Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Logistics Matters with DC VELOCITY

Logistics Matters with DC VELOCITY artwork

Guest: Taivo Pungas of Pactum on balancing AI investments and risks in supply chains; Container handling conditions at U.S. Ports, What demand for machine tools means for manufacturing

Nov 13, 202619 minAudio
ShareXLinkedInFacebookWhatsAppEmail

Our guest on this week's episode is Taivo Pungas, chief information officer at Pactum. Every company is wrestling with the need to incorporate artificial intelligence into their processes. But how do you balance those needs, the costs of the implementations, and the risks involved? Senior Editor Victoria Kickham finds some answers as she welcomes this week’s guest. Our nation’s maritime ports have been humming the past few months as lots of products have been brought in for the holiday shopping season. But we saw a report this week showing that import volume at the nation’s major container ports is expected to taper off for the remainder of the year following an extended peak season that has kept docks busy since the spring. Volumes are also expected to drop in the first few months of 2027 compared to the same period in 2026. We look at what is behind these predictions Some good news this week from the manufacturing industry helped boost the U.S. economic outlook. Despite cost pressures and other challenges we’ve been discussing on the podcast, manufacturing investment continues to grow. Specifically, manufacturers and machine shops booked more than $1 billion in new orders for metalworking machinery in September, setting a record for monthly investment. That kind of investment in new machines indicates new growth in the manufacturing sector. Articles and resources mentioned in this episode: Pactum NRF: Import cargo still strong at end of tariff-extended peak season Machine tool demand hits record high Visit DC Velocity Visit Supply Chain Xchange Send feedback about this podcast to podcast@agilebme.com This podcast episode is sponsored by: Werner

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

SPEAKER_02Balancing AI investments, container handling conditions at U.S. ports, and what demand for machine tools means for manufacturing. Pull up a chair and join us as the editors of DC Velocity discuss these stories as well as news and supply chain trends on this week's Logistics Matters podcast. Hi, I'm Dave Maloney, I'm the group editorial director at DC Velocity. Welcome. Logistics Matters is sponsored by Werner. Unlock the power of a seamless connection.

SPEAKER_02Werner Intermodal combines a massive nationwide rail network and cross-border solutions into Mexico with an expanding private fleet and total digital visibility. Secure your capacity today at Choose Werner.com. As usual, our DC Velocity senior editors Ben Ames and Victoria Kickham will be along to provide their insights into the top stories of this week. But to begin today, every company is wrestling with the need to incorporate artificial intelligence into their processes.

SPEAKER_02But how do you balance that need, the cost of the implementation, and the risks involved? Victoria will find some of those answers as she welcomes this week's guest. Victoria?

SPEAKER_00Thanks, Dave. Today's guest is Tyvo Pungus. He is Chief Information Officer at Pactum, and he's here to talk to us about AI, specifically about how companies can balance investment and risk when applying AI in supply chain operations. Welcome, Tyvo.

SPEAKER_01Happy to be here.

SPEAKER_00Yeah, thanks for joining us. So I just wanted to start by asking if you can give us a brief overview of Pactum. What does the company do and what's its role in the supply chain?

SPEAKER_01Right. So Pactum serves some of the world's largest manufacturers, and we work with their procurement teams to help run the procurement process autonomously. So we give our customers digital workers that take every request from intake to a purchase order. We check the purchases against the company's policies. We do sourcing, we do negotiation, we place the order, and all of this happens 24-7, like an employee that never sleeps.

unknownOkay.

SPEAKER_00Terrific. Thank you. So we see more and more companies investing in or testing AI solutions, like you just mentioned, uh, specifically for their supply chain operations, you know, to varying levels. I wonder if you can tell us from your perspective, what are the some of the most common AI applications you're seeing in supply chain today? Maybe a couple of examples of that.

SPEAKER_01I think the most common is still the general purpose AIs. So your JGPT copilot claude. And uh and this can be used for any type of general office work. And I'm sure most of the listeners here will have used some of them, uh, putting together a quote or um you know working on a on an Excel sheet, drafting some communications, an email. Uh but the limitation on all of this is it is still fundamentally driven by a human.

SPEAKER_01So it's it's an amplifier, um, but there are limits to how far you can get with that.

SPEAKER_00So that that leads me into sort of, you know, we hear a lot of different um terms thrown around with AI, and I wondered if you can talk about the difference between those things, things like agentic AI, autonomous AI. Can you just give us a little bit about each one of those and what what they actually mean?

SPEAKER_01Unfortunately, the terms have all been a bit butchered. So everyone's saying they have agents. Doesn't matter really what the product is, but um but you know you you can try and and distinguish two broad clusters of things. One is and and and we can use the self-driving analogy. So if you're driving a car, you may have quite advanced driver aids, but you still have to hold the steering wheel and pay attention to the road, um, you're still driving. And that's kind of one big cluster. You are driving Claude, you're driving ChatGPD, and even though it does work for you, you know, it won't act on its own.

SPEAKER_01You have to remember to start it. And the other cluster is uh, you know, a car driving you from one place to another autonomously, you don't really even have to remember to order it. You know, it it comes to you based on your calendar at the right time and in the right place, and then takes you to where you need to go. And that that is your you know, more the peck-dum model, the thing that we do uh and and other companies do, where kind of like if you hire an employee, they will do work, they will do their tasks, and it's kind of expected that when you hire someone, they will do their job without you having to like tell them every five minutes to go and do the next task.

SPEAKER_01And uh and that's that's the other broad cluster. So this the second cluster, this is what what we call autonomous, um, where it has broad capabilities, broad authority to act uh over a long period of time and doesn't really require a human to be you know starting and observing um any any small task that it does.

SPEAKER_00Thank you. Um so with any of this, uh you know, any of these investments there comes with risks, especially I think, or the perception is around these autonomous AI solutions. There have been recent um examples of you know systems you know going rogue or making decisions uh that maybe the humans involved didn't you know necessarily want them to make. I wonder if you can talk about some of the risks and how companies can ensure they have the right level of human oversight involved with AI projects.

SPEAKER_01The the headlines have been about the things that I think are don't really um touch or are not really relevant for the companies that uh uh that are not themselves developing very you know cutting-edge agents necessarily. Um the the the I think the more practical risks uh not so so it's not about um an AI breaking out of a box and accidentally hacking a third party, like that that is not the risk that most people have to worry about.

SPEAKER_01Um what is more likely is that you put you give an agent a task and then it goes off and does something that kind of you wouldn't do. Uh and again, if if you if you you know if you're not doing your hiring well, if you hire a human and you don't train them well or you don't pick them well, uh that the same thing may happen. It's just that um AI LLMs tend to have a bit less common sense, a bit less uh you know shared shared background, shared uh experience. So um it's more surprising when they do that. So they might send an email accidentally to the wrong person or with the wrong sort of um of sensitive content.

SPEAKER_01Uh that is more the risk. So an agent taking action that it is in principle allowed to do, just making a mistake about the context or or the content uh that it puts or that it that it um uh chooses to sort of attach to this action. And that's the more practical risk. Uh and the solution to that really is to pick a vendor, pick pick products and tools that are that are intentional about it, that have enough human oversight, human involvement that the product is designed to you know include the best of what the human j can judge, and not just let an agent uh you know go off on its own and and basically have no oversight.

SPEAKER_01So it should be a design goal of the um of the product of the system that you bought.

SPEAKER_00So it sounds like you know, work with companies that know your business, understand what you're trying to accomplish, that kind of thing, uh, and and sort of incorporate that into their the tools that you'll you'd be using, is that right?

SPEAKER_01Correct. And and specifically they understand, you know, what is a high-risk action, what's a lower risk action, right? So sending an email in uh, you know, if you're sending a marketing campaign, sending it to the wrong person, it's not very harmful. But sending uh a detailed um specification of a part to the wrong wrong supplier, or uh that may be much more harmful. Uh you know, you're you're you're sharing information that you're not supposed to. Uh so it's quite contextual and you should work with companies that go deep and understand what you do in a way that they understand what is what's important, what's high risk, what's low risk.

SPEAKER_00That makes a lot of sense. I wonder um you know what advice you'd give to companies as they navigate the AI landscape, in addition to what we just talked about, you know, uh, you know, what would you tell um you know companies or are you know looking into either just starting out on this AI journey or those who are already involved in using different AI tools?

SPEAKER_01So probably the biggest um biggest constant these last couple of years is that the pace of change is is crazy fast. Uh new models and new products keep coming out. So um I would I would aim to be curious and try out many things. And I know I am probably as a technologist, I'm kind of more keen on doing that. Uh, but then also keep optionality. So so today I would not commit to a five-year contract, not even a three-year contract. Even one year is kind of a long time in today's world to commit to a particular vendor or a particular solution.

SPEAKER_01So uh so it is a time of rapid change, and I would I would recommend approaching it with curiosity and kind of keeping your options open a lot.

SPEAKER_00Any other um issues or trends occurring in this this AI world that you'd you'd tell our listeners to uh to pay attention to? Any any it's kind of a crystal ball question I'm asking you, but any you know anything that they should be aware of as this landscape sort of unfolds?

SPEAKER_01The most interesting thing that I think is not widely appreciated is the pace at which the models not just get are getting better in the sense of being more intelligent, but also how fast they are getting cheaper. So the best model a year ago costs today, like the equivalent model today um that would have been top of the world 12 months ago costs about 100 times less today, uh which is crazy. So so if if it seems that some use case of AI today is kind of frivolous or um or it's it's a bad use of tokens, it's a bad use of money, it could well be that in a year like the economics have changed.

SPEAKER_01So it is now actually a good idea to do that. And and that that comes very surprising to um even to myself, honestly, because uh if if if I bring an analogy, um the price of uh like a high-end luxury car, so let's say one year ago, if it comes down 100x, you could get um that car for$2,000 today, let's say 200k, 200$2,000. And uh like in nowhere in in the physical world do we actually experience that kind of pace of progress, and that's why it is quite unintuitive.

SPEAKER_01Um, so things are changing rapidly both on the capability and the cost front.

SPEAKER_00Yeah, the speed of change is is is staggering. You're right. It's it's hard to keep up with. Tyvo, thank you so much for joining us today. We really appreciate your insight and expertise on this.

SPEAKER_01Thank you, Victoria.

SPEAKER_00Yeah, we have been talking with Tyvo Pungus of Pactum. Back to you, Dave.

SPEAKER_02Thank you, Tyvo and Victoria. And now let's take a look at some of the other supply chain news from the week. And Ben, our nation's maritime ports have been humming for the past few months as lots of products have been brought in for the holiday shopping season. And you wrote this week about the current conditions at the ports and what to expect for the next few months.

SPEAKER_03Uh yeah, humming is exactly right. We saw a report this week showing that import volume at our nation's major container ports is finally expected to taper off for the remainder of the year following an extended peak season that has kept those docs really busy since the spring. That assessment comes from a monthly report released by the National Retail Federation or NRF and Hackett Associates. So, why have we had this extended peak season?

SPEAKER_03The answer is a single word tariffs. That's because this series of unpredictable tariff increases from the White House led retailers to rush to bring in their merchandise earlier than usual to be able to do so cheaper before the threatened tariffs could take effect. And if you remember, sometimes those tariffs were often delayed at the last minute and then imposed anyway later on, which kind of started the whole cycle over again. In the words of NRF Vice President for Supply Chain and Customs Policy Jonathan Gold, he said U.S.

SPEAKER_03ports have seen high volume levels throughout this year's longer than usual peak season, and the numbers are still strong even as the season winds down. Unfortunately, he said these volumes have led to backups at some ports, particularly on the West Coast, where cargo has sat for extended periods waiting to be transferred to rail cars. Another complication on the West Coast is the possibility, uh, as we speak, of a strike by certain dock workers, specifically the security watchmen at the ports of uh Los Angeles and Long Beach.

SPEAKER_03At last check, although the union had authorized a strike, that action had been delayed until November 20th to give more time for negotiations.

SPEAKER_02Right, Ben. And we've seen several times in the past that any kind of a strike at the docks can seriously impact entire supply chains. So we hope that we won't see any work stoppage or that can all be avoided.

SPEAKER_03You're definitely right about that. We've done a lot of coverage uh when that has happened in in the past. It puts a big speed bump in supply chains. So hopefully those negotiations go well for both sides. But the NRF also put a positive spin on some of these challenges. Uh they did say, quote, nonetheless, most holiday merchandise is already here and retailers are working closely with their supply chain partners to ensure that the supply of holiday gifts continues to flow smoothly. So that assessment is good news for everybody's personal holidays.

SPEAKER_03And relief could also come because at some point uh this import rush has to slow down. So the report from NRF said that uh the September container import volumes, for example, were a little more than 12% above the same month last year, October about nine percent higher than last year's October. And then the forecast is that things will quickly slow down. So the forecast is that November will be up just one percent, same with December, and then January, February, March will all be down in negative numbers, with volumes actually shrinking compared to the previous year.

SPEAKER_03So that should take the pressure off.

SPEAKER_02It should. And as you mentioned earlier in your report, that a lot of that volume we saw in the early months of the year, um, the current year, have were because of those tariffs. Um, trying to avoid some of those. We won't see those higher volumes again, probably this year, but then again, anything can happen in this environment. So who knows?

SPEAKER_03It sure can. It keeps us all busy.

SPEAKER_02It certainly does. Thank you, Ben. Glad to And Victoria, we've written in DC Velocity this year about new investments that have been made in manufacturing operations. And you found some new examples that illustrate this bit of a U.S. manufacturing resurgence.

SPEAKER_00That's right, I did, and there was definitely some good news this week from the manufacturing industry, and it helped to boost that U.S. economic outlook. Despite cost pressures and other challenges we've been discussing on the podcast in the past few weeks. As you mentioned, Dave, manufacturing investment continues to grow. Specifically, uh manufacturers and machine shops booked more than a billion dollars in new orders for metal working machinery in September, setting a record for monthly investment. Now the information comes from the latest U.S.

SPEAKER_00Manufacturing Technology Orders Report, that's also known as the USMTO report, released Monday by AMT, the Association for Manufacturing Technology. So this report tracks new order values and machine units for metal cutting and metal forming machinery. It serves as an economic indicator for future manufacturing activity as companies invest to support expected production needs. September's numbers continue an upward trend in machinery investment, marking the seventh straight month of orders exceeding half a billion dollars.

SPEAKER_00And that's the longest streak of investment in the report's history, according to AMT.

SPEAKER_02Well, that is good news, and it seems to indicate the U.S. economy is headed in the right direction, but how much of an improvement did the report track?

SPEAKER_00Well, September orders rose 38% compared to August, and they were up nearly 102% from a year ago. The total value of orders placed in 2026 uh so far reached$5.76 billion, and that's 45% above the year-ago period. So some pretty strong increases in there. September's totals were also nearly uh 23% above the previous monthly record, which was set in December of 2025. And orders accelerated across customer industries, including machine shops, aerospace manufacturing, and automotive manufacturing.

SPEAKER_00The other good news is that AMT calls for elevated activity for the remainder of this year and into 2027. Also, this news came on the heels of some other recent positive signals that we've seen. North American robot orders, for example, and those indicate demand for factory automation, were up 25% in October to reach their highest levels this year. That's according to the latest information from A3, the Association for Advancing Automation, which is a group we cover frequently.

SPEAKER_00And of course, we saw uh a jobs report last week that beat expectations. Private sector hiring was surprisingly strong, especially in professional services and, of course, manufacturing.

SPEAKER_02Well, that is all good news, and we hope it signals better things ahead for 2027.

SPEAKER_00Yes, always looking for the silver lining here.

SPEAKER_02Right. Thank you, Victoria.

SPEAKER_00You're welcome.

SPEAKER_02We encourage listeners to go to dcvelocity.com for more on these and other supply chain stories. Also, check out the podcast notes section for some direct links to read more about the topics that we discussed today. And we'd like to thank Tyvo Pungas of Pactum for being our guest. We welcome your comments on this topic and our other stories. You can email us at podcast at agilebme.com. We also encourage you to subscribe to Logistics Matters at your favorite podcast platform. Our new episodes are uploaded on Fridays.

SPEAKER_02And a reminder, Logistics Matters is sponsored by Werner. Unlock the power of a seamless connection. Werner Intermodal combines a massive nationwide rail network and cross-border solutions in New Mexico with an expanding private fleet and total digital visibility. Secure your capacity today at Choose Werner.com. We'll be back again next week with another edition of Logistics Matters. Be sure to join us. Until then, have a great week.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Guest: Maxime Darmet of Allianz Trade on the economy after the election; High hopes for the holiday season; What it takes to have an exceptional healthcare supply chain

    Nov 20, 202624 minAudio

    Our guest on this week's episode is Maxime Darmet, senior economist at Allianz Trade. With the determination earlier this week of the Alaska Senate race, the midterm elections in the U.S. are now in the books. It was an...

    Credit the people on this episode

  2. Guest: Rupal Deshmukh of Kearney on how the election results may affect supply chain costs; A look at supply chains of the future; The latest health of the logistics industry

    Nov 6, 202618 minAudio

    Our guest on this week's episode is Rupal Deshmukh, Kearney partner and global lead of the Kearney Supply Chain Institute (KCSI). The midterm election is over and the people have spoken, signaling what could be dramatic...

    Credit the people on this episode

  3. Guest: Rebecca Brewster of the American Transportation Research Institute on trends in trucking; How will the midterms affect infrastructure funding? MIT reports on industry sustainability

    Oct 30, 202621 minAudio

    Our guest on this week's episode is Rebecca Brewster, President & COO, American Transportation Research Institute. The trucking industry has recently faced several rocky years, from an ongoing freight recession to the...

    Credit the people on this episode

  4. Guest: Vikram Sekhon of Piston on preventing fuel fraud; Seeking greener freight transport; Dual fuel ships sail ahead

    Oct 23, 202623 minAudio

    Our guest on this week's episode is Vikram Sekhon, CEO and co-founder of Piston. Diesel fuel costs have nearly doubled this year, largely driven by the impacts of the war with Iran. Whenever costs get more expensive,...

    Credit the people on this episode

  5. Guest: Chrissy Sailstad of Durham Exchange Club Industries on hiring people with disabilities; How Walmart approaches automation; Key trends affecting global shipping through 2027

    Oct 16, 202620 minAudio

    Our guest on this week's episode is Chrissy Sailstad, CEO, Durham Exchange Club Industries. October is National Disability Employment Awareness Month. This annual event was begun in 1945 by the U.S. Department of Labor...

    Credit the people on this episode