data for decision making, how they implement the changes within their teams. So it's, the
learning, but learning by doing,
Speaker 2 33:05
and then sharing those success stories.
Speaker 1 33:08
Yeah, so you built in feedback then. This was an interactive training. What audience were you selling this to internally? And then can
you talk about some of the benefits of upskilling Coca-Cola with data literacy?
Speaker 2 33:28
Yeah, so we started off with the marketing commercial teams and the supply chain teams, who were like the key engines of
Coca-Cola, the ones that sell the Coke and the ones that make the Coke. And they're also big creators of data and big usage of data,
and then everyone in between. But the program targeted the data catalyst.
Speaker 2 33:54
So we talked about the champions that influence their teams. So this was like a hymn sheet to give everyone in terms of, okay, this
is what we're doing. This is a data strategy. This is how, these are the behaviors we want the company to display in terms of how
they understand data,
Speaker 2 34:13
interact with data, talk about data, make decisions based on the data. And, having that cross-functional interaction, as well as
different territories across the globe, allowed us to view the challenges and the success stories across the different functions.
Speaker 2 34:33
So supply chain might go, well, I had stopped the floor interactions with the teams, and we were like, okay, what's the dashboard
telling us today? How do we use dashboards to change how we make decisions today? And then the commercial guys would go,
Speaker 2 34:52
okay, well, these are the learnings that you guys are taking place. Maybe I'll take that back to my teams in a different spin on how to
influence my teams. So it was that cross-functional learnings in terms of challenging how people change their behaviors, but then
also understanding the data lineages
Speaker 2 35:10
across the different functions. So having that awareness of the data impact down the line or up the line. So it's not so siloed.
Speaker 1 35:20
Yeah. Interesting. Thank you, Rosanne.
Brian Booden 35:23
I was just gonna say that that continuity, ironically, after that is very important, but often you see it brought in at a later point during a
program, right? You're very much taught the basics or the practical, and then communication is brought in as an afterthought and in
between departments.
www.sembly.ai Page 10 of 16
Brian Booden 35:39
So I think you're really right on something there with getting that part in first, so that there's no... You can't just train one part of an
organization, then another part, and then another part, and then here's the glue that'll stick you together at the end because it's...
Brian Booden 35:56
It's a house of cards that's built to fall, right? So then you get pockets of stuff happening in your organization but you can't sustain it.
So I think what you're saying there is really important.
Speaker 2 36:07
Yeah, we get feedback from the data catalysts in terms of they get to speak to people outside of their teams. So it's like data's not
just within their teams, it's actually a wider organizational business initiative
Brian Booden 36:26
And just a quick word for the chat as well, we're seeing everything that's coming in there and we're gonna come back to all of it in the
QA, so just be patient with us and hang on, there's some great stuff coming in there. So like, if you nail it down to some specifics,
then Roseanne, right, there must be some critical skills
Brian Booden 36:45
that are needed, you know, to make your workforce sort of future ready for this type of, because I think some things we forget, right,
we've been through a pandemic and before that, there was a, and a lot has changed with the combination of AI and a pandemic, a lot
has changed technology-wise, right?
Brian Booden 37:04
And we need to remember that when we're thinking about the skills that our workforce needs, even the skills that were needed five to
10 years ago pre-pandemic are maybe not quite the same profile as they are now, so what are the things we should be focusing on
right now in terms of those critical skills?
Speaker 2 37:22
Thank you. I guess what we've seen is, especially with the rise in AI, is sometimes the hard technical skills don't just cut it anymore.
You really do need those soft tug-of-war power skills to bridge the gap. I call them the five Cs, you know, the curiosity, the creativity,
the critical thinking, communication and collaboration. And they're things that I have not seen AI been able to replicate well, right?
Speaker 2 38:00
And you've got curiosity, so how do we dig deeper into the new and existing data? How do you explore with it the creativity? How do
we do new ways of doing things with the existing data? And when we, especially with AI, what I'm seeing is, you know, the whole
traditional way of learning, you know, we're taught what to learn at school. And when AI comes about and people are trying to adopt
AI, they're like, well, tell me what to put in the system.
Speaker 2 38:31
And it's like, well, use your curiosity, use your creativity. What would you put in the system? And I think that's something that people
are struggling with when it comes to, you know, adopting AI. And I think it's tapping into that curiosity, creativity on how to generate a
prompt, how to generate the prompt to get the Coca-Cola truck come through. But I also think...
Speaker 2 38:55
When it comes to AI outputs, your critical thinking is really important there. are we critically thinking about the outputs in terms of is
there bias in the output, is it hallucinating, you know, is it really AI generated, as does it have AI generated words in there that sounds
really, you know, robotic.
Speaker 2 39:21
So I think it's using AI, but having that kind of AI sandwich. So you're like, you know, having your curiosity and creativity, and then
you've got the AI that generates the output, and then using your critical thinking skills to then, you know, get the best out of AI.
www.sembly.ai Page 11 of 16
Speaker 2 39:39
And you mentioned the communication and collaboration, they're the bits that, you know, that human connection and that human
communication bit that, you know, gels us all together, that emotional connection that AI is probably not quite able to replicate at the
moment.
Speaker 1 39:55
Yeah. And Rosanna, have you seen organisations now weaving AI training into their data literacy programmes?
Speaker 2 40:06
Yes, but you need to start to do AI, you need to have your data right. So if you have bad data and you put AI on top of it, you get
super turbo bad, you know, results. So yes, AI literacy is important in terms of like how do we, the cognitive ability in terms of
prompting, but also in order to have a really good foundation,
Speaker 2 40:34
you need to understand your data literacy that then feeds into the AI.
Speaker 1 40:39
Yeah. So it's sort of, it starts with a foundation of data literacy then, on top of good data within your organisation. And then after
that, you can start learning and understanding how to use prompts to get what you need from data.
Speaker 2 40:55
Thank you.
Brian Booden 40:56
Yeah, wow great world. I think and also I just think having that alliterative Methodology as well. I think is like for those who are neuro
diverse I think having something like the five C's or something like that that you can latch on to is a really important thing and it
shouldn't be underrated because more and more of the workforce are being highlighted as neuro diverse and they're actually that's
actually being diagnosed and It's it's just a thing nowadays. It may not and again, I don't even think
Brian Booden 41:31
Pre-pandemic or in the last 10 years. It was really recognized as a this is a genuine part of the workforce So I think to have anything
that you can latch on to with that is super important as well Is that something that you notice is that when you have those
frameworks? They tend to embed a bit better because of that as well and having that structure
Speaker 2 41:50
Because it's also embracing the personalised, or the ability to personalise the learnings. People learn differently, so, and then having
those avenues to learn in different manners then allows that upskilling to take on a lot faster and that adoption as a result of it.
Brian Booden 42:12
Well, that's amazing. I think that's pretty awesome insights, and I'm just aware that the chat is blowing up here on the side, so I'm
quite keen to give us a bit of time to look at some of those if you're up for it, Roseanne. Let's just have a bit of a recap and Q&A here.
What we'll do is we'll just move down some of the questions,
Brian Booden 42:34
and we've got Joe here to start with. I could have flashed this up quite a lot earlier, but how important would you say data is in the
age of AI? You talked about baselining, but anything more to say on that, Roseanne?
Speaker 2 42:50
Yeah. In the stage of AI, you've got to get your data ready. It's not just the amount of data, it's the quality of the data, and then also
have you got the processes in place to do it. So very important. Hi, Joe.
Brian Booden 43:10
George, I think you might be interested in this one as well, but again, this one's more on the edge of sustainability and data needs to
be sustainable. So sustainability needs data and data needs to be sustainable. George?
www.sembly.ai Page 12 of 16
Speaker 1 43:22
Yeah. Actually, I listened to a very relevant podcast with the CEO of Scale AI. Scale AI make synthetic data or pull together very clean
and very focused datasets for the AI industry. So this is what a lot of the mainstream data models are built on. And we've heard
people like,
Speaker 1 43:45
I think Elon Musk said he's built, what did he call it? Colossus, and he was gonna have 100,000 GPUs in a data center. And then he
said, no, wait a minute, I'm gonna have 200,000. Actually, I'm gonna have a million GPUs
Speaker 1 44:00
in this data center. The point is that I think the limitation is not gonna be how many GPUs you can stash into your data center. It is
how can you get the quality data to train your models on? That's the limiting factor.
Speaker 1 44:16
And I think that's exactly what you've been saying, Rosanne.
Speaker 2 44:18
Yeah, and also when you come across problems, is AI really the right tool for the job? I think you gotta understand what the real
problem to solve is. And I've seen companies, they build all these AI tools where actually you can probably solve it with some really
simple automations,
Speaker 2 44:43
but they wanted to go into the shiny tool box to play with AI. And it's really finding whether AI is really the tool for the job and then
invest in it.
Speaker 1 44:58
Yeah, I think I completely missed Lisa's point there. She was talking about sustainability and data. So hopefully that tied in the fact
that we've not got, at the moment, data is the limiting factor within what we're doing here.
Brian Booden 45:20
Yeah, and we have a few others as well. I'm going to bring up Angie again here at the bottom. So what are some quick-win strategies
that you've seen work in terms of bedding in the new system fast? What are the things that we can do to make that work? Not
putting you on the spot or anything, Roseanne, but yeah.
Speaker 2 45:41
I'd say start small but be strategic in terms of what you're doing and so we're not overwhelming your workforce at once but then at
the same time you're starting, you're starting with experimenting, you're starting with the change, you're getting that feedback in a
quick manner and then I always say use your champions and catalysts because they're the ones that are able to kind of give you the
feedback but then also the microphone of the change into the masses so it's like start small but strategic and then making sure you're
using your catalysts, your change agents to kind of really drive that, ripple that change across the organisation
Speaker 2 46:29
and just make sure you're equipping them with the right tools as well so the learnings, the always-on collaboration so that, you know,
that change is more of a pull than a push.
Brian Booden 46:42
And we've got time for one more, I think, George. And we know Nathan's from the Qlik ecosystem, so we know Nathan well. Soft
skills, key for the success of AI in the future. And I think my take on this is that I still see so many people just putting a one-line
prompt into GPT and just expecting the answers to pop out.
Brian Booden 47:07
It's still so common. And we all do it, right? We all do a quick hit to see if we can just get a quick answer. But. What we're doing is
we're misguiding these tools and we're teaching them to misguide themselves by not giving them.
www.sembly.ai Page 13 of 16
Brian Booden 47:21
And the data learning process or something like that is incredibly, an incredibly long period of time and a massive amount of data.
So the better, when we start to learn how to prompt AI, that's when we'll see the benefit of AI most in my opinion. I think we're not
quite there yet, but I think that's why having the frameworks
Brian Booden 47:41
is really important because you really need to stop before you get on your keyboard and start going do, do, do, do, do, and to chat
GPT, you really need to think, am I being creative enough with my prompt? Am I being critical enough with the question that I'm
asking?
Brian Booden 47:55
Or am I just feeding in a diagram and saying, just give me the solution, chat GPT. And I do worry about our next generation about
whether their brains are gonna work in the same way because of that as well. I think there's a real cognizance issue for the next
generation too.
Brian Booden 48:10
So I'd be interested to hear what you think about that and the cognizance of our younger generation that are growing up with all this
tech and how it changes.
Speaker 2 48:17
Yeah, well, I'd like to probably say one thing in terms of like the prompting side of things. One thing I found really worked well is, and
I'm experimenting with this, is asking Claude or ChachiBT, like, this is my prompt, can you tell me what I'm missing to generate a
better prompt? And it will go, well, you're missing the tone or what type of output would you like? And that's kind of helped me then
reframe that prompt. So I've been kind of brainstorming with AI tools in terms of how I can improve my prompt
Speaker 2 48:52
and testing with that as well. With the younger generation, one thing I have thought is, we were looking at the Coca-Cola ad before, I
find that they are probably more in tuned with understanding what is fake and what is not, better than what we would be because
they're a lot more exposed to it. And I guess in terms of people can cheat now with their learning styles. So if they want to cheat with
AI tools, then they can. And so I think it's really tapping into that kind of curiosity,
Speaker 2 49:35
the creativity side of things that then helps build their learnings on it. But I think if they want to cheat, they can do that now. And AI
tools is just going to be another tool for them to go around it in a way.
Speaker 1 49:51
Yeah. But I guess it was a little bit like, I mean, we were told not to use calculators at primary school when I was, I mean, I'm going
back a long time. But now calculators are allowed in exams. And I think as long as you have the, what we're testing is, do you know
under the hood, do you know what you're asking? Ask a better question and you'll get a better answer.
Speaker 1 50:14
And the AI is not, I think, as you said, it's not generating new knowledge at the moment anyway, before it can start running its own
experiments. And so staying curious, staying creative, that's got to be part of your keys for success.
Speaker 2 50:31
Exactly. And being critical about the results as well. And just kind of pushing that status quo.
Speaker 1 50:38
I love your point about actually asking it. How would you like me to ask this question better?
Brian Booden 50:43
Yeah, and I think you know having don't miss your C's George, but but yeah, I think that's a that's a good that's a great way to sort of
finish off off here and I've learned so much in this last sort of I'm looking at the clock well 52 minutes. Well, it's just zoomed past We
need to be aware of what's happening. We didn't manage to get through all the comments, but be sure we'll head back through the
YouTube and LinkedIn Live afterwards
www.sembly.ai Page 14 of 16
Brian Booden 51:10
and make sure that we give you all a buzz on the way. We've got a few things to just quickly cover off, so just stay with us for a
minute, Rosanna, and we'll come back to you. But look, guys, I know we mentioned it at the start of the show, but if you enjoyed what
you saw today,
Brian Booden 51:25
please do support us. Please give us a like or subscribe on YouTube, Spotify, Apple. It really does make a difference. So thank you
very much for that. And thank you for everyone that's joined us on LinkedIn Live today.
Brian Booden 51:39
George, I'm gonna let the audience know about our next episode as well. I don't know if you want to just give a quick mention of that
for the next one when I flash that up. Let's do that. All right, tell us who we've got next, mate.
Speaker 1 51:50
So we've got Mitchell Tan on, is it the 30th? Sorry, I should be reading this.
Brian Booden 51:57
Yeah, it's the 30th. First meet, it's Hogmanay, yeah.
Speaker 1 52:00
This is on Hogmanay, as we say in Scotland, or New Year's night, as you probably say in the rest of the world. So Mitchell, we
actually spoke to about three weeks ago. Fascinating chap. He's built an interface for LinkedIn messages, and he's going to talk a
little bit about bootstrapping a brand new business,
Speaker 1 52:27
the data elements of it, how his APIs work, and we're going to go a little bit under the hood as to how he's created this business and
what it means for the data world, especially the organization of data in a usable format.
Brian Booden 52:44
I think it's a really important episode. The LinkedIn inbox is not a fun place to be, and Mitchell is changing that, and as someone who's
used it already, it really is making a difference. So if you spend a lot of time on LinkedIn, if your DMs are a mess, come and join us on
the 31st, and Mitchell and we will guide you through that as well. So thank you very much. We've actually not got a barcode for this
yet, but we'll pass on the details to everyone after the show. So, all that remains really is to say thank you so much,
Brian Booden 53:19
Roseanne, for popping along and giving us all your experience and insight and the holidays are coming knowledge. We're very
excited about all of that. So, thank you so much.
Speaker 2 53:29
Thank you so much for having me. It's been great. Thanks so much, it's been fun.
Speaker 1 53:33
Thank you, Roseanne. And it really has been an eye-opener. And one thing I'm going to take away is Martech's law. That's brand new
information for me.
Brian Booden 53:42
Thank you. And for me, it's the five Cs, so fantastic stuff. That's a wrap for us, guys. Thank you very much for joining on the live
streams. And Roseanne, thank you so much. And George and I will see you in two weeks for our next episode of the Datamix.
Brian Booden 54:00
Thank you very much, everyone. Thank you.
www.sembly.ai Page 15 of 16
www.sembly.ai Page 16 of 16