Chief Data Officers, AI, ML, and Influencers on The Cube [Video]

Chief Data Officers, AI, ML, and Influencers on The Cube

I joined an all-star team of IBM social influencers to speak on Silicon Angle’s The Cube at the IBM CDO Summit recently. We discussed artificial intelligence, machine learning, neural networks, predictive analytics, and so much more. Hear what Tripp Braden, Mike Tamir, Bob Hayes, and I had to say:

Thanks to IBM and Silicon Angle for having us!

Closed Caption Auto-Transcript:

live from Boston Massachusetts
 it’s the queue coverage IBM chief data
 officer summit brought to you by IBM
 welcome back to the cubes live coverage
 of IBM’s chief data a strategy summit
 I’m your host Rebecca Knight along with
 my co-host Dave Vellante we have a big
 panel today these are our social
 influencers starting at the top we have
 Christopher Penn VP Marketing of shift
 communications then Tripp Braden
 executive coach and growth strategist at
 strategic performance partners Mike
 Tamir chief data science officer
 attacked Bob Hayes president of business
 over Broadway thanks so much for joining
 us thank you yeah so so we’re talking
 about data as a way to engage customers
 a way to engage employees what what
 business functions would you say stand
 to benefit the most from using from
 using data nothing too active that if
 it’s the biggest function but I think
 that the customer experience and
 customer success how do you use data to
 help predict what customers and
 customers will do and how do you then
 use that information to kind of
 personalize that experience for them and
 drive up recommendations retention
 absolutely things like that so it’s
 really it’s really the customer
 experience that you’re focusing on yes
 and I just just released a study I found
 that that analytical leading companies
 tend to use analytics to understand
 their customers more than say analytical
 a Gerdes so those kind of companies who
 can actually get value from data they
 focus their efforts around improving
 customer loyalty and just gaining a
 deeper understanding about their
 customers Chris do you want to jump in
 here and you say as many of us said do
 the weave have 3 things you really care
 about us as business people right we
 want to save money save time or make
 money and so any function that those
 company that meets those qualifications
 are this function will benefit from data
 I think there’s also another engineer
 interesting dimension to this when you
 start to look at this the leadership
 team in the company now having the
 ability to anticipate the future I mean
 now this we are no longer just looking
 at static data but we are now looking at
 anticipatory capability and seeing
 around corners so that the person comes
 to the team they’re bringing something
 completely different than the team in
 them has had in the past
 this whole competency of being able to
 anticipate the future and then take from
 that where you take your organization in
 the future so so follow-up on that trip
 just does this data now finally trump
 gut feel remember the HBR article of
 10–15 years ago can’t beat gut feel is
 that we hit a new era now well I think
 we’re moving into an era where you have
 both I think it’s no longer neither or
 you happen to wish you know we have data
 now we have both the organization’s who
 can leverage both at the same time and
 develop that capability and earn the
 trust of the other members by doing that
 that I see the chief data officer really
 being a catalyst for organizational
 change so dr. Tim another one if I get
 asked you a question maybe the whole
 panel but so we’ve all followed the big
 data trend and and the meme AI deep
 learning machine learning same wine new
 bottle or is there something substantive
 behind it or our capabilities are
 growing our capabilities and machine
 learning and I think that’s part of why
 now there’s this new branding of ai ai
 is not what your ear what your mother
 might have thought is it’s not robots
 and Cylons and that sort of thing that
 are going to that are going to be able
 to think intelligently they just did
 intelligence tests on I’m too different
 like Siri and Alexa quote a eyes from
 different companies and they port at
 least be scored horribly that they
 scored worse than in my mind what much
 worse than mice very intelligent
 seven-year-old and that’s not a comment
 on
 the deficiencies in the in Alexa or in
 serious a comment on these are not
 actually artificial intelligences these
 are just tools that apply machine
 learning strategically so who you are
 all thinking about data and how it is
 going to change the future and one of
 the things you said trip is that we can
 now see the future it talked to me about
 some of the most exciting things that
 you’re seeing that companies do that are
 anticipating what what what what
 customers want okay so for example in
 the customer success space a lot of SAS
 businesses have a monthly subscription
 so they’re very worried about customer
 churn so companies are now leveraging
 all the user behavior to understand
 which customers are likely to leave next
 month and if they know that they can
 reach out to them with maybe some
 retention campaigns or even even use
 that data to find out who’s who’s in the
 who’s most likely to buy more from you
 in the next month and then market to
 those and effective way so don’t just do
 a blast for everybody focus on
 particular customers their needs and and
 try to service them or market to them in
 a way that resonates with them that
 increases retention of selling and
 recommendations so they’ve already seen
 certain behaviors that show a customer
 is maybe not going to exactly so she
 just you throw this data and machine
 learning right you find the predictors
 of your of your outcome that interests
 you and then using that information you
 see oh maybe predictors a B and C are
 the ones that actually drive loyalty
 behaviors they accuse an information to
 segment your customers and market to
 them appropriately it’s pretty nice
 pretty cool stuff
 February 18th 2018 okay so we did a
 study recently just for fun of when
 people search for the term outlook
 out-of-office
 yeah and you really only search for that
 term for one reason you’re going on
 vacation and you and you want to figure
 out how the trends feature on so we did
 a five year data poll of people with a
 search box with it and then inverted it
 so when do people search least for that
 term that’s when they’re in the office
 and it’s the week of every 18 2018 will
 be that time when people like yeah
 I mean I’m at the office saying I work
 and knowing that the prediction and day
 to give us specificity like yeah we know
 the first quarter is busy we know
 between Memorial Day and Labor Day is
 not as busy in the b2b world but as a
 marketer we need with specificity data
 and predictive analytics gives us
 specificity we know what week to send
 our email campaigns what week to turn
 our ad budgets all the way to fall and
 so on and so forth if someone’s looking
 for the cue when do when will they be
 doing that you know going forward that’s
 the power of this stuff is that
 specificity we’re gonna search for word
 before we searching I’d like to know
 where I’m gonna be next week why that
 date do not see the date that people
 least search for the term outlook out of
 office okay so they’re not looking for
 that feature which logically means
 there’s a good point on not just what
 you’re predicting for interactions right
 now but also anticipating the trends so
 Bob brought up a good point about
 figuring out when people are turning to
 the flip side of that which is how do
 you get people to be how do you get your
 customers to be more engaged and now we
 have really an explosion in the
 enforcement learning in particular which
 is a tool for figuring out not just how
 to interact with you right now as a
 one-off statically but how do I interact
 with you over time this week next week
 the week after that and using
 reinforcement learning you can actually
 do that this is the sort of technique
 that they used to beat alphago or to be
 humans with alphago machine learning
 algorithms supervised learning works
 well when you get that immediate
 feedback but if you’re playing a game
 you don’t get that feedback that you’re
 gonna win 300 terms from now right now
 you have to create more advanced more
 advanced value functions in ways of
 anticipating where things are going this
 move so that you see that you’re on
 track for winning and 20–30–40 moves
 down from down the road and it’s the
 same thing when you’re dealing with
 customer engagement you want to
 you can make a decision I’m going to
 give this customer coupon that’s going
 to make them spend 50 cents more today
 or you can make decisions
 algorithmically that are going to give
 them a 50 cent discount this week next
 week then the week after that that are
 going to make them become a coffee
 drinker for life or customer for life
 it’s about finding those customers for
 life IBM uses the term cognitive
 business we go to these conferences
 everybody talks about digital
 transformation at the end of the day
 it’s all about how how you use data so
 my question is if you look think about
 the bell curve of organizations that you
 work with how do they you know what’s
 the shape of that curve part one and
 then part two is where do you see IBM on
 that curve well I think a lot of my
 clients make a living predicting the
 future their insurance companies their
 financial services that’s where the CDO
 currently resides and they get a lot of
 benefit but one of the things we’re all
 talking about but talking around is that
 human element so now how do we take the
 human element and incorporate this into
 the structure of how we make our
 decisions and how do we take this
 information and how do we learn to trust
 that and the one thing I hear from most
 of the executives I talk to when they
 talk about how data is being used in
 their organizations is the lack of trust
 now when you have that and you start to
 look at the trends that we’re dealing
 with and we call them data points but
 she’s calling them people now you have a
 problem because people become very
 almost anonymity analytically challenged
 right so how do we get people to start
 saying okay let’s look at this from the
 point of view of it’s not a neither or
 solution in the world we live in today
 cognitive organizations are not going to
 happen tomorrow morning even the most
 progressive organizations are probably
 five years away from really deploying
 them completely but the organizations
 would take a little bit of an edge so
 five 10% edge out of there
 they now have a really a different
 in their markets and that’s what we’re
 talking about hyper critical thinking
 skills I mean when you start to say how
 do I think like Orrin Buffett how do I
 start to look and make these kinds of
 decisions analytically how do I sit how
 do I recreate an artificial intelligent
 machine learning practice and in program
 that’s going to solute profit solution
 for people and that’s where I think
 organizations that are forward leaning
 now are looking and saying how do I get
 my people to use these capabilities and
 ultimately trust that the data that
 they’re told
 so I forget who said it but it was early
 on in the big data movement somebody
 said that we’re further away from a
 single version of the truth and than
 ever and it’s just gonna get worse but
 so was the data scientist Oh what say
 you not familiar with the truth code but
 I think it’s very relevant and very
 relevant to where we are today with
 there’s almost an arms race you hear all
 the time about automating putting out
 fake news putting out misinformation and
 how that can be done using all the
 technology that we have our disposals
 for dispersing that information the only
 way that that’s going to get solved is
 also with algorithmic solutions with
 creating creating algorithms that are
 going to be to be able to detect is this
 news is this something that is trying to
 attack my emotions and convince me just
 based on fear or is this an article
 that’s trying to present actual facts to
 me and you can you you can do that with
 machine learning algorithms but of now
 we have the technology to do that better
 algos them like and share from a
 technological perspective to your
 question about where IBM is IBM has a
 ton of stuff that’s called AI as a
 service essentially where if you’re a
 developer on bluemix for example you can
 plug in to the different components of
 Watson at literally pennies per usage to
 say I want to do sentiment Alice I want
 to do tone analysis I want personality
 insights about this piece of know who
 wrote this piece of content
 and to dr. communes point this is stuff
 that you we need these tools to do
 things like you know fingerprint this
 piece of text did the supposed author
 actually write this you can tell that so
 of all of the the format I recall the
 Microsoft Amazon Google IBM getting on
 board and we’re adding that five or ten
 percent edge that that trip was talking
 about is easiest with IBM bluemix great
 one of the other parts of this is you
 start to talk about what we’re doing and
 you start to look at the players that
 are doing this they are all
 organizations that I would not call
 classical technology organizations they
 were 10 years ago you look at the
 Microsoft but you look at the leadership
 of Microsoft today and they’re much more
 about figuring out what the formulas for
 successful business and that’s the other
 place I think we’re seeing a
 transformation occurring and the early
 adopters is they have gone for the first
 generation in the pain you know of
 having to have these kinds of things and
 now they’re moving in that second
 generation where they’re looking for the
 game and they’re looking for people can
 bring them capability and have the
 conversation and discuss them in ways
 that they can see the landscape I mean
 part of this is we could get caught in
 the bits and bytes you missed landscape
 pitches should be seeing in the market
 and that’s where I think there’s a
 tremendous opportunity for us to really
 look at multiple markets off the same
 data and imagine looking and saying
 here’s what I see everyone in this group
 would have a different opinion in what
 they’re saying but now we have the good
 to see at five different ways and share
 that with our executive team and what
 we’re seeing so we can make better
 decisions I wonder if we could have a
 frank conversation honest conversation
 about the data and the data ownership
 you heard IBM this morning saying we’re
 gonna protect your data but I love you
 guys as independents to weigh in you got
 this data the data you guys are involved
 with your clients building models the
 data trains the model I got I kind of
 believe that that model gets used at a
 lot of different places within an
 industry like insurance or across retail
 whatever it is so I’m afraid that my
 data is gonna my IP is gonna seep across
 the industry should I not be worried
 about that I wonder if you could guys
 could weigh in well if you work with a
 particular vendor sometimes vendors have
 have a stipulation that we will not
 share your models with other clients so
 then you just got to stick to that so
 I mean but in terms of science and you
 build a model right you want to
 generalize that to other businesses so
 so maybe if you could work with work
 somehow with your existing clients say
 here just this is what you want to do
 you just want to you’d want to elevate
 the water for everybody right so
 everybody wins when all boats tries
 right so if you can get that if you can
 kind of convince your clients that we
 just want to help the world be better
 and function better make employees
 happier customers happier let’s let’s
 take that approach and and just use the
 models in a that maybe generalize to
 other situations and use them and if you
 don’t then you just don’t launch a
 transparent about it exactly good yeah
 I’m not super you David and Tripp and I
 are all dressed similarly right we have
 the model of if I put on your clothes we
 wouldn’t but if I were to put on clothes
 it would not be you it’s the same model
 it’s just not gonna be the same outcome
 it’s gonna look really bad okay so yes
 companies can share the models and that
 the general flow is tough but there’s so
 much if a company’s doing machine
 learning well there’s so much feature
 engineering that unions then company
 they’re trying to apply that somewhere
 else just gonna blow up
 yeah we could switch ties
 Chris Tripp Mike and Bob thanks so much
 for joining us this has been a really
 fun and interesting panel thank you very
 much thanks you guys we will have more
 from the IBM Summit in Boston just after
 this



Want to read more like this from Christopher Penn? Get updates here:

subscribe to my newsletter here


Originally published at Christopher S. Penn Marketing Blog.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.