Latest / Beyond Fulfillment Podcast / How Failure Fueled a $700K Supply Chain Win | Kelly Breakstone Roth

Transcript

How Failure Fueled a $700K Supply Chain Win | Kelly Breakstone Roth · Beyond Fulfillment Podcast

• 0:00
  investors are chasing you. I mean that had to feel good after.
• 0:02
  my god. I literally had a call that started with this is a shut up and take
• 0:07
  my money kind of call. Which I've heard about before.
• 0:10
  Never thought that it would come.
• 0:12
  so that was really nice.
• 0:13
  building a business, building a startup,
• 0:15
  the whole
• 0:15
  family is involved and my daughters as well.
• 0:17
  They've been very encouraging and very supportive.
• 0:20
  And I feel that also part of my job as a mother is to be that role model
• 0:24
  for them and show that you know you fall,
• 0:26
  but you get up and you try again,
• 0:28
  you try again. And so that was the wind beneath that kept on pushing me forward.
• 0:37
  All right, Dave Gulas back here at Shop Talk,
• 0:40
  and I am with Kelly Brakestone Roth,
• 0:42
  who is the co-founder of Prysmic.
• 0:45
  Hi Kelly. Hi. Good to be here.
• 0:47
  Yes, glad you could make it. for the audience,
• 0:50
  tell us what is Prysmic? Prysmic is the operating system for profitable growth.
• 0:54
  So what we're seeing is that supply chain leakage is basically growth capital just
• 1:00
  hiding in plain sight. And what we do is we deploy fleets of AI agents across
• 1:05
  the company's supply chain from procurement to inventory to logistics to finance.
• 1:09
  We find the areas where margin is bleeding,
• 1:12
  where capital is tied up, where sales are lost,
• 1:15
  and then our AI agents go and fix it.
• 1:17
  They actually act within a company's stack,
• 1:19
  they update plans, they chase suppliers.
• 1:22
  They reallocate inventory, they manage and dispute overcharges.
• 1:27
  So really just helping get that revenue,
• 1:30
  that capital growth back to the hands of the business.
• 1:33
  that sounds scary when when an AI agent is chasing I I would hate
• 1:38
  to be that supplier that a AI agent is chasing.
• 1:41
  Right, because they're relentless.
• 1:42
  They're relentless. And they don't forget.
• 1:44
  Yeah, I bet. I bet. Okay. Well,
• 1:46
  I love the vision. I love how you explain that.
• 1:48
  I'm in.
• 1:49
  Right on the idea. So tell what what made you decide to start this?
• 1:52
  So I've spent most of my career inside complex operations.
• 1:56
  first for the first part of my career I was working for Israel's intelligence,
• 2:01
  I was a lieutenant colonel, and I was leading cross-border operations with heavy
• 2:05
  logistics involved and heavy technology.
• 2:07
  So that's where I kind of encountered this world in the first place.
• 2:11
  And then about six years ago, I took the leap into the entrepreneurial world.
• 2:15
  I started a company which I loved.
• 2:17
  I was pursuing a childhood dream.
• 2:18
  It was a global marketplace connecting top-tier artisans in the developing world
• 2:23
  to the US consumer. So working with hundreds of suppliers from Morocco
• 2:26
  and Mexico and Colombia, managing all their supply chain end-to-end,
• 2:30
  everything from international shipping to customer support,
• 2:33
  refunds, and returns. It was a very complex endeavor,
• 2:36
  as you can imagine.
• 2:37
  we were a very small team at the time,
• 2:38
  so we kind of had to build a lot of automations and we were using
• 2:41
  the tools that were available to us at the time,
• 2:43
  mostly RPA tools like Zapier, which helped us a lot,
• 2:46
  but RPA tools are not perfectly fitted for the world of supply chain,
• 2:51
  which continuously breaks. Like in supply chain,
• 2:53
  20% is taking the happy path, and about 80% is exceptions and edge cases
• 2:58
  and things that just get really,
• 2:59
  really messy along the way. And so these automations that we had kept on breaking.
• 3:03
  And it consumed all of our time and our attention,
• 3:06
  and it really ended up being a ceiling to our company's growth.
• 3:09
  ultimately we closed the company.
• 3:10
  the days of post-COVID, we weren't able to bring in enough capital
• 3:14
  to take the company to the next level.
• 3:15
  But bearing those scars, I looked to see how we can change this,
• 3:20
  how AI can actually move those constraints.
• 3:22
  And I was really fortunate to meet my co founder.
• 3:25
  and CTO Dana Geva, who is the most brilliant human on this planet.
• 3:29
  And then all of a sudden all these ideas became something that we could reach
• 3:33
  and that could really change this industry.
• 3:35
  And so that's how I got here. Okay.
• 3:37
  Wow. Amazing. Amazing. So much to talk about there.
• 3:40
  Real quick, you said it was a childhood dream to be an entrepreneur.
• 3:43
  So t talk about that more if you would.
• 3:46
  So it's not the childhood dream
• 3:48
  dream to be an entrepreneur. I think I was very entrepreneurial from the
• 3:50
  day I was born and I see that in my daughters.
• 3:52
  There's you know some are more entrepreneurial than others.
• 3:54
  It's just I think something that you're born with.
• 3:56
  The dream was being able to connect with incredible makers around the world.
• 4:01
  there's an artist in there. I do pottery.
• 4:04
  I love these things. I was traveling the world a lot.
• 4:06
  The thing that I love doing more than anything else was kind of getting lost
• 4:10
  in the streets and finding these beautiful
• 4:12
  little you know craft shops and then bring those products home with me.
• 4:16
  so the dream was how do we create access to this magnificent talent around
• 4:20
  the world and make it accessible to everybody.
• 4:22
  So that was the dream that I was pursuing.
• 4:24
  you know it was incredibly exciting and beautiful and inspiring,
• 4:28
  but it was also very difficult to manage supply chain across multiple continents
• 4:32
  with hundreds of suppliers.
• 4:34
  Many of them not near the main cities.
• 4:36
  I mean that was quite an endeavor to take.
• 4:37
  So you meet your co-founder, the most brilliant person I I've I have
• 4:43
  to meet this person now. She's exceptional.
• 4:45
  Okay, all right. And then so it was you realize that all these ideas
• 4:49
  and these visions that you had you could make a reality.
• 4:52
  so
• 4:52
  What's it like? Because a lot of people have ideas,
• 4:55
  right? They never get out of their own way.
• 4:56
  They don't start, fear like obviously you'd been in business before,
• 5:00
  like you said, you were bearing the scars and the ultimate failure of
• 5:03
  of that e commerce brand kind of gave you the the idea for what Prysmic became.
• 5:08
  So like what was the process like when you said,
• 5:10
  Okay, we're gonna do this and you just you just jump in and you start
• 5:13
  a company like this where you're serving like nine figure brands?
• 5:17
  So first of all I'll say it it not only left the scars of what it took
• 5:20
  to build that out, but I was left with a very big chip on my shoulder.
• 5:24
  Not being able to take that first company and turn it into something that
• 5:27
  I I had hoped it would be. and so when I closed the company,
• 5:30
  I was very driven to figure out what else I could do,
• 5:32
  how I can take those learnings and turn it to something useful and something that
• 5:36
  I would be proud of.
• 5:37
  So I started co-founder dating.
• 5:39
  I like for two years just met anybody I could who was interested
• 5:42
  in building something and try to find those areas that we had in common,
• 5:46
  those pains that we both shared.
• 5:47
  and then I met Dana. And Dana and I met it's probably about two and
• 5:51
  a half years ago. And we didn't build Prysmic from the first moment.
• 5:54
  We tried out a couple of different ideas,
• 5:56
  building out MVPs, speaking to potential customers,
• 5:58
  speaking potential investors.
• 6:00
  And every time deciding, okay,
• 6:02
  that's not what we're gonna be investing the next five,
• 6:04
  ten years of our life in, and putting that back into the drawer,
• 6:07
  which is by the way, a very excruciating process because every time
• 6:10
  you have this idea, you get really excited,
• 6:12
  you put effort into it, and then you're like,
• 6:14
  no, that's not it. And then we came up with the idea of Prysmic
• 6:17
  and everything just kind of clicked into place.
• 6:19
  Like investors came after us. we were able to get design partners very,
• 6:23
  very quickly to build the system together with them.
• 6:26
  and then we were able to find customers.
• 6:28
  So it was a really long journey.
• 6:30
  It wasn't like I woke up one day and said,
• 6:31
  this is the best idea and I'll pursue it.
• 6:33
  There was a lot of trial and error there.
• 6:35
  Okay, wow. So appreciate all that context.
• 6:38
  And so like you said, that that ultimate e-commerce failure,
• 6:41
  you used it as fuel for the next like so many people even in that position might
• 6:45
  have just given up, went to get a job and just said,
• 6:48
  This isn't for me, but you said,
• 6:49
  No, I'm gonna build something.
• 6:50
  And then top of that too, you spend
• 6:53
  next couple of years just trying to figure out what it's gonna be,
• 6:56
  finding the right co-founder, testing different things out.
• 6:58
  Like when you're going through that whole process,
• 7:00
  was there ever a point in time you thought about giving up?
• 7:02
  every day. ~ absolutely I think that I'm first of all I'm very privileged.
• 7:09
  I have a husband who's very supportive of this journey and kept on pushing
• 7:12
  me and just, you know, reassuring me that it's gonna work out.
• 7:15
  So just you know building a business,
• 7:17
  building a startup, the whole
• 7:19
  family is involved and my daughters as well.
• 7:21
  They've been very encouraging and very supportive.
• 7:24
  And I feel that also part of my job as a mother is to be that role model
• 7:28
  for them and show that you know you fall,
• 7:29
  but you get up and you try again,
• 7:31
  you try again. And so that was part of like I guess the the wind beneath that kept
• 7:36
  on pushing me forward. Okay. Okay.
• 7:38
  Amazing. All right. So when you finally you you link up with Dena,
• 7:42
  you start Prysmic
• 7:44
  Everything starts falling into place.
• 7:45
  investors are chasing you. I mean that had to feel good after.
• 7:48
  my god. I literally had a call that started with this is a shut up and take
• 7:53
  my money kind of call. Which I've heard about before.
• 7:56
  Never thought that it would come.
• 7:57
  so that was really nice. I you know,
• 7:59
  I think there's also a lot of luck involved.
• 8:01
  Like we came across this idea at a time where everybody was looking to
• 8:05
  see how they can implement AI.
• 8:07
  when I built the marketplace, the the Nopo,
• 8:09
  my previous company, we knew from the first day that we were going against
• 8:12
  the current. Like marketplaces were starting to die off,
• 8:15
  people, you know, it it just wasn't the right time to be building that
• 8:18
  and we kept on feeling that negative,
• 8:20
  like vibe wherever we went. And here it was very different.
• 8:23
  So, you know, just finding the right idea at the right time is huge.
• 8:26
  I think that's gonna be the episode title,
• 8:28
  Shut Up and Take My Money.
• 8:34
  That's amazing. Okay. All right.
• 8:59
  So now you got the the funding,
• 9:01
  you got the partners, and you even got customers and you just celebrated your first
• 9:06
  birthday, is that right? Okay.
• 9:07
  So company's one year old, so still very new.
• 9:10
  So what's it like with something that new and then you're heavily leveraging
• 9:15
  AI and the concept sounds phenomenal,
• 9:17
  but I can only imagine there's a lot of still complexity and when you're dealing
• 9:21
  with such big customer you know what I mean,
• 9:23
  you're dealing with enterprise level customers right off the bat as a new company.
• 9:27
  What's that process like? I mean first of all it's it's very different,
• 9:29
  like it depends who you meet at a specific company.
• 9:33
  it can be very easy, it can be very quick.
• 9:35
  Like our sales cycle is anywhere between three weeks and seven months.
• 9:39
  So it really depends on a lot of things that are exterior to you,
• 9:43
  exterior to what you're offering.
• 9:44
  again, luck, timing, finding the right ICP inside the company.
• 9:48
  and you know, what we were selling on day one is not what we're selling today.
• 9:52
  So your story keeps on evolving.
• 9:54
  Understanding the needs changes all the time and we spoke about this before,
• 9:58
  someone shared with me that the
• 9:59
  The surface of knowledge from a distance looks very smooth and it seems very clear.
• 10:03
  But the closer you get to a problem,
• 10:05
  the more edges you see, the more complex it becomes.
• 10:08
  And it's the same thing here. So,
• 10:10
  you know, we had this brilliant idea and we started working on it,
• 10:13
  and the closer you get to the problem,
• 10:14
  the more complex you see it is,
• 10:16
  but also it kind of encourages that the problem is so big and so deep that
• 10:20
  it just motivates you to continue building towards the solution.
• 10:24
  Okay.
• 10:25
  All right, and so even in this first year,
• 10:27
  with the company and some some of the brands you've dealt with,
• 10:30
  what's been like some of the biggest wins that you've experienced
• 10:33
  for for the customers? like KPIs that we've achieved for them.
• 10:36
  Yeah. ~ I probably like the the most standout number is ~ nine figure multi brand
• 10:42
  company where we were auditing all of their freight invoices and we were able to
• 10:48
  capture seven hundred thousand dollars of overcharges.
• 10:51
  So that's like that's money that you you know goes straight into growth capital.
• 10:55
  That's a huge number. But across the board we're seeing anywhere between
• 10:58
  ten to twenty percent of recoverable margin.
• 11:02
  So again, freight forwarding, customs,
• 11:04
  suppliers, carriers, three PLs,
• 11:06
  and then inventory that we're
• 11:09
  fixing the projection by more than thirty percent.
• 11:12
  So that's can you can imagine how much capital is tied into into that.
• 11:15
  Yeah, and so talk more about that too,
• 11:17
  like how the AI can streamline the inventory I guess the inventory control,
• 11:22
  the demand planning, the forecasting,
• 11:24
  right? And you have these agents placed in all because you were telling
• 11:28
  me to the channels that all this happens are not necessarily in the ERP,
• 11:32
  they're in email, they're in Slack.
• 11:33
  So you put agents everywhere.
• 11:35
  And then they basically collaborate in So our agents are able
• 11:39
  to capture thousands of signals across the supply chain.
• 11:42
  And they're sitting inside all of your systems,
• 11:45
  whether it's your ERP, your TMS,
• 11:47
  your IMS, your partner portals,
• 11:49
  but it's also sitting where operation activity actually happens
• 11:52
  off system and cross-system. So in emails,
• 11:54
  share drives, Excel sheets, Slack,
• 11:57
  WeChat. And so we have expert agents for every type of role,
• 12:01
  responsibility, and task.
• 12:03
  And they're just sitting there waiting for basically for an event to happen
• 12:07
  for some sort of trigger. All these agents are sharing this I would
• 12:10
  say like operational cortex and operational brain.
• 12:13
  So once a new signal enters or or one of the agents receives one of these signals,
• 12:17
  it automatically sort of sends a message to other agents across the supply chain.
• 12:22
  So everybody understands how that will impact them.
• 12:24
  So when the shipment is delayed.
• 12:26
  the impact for this goes all the way downstream for marketing,
• 12:29
  for something that you're holding for a retailer.
• 12:31
  So you need to be able to pass that signal along,
• 12:34
  understand, interpret the impact across the supply chain and then come back with
• 12:38
  the best course of action. and I think that one of the most important things that
• 12:43
  we learned is that the quicker you can move from signal to decision,
• 12:46
  from signal to action, the better options you still have on the table.
• 12:50
  So that speed of response is not just an efficiency metric,
• 12:54
  it's an economic metric. Because if something happened in reality that impacts
• 12:57
  the plan and now I can make a better or faster decision,
• 13:01
  I'm gonna have better options that will save my margin in the long term.
• 13:05
  Now that's interesting too because so often,
• 13:07
  These massive companies, right?
• 13:08
  There's so much bureaucracy and you hear that decisions happen slow
• 13:12
  and some people it's like it's not my problem or it's not my job.
• 13:17
  So now that you have these agents everywhere that are communicating
• 13:21
  and you get these signals and it's happening in real time,
• 13:23
  have you seen a big change in how you can motivate action based on that?
• 13:27
  100% because this is not just based on someone's intuition or tribal knowledge.
• 13:32
  every decision or every recommendation has a full audit trail,
• 13:36
  comes with full evidence. Now everybody's speaking the same language.
• 13:39
  And what we see a lot of times is different stakeholders throughout
• 13:41
  the organization. They're seeing different parts of the picture.
• 13:44
  A lot of the systems are showing different parts of the picture,
• 13:47
  sometimes conflicting parts of the picture.
• 13:49
  And so once that agent has reconciled all that data,
• 13:52
  put it together, resolved the contradictions,
• 13:55
  it is a lot easier to get stakeholders to work
• 13:57
  together.
• 13:57
  But I'll still say, like bringing in an AI system like this still is very much
• 14:01
  a political issue. And when you decide on what you want to deploy and what your KPRs
• 14:06
  are and what you want to achieve,
• 14:07
  you have to understand that you have to have all the stakeholders
• 14:09
  in the room together. We've very early on kind of suffered from that
• 14:13
  by building something out that operations wanted and then finance came in and said,
• 14:18
  no, that's not the way we wanted to do.
• 14:19
  So all our work was okay, who's the boss of this AI?
• 14:22
  Who determines
• 14:23
  What it should be doing, how you evaluate its success and so forth.
• 14:26
  Have you noticed too, are people apprehensive,
• 14:29
  like even at the executive level in terms of people protecting their like thinking
• 14:33
  that they're gonna lose their jobs if this AI system comes in,
• 14:36
  that type of thing? Have you seen that?
• 14:37
  it's there, but it's not there.
• 14:40
  I mean it hasn't really been a barrier.
• 14:42
  We have found
• 14:44
  I would say like more like entry level jobs that are a little bit more concerned
• 14:47
  and won't necessarily collaborate because there's this QA process.
• 14:50
  You want to make sure like one of our first KPIs is fidelity.
• 14:53
  So it's like making sure that everything is not only accurate
• 14:55
  but really perfectly imitates the business logic of the organization.
• 14:59
  And so you need different stakeholders to actually evaluate what the agent
• 15:02
  is coming up with. And we've had some objection from more entry level positions,
• 15:07
  but once
• 15:08
  we really start working with them and they see the benefit,
• 15:10
  then we win them over. part of it it's just like building trust that
• 15:13
  you understand that we're there to work with you rather than against you.
• 15:16
  And even like at the CEO or COO levels,
• 15:18
  these AI agents could theoretically eliminate certain jobs,
• 15:22
  but that's not what anybody has in mind.
• 15:24
  No one's looking how can I cut my staff.
• 15:26
  They're looking at how I can, you know,
• 15:29
  build this operational leverage,
• 15:30
  how I can grow better, how I can grow faster.
• 15:33
  So yes, there's some objection there,
• 15:35
  but it's not major. Okay. All right.
• 15:38
  And so thinking back on the journey too,
• 15:40
  where you've had this dream to kind of be out there building
• 15:42
  and connecting with other makers.
• 15:44
  You had the e com failure, you went through the process of years trying
• 15:47
  to find the right thing. You meet your co founder,
• 15:49
  you build Prysmic and you're you're a year in and things are going great
• 15:53
  and you're you're really excited about the future.
• 15:54
  What what's been the best part of this journey for you thus far?
• 15:58
  couple of things. First all working with my co-founder.
• 16:00
  we're very close friends. I admire her.
• 16:02
  think one of the things that I kind of wished for myself before starting this
• 16:05
  company was to find the best people to work with.
• 16:08
  And so that has been incredible.
• 16:10
  And I think you know we're all sort of navigating this shift into this
• 16:14
  new era and we don't know what exactly it's gonna look like.
• 16:17
  So
• 16:18
  You know, being able to kind of drive the car and be part of the I would
• 16:23
  say the companies that are leading this transition has been incredibly cool.
• 16:27
  you know, I'm forty-six years old,
• 16:29
  being part of this thing that is like changing the way operations are gonna
• 16:32
  be run in the future is is pretty insane.
• 16:34
  I love that. and one of the funnest things that we've done,
• 16:37
  like really from day one, was start building these communities or community events.
• 16:41
  Bringing together either technology companies or brands to really discuss,
• 16:46
  you know, what this shift looks like,
• 16:49
  learn from each other. So we do a lot of dinners,
• 16:51
  we've done panels, so that has been really great.
• 16:54
  Just like talking to people, getting like energized by by the
• 16:57
  the overall like excitement has been really good.
• 17:26
  Amazing. Okay. All right. And Kelly,
• 17:28
  if people want to find out more about Prysmic,
• 17:30
  where do they go? getprysmic.ai and please follow me on LinkedIn,
• 17:35
  Kelly Breakstone Roth
• 17:37
  we create a lot of content, so we have what we call our ops AI Academy.
• 17:40
  So it's also interviews with supply chain leaders and learning about
• 17:44
  how they're managing this shift and then also just like general insights
• 17:48
  and learnings that we've had along the way.
• 17:50
  So definitely look in our blog.
• 17:51
  Okay. We'll link all that in the show notes.
• 17:53
  Kelly, thanks so much. Thank you.
• 17:55
  This was fun.