Latest / The Retail Podcast / AI Won’t Replace Merchandisers — It’ll Make Them Superhuman

Transcript

AI Won’t Replace Merchandisers — It’ll Make Them Superhuman · The Retail Podcast

• 0:00
  Again, I don't think this is going to replace anybody.
• 0:03
  I don't think this is going to suck the creativity out of
• 0:05
  people. It's not going to reduce that at
• 0:07
  all. What it's going to be able to do
• 0:09
  is just make you better, more accurate and have a lot of
• 0:13
  financial implications. Hello, welcome back to the
• 0:27
  Retail Podcast. Now you can't leave a conference
• 0:30
  or scroll on LinkedIn without seeing how AI is having such a
• 0:35
  disruptive impact on the industry and specific identity
• 0:40
  AI. And I thought, why don't we get
• 0:41
  some experts to come and tell us what's really happening so you,
• 0:46
  Mr Mrs Retailer, can understand, you know, where am I going?
• 0:51
  Am I on the right track? Am I not?
• 0:53
  Should I be looking at this? Should I not, what's the reality
• 0:56
  right now? Because I think again, sometimes
• 0:58
  I've, I've been at conferences thinking that is this really the
• 1:01
  reality that we where we are now or it's just people's sort of
• 1:05
  biassed or built impressions? And I think there's no better
• 1:08
  person to have this conversation with.
• 1:10
  Noah Hirshman, ex Microsoft colleague.
• 1:13
  So Full disclosure, but someone who I've worked with and someone
• 1:17
  who can, you know, hand on heart saying that he's got some
• 1:20
  merchandising experience just a little.
• 1:23
  Noah, Hello, welcome to the show.
• 1:25
  Hey Alex, good to be here. Thanks for having me.
• 1:27
  Well, it's an absolute pleasure. So Noah, why don't we frame this
• 1:30
  for people who are like they're super interested in AI?
• 1:34
  They are. From a merchandising background.
• 1:37
  And so therefore, their lens is not necessarily the same as
• 1:40
  someone from tech because tech will probably take a different
• 1:44
  view on this and just frame the conversation that we're going to
• 1:47
  have today, if you don't mind. So my background is really about
• 1:51
  retail merchandising and I've been doing it since many, many,
• 1:55
  many years. I started out in the electronics
• 1:57
  business and I also will help build iconic businesses at
• 2:01
  Amazon and at eBay. But merchandising obviously
• 2:04
  extends to all categories, grocery, fashion, you know, all
• 2:09
  sorts of all, you know, Home Depot kinds of things.
• 2:12
  And so it's really, in my opinion, and of course, I'm
• 2:16
  biassed, right? It's the hub of the wheel.
• 2:19
  I mean, I always make. He added to Yeah, sure, the
• 2:22
  Apple Store has great fixtures and fantastic customer service,
• 2:27
  but if it didn't have the iPhone 17, nobody would care.
• 2:30
  Yeah, right. It's really about the product.
• 2:32
  And so it starts from there. And that is sort of the hub of
• 2:36
  the wheel of retail because once there's an assortment and the
• 2:40
  product is great and somehow better than the competitors
• 2:43
  price better, better quality, more features, etcetera, then
• 2:47
  that can be pushed out to the rest of the organisation,
• 2:50
  including obviously supply chain and things around vendors and
• 2:54
  manufacturers, marketing in store merchandising and and and
• 2:58
  set up sales training, e-commerce, etcetera, etcetera.
• 3:03
  So really getting the assortment planning right and the product
• 3:07
  and the pricing right is the whole all what?
• 3:10
  And once that and gone, you can move it out.
• 3:13
  Now the problem is, is that that has heretofore been a very
• 3:17
  manual process, right? And, and you know, for me coming
• 3:21
  back and you know, even as far back as the 80s and the 90s
• 3:24
  where you had to do things on spreadsheets and do manual
• 3:27
  calculations. Let's see, I have 50 in stock.
• 3:29
  I sold 30. I need to buy another 20.
• 3:32
  Don't let kind of stuff and all the way to, you know, it's on
• 3:36
  sale, you know, it's usually 999.
• 3:38
  Now it's on sale for 8, 99 and pushing all those things out is
• 3:42
  very cumbersome thing. So yeah, AI is a great use.
• 3:48
  Now I am going to be very opinionated about this one part
• 3:52
  of AI, which is AI can be, you know, something that can be, you
• 3:58
  know, take, take, take people's jobs or take people's things
• 4:01
  that'd be more efficient and therefore save working hours.
• 4:06
  But that's not really, in my opinion, the great thing about
• 4:11
  AI. The great thing about AI is that
• 4:13
  it could do your job. It could do a better job for you
• 4:17
  as an assistant. You know, Microsoft has Co
• 4:20
  pilot, obviously, and a lot of people are thinking about, OK,
• 4:23
  this is sort of my the person who works for me.
• 4:26
  This is my underlaid, you know, that that that that is in in my
• 4:30
  department. Now.
• 4:30
  They don't they're not a human, you know, they're an algorithm,
• 4:34
  but still like kind of work for me.
• 4:36
  And the truth is, is that when you look at the creativity, the
• 4:40
  that goes around buying and merchandising, it's really a
• 4:44
  half, half night. It's got your right brain, which
• 4:46
  is what's hot. What's my gut feel?
• 4:49
  What's the fashion trend going to be?
• 4:50
  What's the colour next season going to be?
• 4:53
  And can that be replaced by AII? Don't think so.
• 4:56
  I mean, I think that that is the human to human conduct,
• 4:59
  understanding your buyer, understanding your customer in
• 5:01
  your market. Can AI hope you get much more of
• 5:05
  the information of the trend prediction faster?
• 5:08
  Absolutely. But you use that as as a tool,
• 5:11
  right? Yeah.
• 5:12
  So then, you know, the next thing is how do we price it?
• 5:15
  Well, right now a lot of people are still pricing on gut feel.
• 5:18
  Oh, or based on the competitor. But the truth is there's a whole
• 5:22
  other thing about pricing elasticity, which is a very kind
• 5:26
  of long and involved mathematical analytics process
• 5:30
  that most people either don't know how to do or they just
• 5:34
  don't want to be bothered with it because it's really painful.
• 5:37
  So that's something that AI could do very, very quickly.
• 5:40
  And that doesn't really replace something.
• 5:42
  It just enhances or augments what you already do.
• 5:45
  And there's a million examples besides that as well.
• 5:48
  So what I hear is you still need a human in the loop to bring
• 5:53
  that element. But obviously the data that
• 5:55
  human has access to is faster, wider, broader and that can be
• 6:01
  bought to them, right. For this sort of looking at the
• 6:04
  future of trends and how am I going to what am I going to
• 6:07
  focus on in next season or what, what are the products that I
• 6:10
  need or colours. But in my mind they're still
• 6:13
  losing the signals because when I look at.
• 6:17
  I mean it's. Such a silly thing, but when I
• 6:19
  look at for example, K Pop Demon Hunters and now most, most
• 6:24
  retailers missed that opportunity because they weren't
• 6:27
  they didn't have the signals built into their business to
• 6:31
  pick up on trends, real trends, not what Disney's telling,
• 6:35
  right? They completely missed the
• 6:36
  marketing opportunity. They weren't able to bring out
• 6:39
  product, they weren't able to bring out relevance.
• 6:41
  And so therefore culturally during that season, they're a
• 6:45
  little bit less relevant, they're a bit more dated.
• 6:47
  And so I obviously, I think maybe that takes the
• 6:50
  conversation in a different way, but I, I just think that the
• 6:53
  signals that retailers need are still lacking.
• 6:57
  So if we zoom in a little bit more into merchandising and how
• 7:02
  the merchandiser of the future take me through what that looks
• 7:05
  like, what does the merchandiser of the future look like?
• 7:08
  Yes, So, you know, first of all, you're spot on with the K pot
• 7:11
  even hunters thing, right? I mean, this is really about
• 7:15
  identifying trends. You know, there's companies that
• 7:17
  will scrape 10 million Instagram photos to try to figure out what
• 7:21
  the next colour of the feature is going to be.
• 7:23
  And those are the kinds of things that again, is that mix
• 7:27
  between left and right brain movement, creativity and
• 7:29
  analytics. Be able to have the data, but
• 7:32
  not only just the data, sort of the compute power and the
• 7:35
  processing power to make sense of that.
• 7:37
  So how you find those trends? OK, well, once you can do that
• 7:41
  and either you're a manufacturing retailer, what we
• 7:45
  call an own brand retailer like a Gucci or River Island or
• 7:50
  something like that, that makes their own stuff, you have to go
• 7:52
  to get that made and you have to be able to find the right
• 7:55
  colours, the right styles. Then there's the guys who are
• 7:58
  the buy sell retailers like a Best Buy, you know mostly by
• 8:02
  name brand products from suppliers.
• 8:05
  You have meal that identify which suppliers you're missing
• 8:09
  and what things are going to be hot.
• 8:11
  And so this is really where the all of this trend analysis comes
• 8:15
  in and where AI, the particularly agentic AI is able
• 8:18
  to be able to help. But that's only number one part.
• 8:22
  Number two-part is OK, well, where does it go and, and how
• 8:26
  much should it be priced for and how should it be displayed in
• 8:31
  each store? I mean, then when you have 1000
• 8:33
  stores or something like that, that becomes a very cumbersome
• 8:36
  process. So how do you price it?
• 8:39
  What's the information coming from the competitors?
• 8:42
  How do you have pricing that's in a logical line step?
• 8:45
  And then when you look at the whole product life cycle, how do
• 8:48
  you price it so that you don't get stuck at the end with a
• 8:52
  whole bunch of overstock that you have to mark down
• 8:54
  precipitously and lose a bunch of money on because the new
• 8:57
  product is coming and you're not, you know, able to look at
• 9:00
  that? A lot of this stuff is basically
• 9:03
  human capacity. I just don't have the capacity
• 9:06
  to do all this stuff myself, right?
• 9:09
  I need an agent and a digital assistant, as it were the
• 9:13
  helping, right? And also, I don't know how to do
• 9:16
  it that well. I can have the data, but then
• 9:19
  what do I do with this data? How do I process it?
• 9:22
  How do we turn it into something that's actionable?
• 9:25
  Yeah, I think, I think for me and I maybe when we come into
• 9:28
  what the solution is, is how do I trust the data?
• 9:31
  I don't have faith that this is actually, you know, if I'm
• 9:34
  making business decisions on it, there's no, there's no
• 9:37
  hallucinations, there's no, you know, left field data thing
• 9:41
  data, broken data that is giving out a bad answer to me.
• 9:46
  So I'm curious. So tell me what, what does that
• 9:49
  look like? I appreciate you work for
• 9:52
  Intellio and obviously you guys have a, an agenda model, but I'm
• 9:57
  forgetting, forgetting in telly of one second.
• 9:59
  What do you, what do you think the agentic agent can do for me
• 10:03
  then? As a merchandiser in today?
• 10:05
  What should I be expecting from technology?
• 10:09
  Right now, you know, what it can do is it can first of all help
• 10:12
  you create a very comprehensive merchandise financial plan, you
• 10:16
  know, and that's a lot of times is done in cahoots between the
• 10:20
  CFO and the EVP of merchandising or whoever is it in charge of
• 10:24
  that. So how do I say, OK, look at all
• 10:27
  of this stuff that's going on in the world.
• 10:29
  Look, and what's been happening in my company specifically, how
• 10:33
  do I make those two things kind of come together and build
• 10:37
  something that is going to be able to take to the board if
• 10:40
  we're public to be able to, you know, you used to be able to you
• 10:43
  know, predict future performance, etc.
• 10:45
  And then how do I turn that into actionable?
• 10:48
  We call the bottoms up kinds of forecasts.
• 10:51
  So that's the first thing you could do.
• 10:53
  Now the thing about AI in this a gentig AI or generative AI is
• 10:58
  that it can reason. But more importantly, you could
• 11:01
  do these instantaneous what if scenario planning.
• 11:04
  So you can say, OK, what if I wanted to bring my margin up by,
• 11:08
  you know, a 200 basis points or 20 basis points or something
• 11:11
  like that. What are the other things that
• 11:14
  would need to change in order for that to happen?
• 11:17
  Or what happens if I say, hey, you know, I'm going to close
• 11:20
  these three stores next year or I'm going to open this or other
• 11:24
  kinds of competitive factors. This allows you to do, you know,
• 11:27
  super instantaneous modelling and then be able to kind of come
• 11:31
  up with a plan that makes a lot of sense through what this would
• 11:34
  normally take. If you had, you know, in the
• 11:36
  olden days with Excel and financial planning guys, you
• 11:40
  know, it would take weeks to be able to build these kinds of
• 11:43
  model out these scenarios where it's in, you know, it's in only
• 11:45
  in seconds. I got you and so therefore give
• 11:50
  me some outcomes. What can I?
• 11:52
  So I expect my day will be shorter in terms of because now
• 11:57
  I'm getting to the conclusion a lot.
• 11:59
  Awesome. What's the what's the outcomes
• 12:01
  that I'll be getting from this? What did the you know the the
• 12:03
  gas will expand and hit the you know, whatever the the vehicle,
• 12:08
  the vessel is that I think no way your day is going to get
• 12:12
  shorter authentic. What's going to happen is you're
• 12:14
  just going to be better, right? You know, the output's going to
• 12:17
  be better. It's going to be more accurate.
• 12:19
  The other thing about AI, as you know, Alex, is that it learns
• 12:22
  from its mistakes, right? It continually gets this through
• 12:25
  the feedback loop. And so it, you know, you can
• 12:28
  say, all right, I'm, I'm going to forecast this.
• 12:30
  And then it looks at the actual and says, whoops, you know, and
• 12:33
  then it makes adjustments with the algorithm so that it it's
• 12:36
  more accurate over time, right? Yeah.
• 12:39
  So again, I don't think this is going to replace anybody.
• 12:43
  I don't think this is going to suck the creativity out of
• 12:45
  people, you know, as that was my thing.
• 12:48
  I always thought I could buy the best deals and the hottest stuff
• 12:51
  and beat my competition. And this is even more true in
• 12:54
  the world of fashion and things that really require a lot of
• 12:57
  vision and creativity. It's not going to reduce that at
• 12:59
  all. What it's going to be able to
• 13:01
  lose just make you better, more accurate and have a lot of
• 13:04
  financial implications. Because if you're, look, if
• 13:08
  you're out ahead of your product life cycle and you are saying,
• 13:11
  OK, I'm going to mark this down 10% now so that by the end of
• 13:17
  the cycle, I'm down to 0 inventory or whatever, that is a
• 13:21
  much better financial outcome. Then if I get to the thing, I
• 13:25
  still have 5000, I need you to mark it down 50% below below
• 13:28
  cost, right? So there's immediate financial
• 13:32
  implications around not being out of stock, having successful
• 13:36
  line steps, managing your product life cycle, being very
• 13:40
  much more accurate in terms of your overall financial plan.
• 13:43
  I mean, I think it's really the future.
• 13:45
  It's per regular. Obviously you, you know this.
• 13:48
  I'm not a merchandiser, but if I was a merchandiser, what's the
• 13:52
  one question I would have asked you that I've not asked you?
• 13:58
  How does it affect well, you know, how wouldn't affect my
• 14:00
  open to buy might be one of the bits right, Because, you know,
• 14:03
  you've got and as a merchandiser, you know, a lot of
• 14:07
  people separate out supply chain from, you know, merchandising or
• 14:12
  buying or what they call Venyor management lower than Amazon.
• 14:15
  We had a vendor managers and supply chain.
• 14:17
  But the truth is, is that how you know, how can this help me
• 14:21
  bet you'd be better at managing my open to buy?
• 14:24
  Because a lot of times what happens is, is that you get
• 14:26
  stuck with core sellers and that takes away your ability or
• 14:31
  chequebook to write something for hot sellers.
• 14:34
  And then you start to really lose traction there.
• 14:37
  And so, you know, this is the kind of thing where you know,
• 14:40
  where AI, especially agentic AI, can, you know, help you get out
• 14:44
  ahead of that game. Another thing is running, how
• 14:47
  does it affect market? How can we do better promotions
• 14:51
  using AI? And I think the future of this
• 14:53
  is, you know, people say, OK, you know, we always do this
• 14:56
  every year we do a buy one, get one or buy X, get Y or whatever
• 15:01
  the, you know, these, these promotions are.
• 15:03
  And you know, is this, is this really the right promotion?
• 15:06
  How do we know that a straight discount or a coupon or
• 15:10
  something like that, is he going to have a better result, meaning
• 15:14
  that it's going to sell more, but at A at a higher margin
• 15:17
  because we're just used to doing what we're always doing.
• 15:20
  But this kinds of things can really challenge everything, it
• 15:23
  can leave no stone unturned and it can really maximise or
• 15:27
  optimise every kind of every part of the business.
• 15:30
  I got you. Okay, well listen, no my brains
• 15:33
  boggling over thinking OK all right, I got you and if I was a
• 15:39
  merchandiser I'd be thinking I need to pick up the phone and
• 15:42
  call you and and ask you well tell where we're intellio based.
• 15:46
  Oh, we're based in New York. OK, fantastic.
• 15:49
  All right. Well, listen with that, no.
• 15:50
  Thank you so much for coming on the show and I look forward to
• 15:53
  hopefully seeing you at NRS in New York.
• 15:56
  It's always a pleasure, Alex. It's nice to see you.