LSPs are under more pressure than ever to deliver a seamless, Amazon-level customer experience across fragmented networks, multiple carriers, and dozens of data formats. But the tools and AI investments they are turning to can’t solve that problem because the data underneath them is inconsistent, unstandardized, and siloed. They are trying to build on a foundation that does not exist yet.
Jamie Andrade’s expertise sits exactly at that intersection. She’s the SVP of product management at SEKO Logistics and spent nearly a decade running operations teams before moving into product and technology leadership. So she understands what LSP systems can and cannot do.
Her argument throughout this episode of Deliver: The Last-Mile Performance Podcast is that the industry keeps reaching for the next tool when the actual problem is one layer below: the data model itself. Hosts Bringg CEO Guy Bloch and Sales Engineering Lead Raquel Zanoni sit down with Jamie to unpack why integration stays broken at most LSPs, what scalable last-mile technology looks like in practice, and why AI investments keep underdelivering without a clean data foundation underneath them.
Key takeaways
LSPs are not failing on capability. They are failing on data and integration
- The gap between what customers expect and what most LSPs deliver is a data consistency problem, not an execution problem. A single shipment can touch a first-mile carrier, a line haul, a customs broker, and a warehouse, with no standardized data layer to reconcile them.
- LSPs that close this gap become partners in their clients' supply chains. LSPs that do not become vendors who get replaced when something goes wrong.
Dynamic flexibility is now the survival requirement
- Automation, dynamic routing, and real-time exception management are no longer capabilities that set LSPs apart. They are the baseline for staying competitive.
- Clients do not want to log into a portal to find out where their freight is. They expect proactive communication built into the workflow, and that expectation is only going to increase.
Scalable last-mile technology means more open configuration
- Scalability means building systems where operations teams, not developers, can edit mappings, create events, and manage carrier connections without submitting a ticket.
- The LSPs that never stop re-architecting are the ones that started with a pilot site instead of a standardized operational process.
Most LSPs are not ready for AI, and buying tools before fixing data makes it worse
- LSPs that find real results with AI identified internal pain points first, then found the tool. Those that start with the vendor demo consistently fail to get value.
- The near-term AI wins are unsexy: delay prediction, automated exception management, and route optimization. But, they’re the investments that build the data foundation needed for anything more ambitious later.
Data is a product that has to be maintained, not a by-product of operations
- Treating data as something that emerges from a shipment, rather than something that has to be actively governed and standardized, is the root cause of most LSPs’ AI underperformance.
- LSPs that build the data model correctly from the first brick find that every new tool and carrier they add feeds into a platform that already knows how to handle it.
How are macroeconomic pressures changing how LSPs think about last-mile strategy?
The pressures hitting LSPs right now—fuel volatility, labor shortages, capacity constraints—are not new. What is new is that clients have run out of patience for them. Where an LSP could once absorb disruption quietly and resolve it after the fact, clients now expect to be told what is happening in real time, before they have to ask.
Jamie argued that this shift has made fixed network models unworkable. The LSPs still running set carrier lanes and static routing are structurally unequipped to respond when something changes, and something always changes.
"You can't just have a fixed network anymore and do things the way we've always done them," Jamie said. "Everything is driving towards dynamic flexibility. The competition is fierce. We are all chasing very thin margins, and the clients have very little tolerance for anything that is inefficient or unstable because it impacts their bottom line downstream. You've got to have tech that enables you to be proactive and enables you to manage by exceptions, not manage every single step of the process."
Automation and dynamic routing have crossed a threshold where they are no longer capabilities that differentiate an LSP in a sales conversation. Today, they’re baseline requirements for staying in business. The LSPs treating them as competitive advantages are already behind.
What do LSP customers actually expect today, and where is the gap?
Amazon reset expectations across the entire supply chain. The upcoming generation of supply chain leaders at LSP clients grew up ordering something at noon and watching it arrive by dinner. That experience is now the reference point for every vendor interaction, regardless of freight type, geography, or complexity.
Raquel pushed on what this means practically for LSPs and asked Jamie to separate table stakes from differentiation. Jamie's said real-time visibility at the carrier and stop level is no longer a premium offering, it’s the baseline.
"Everybody compares to that Amazon-level of visibility and reliability. You have to deliver that experience across all of your tech and every interaction, regardless of which of your offices or stations they're interacting with, or which trucking partners are supporting that move," Jamie said. "The gap isn't in the execution. It's in the consistency of the data you're getting and the experience you're delivering to your customer."
The visibility problem compounds across a fragmented network. A single shipment can touch a first-mile carrier, a line haul, a customs broker, and a warehouse before it reaches the client. Each party holds a piece of the visibility picture and sends it in a different format. Reconciling that into a single, coherent update to the client is the LSP's job, and most are not set up to do it reliably.
"I've got to take people who are all at different levels of maturity, understand what data they can give me and when, and put it in a singular format that makes sense to my customer," Jamie said. “A delivery is a delivery no matter what event code these seven different people send me. I have to do that across lanes, across regions, across carriers.
“The data is the most important part of this—how I can get it and how quickly I can get it. My job is to deliver it through a single lens to our customer, because they don't care about all of that. That's my job to solve for them."
Why is integration such a persistent challenge for LSPs?
The real integration problem sits one layer below the technology. Carriers use different event codes. Clients want to connect in different formats. Every new market or service line adds more variation to an already fragmented data environment. Technology can process that complexity, but it cannot resolve it. Standardization has to come first.
Jamie pointed out that even when standard APIs exist, most clients don’t want to use them. They want the LSP to adapt to their format, on their timeline. That dynamic means every new client relationship can become a new integration build if the LSP has not built a flexible enough foundation to absorb it.
"The customer doesn't want to work with your APIs," Jamie said. “They want you to connect to their format. So we invested a long time ago in having an in-house integration team to build those branches to us.”
SEKO's response was an internal integration platform that centralizes all connectivity across domestic, international, cross-border, and warehouse operations. One inbound message hits every system regardless of service type or region. But building it required going back to first principles before writing a line of code.
"You have to start with defining every single piece," she said. “What is a shipment and what does that look like in all of the systems? What are my event codes that I want to have as standardized? Because every carrier is going to have different ones.”
Catch up on every episode of Deliver: The Last Mile Performance Podcast
What does scalable last-mile technology actually look like in practice?
Scalability in practice means reusability. A system is scalable when the business can configure it without returning to the development team every time something changes. That distinction, configuration versus custom code, is where most LSP technology investments break down.
When SEKO deployed a new last-mile technology platform across more than 30 US sites, spanning corporate-owned stations, strategic partners, asset-heavy and asset-light operations, the temptation was to start with a pilot and learn from it. Jamie pushed back on that approach.
"You start at the other side, get one pilot site live and learn from it, and then site two has a thing you have to account for. You re-architect. Then the fourth site has another thing. You're just forever developing," she said.
Instead, SEKO spent significant time upfront mapping the operational process across every site model before any deployment began. The goal was to identify what was consistent across all sites, what needed local configuration, and what the technology needed to produce at each step. That groundwork allowed them to deploy at a rate of several sites per week once the rollout started.
Within their integration portal today, operations teams and account managers can edit messages, resend transmissions, and create new event mappings without submitting a development ticket. When a geopolitical disruption creates a new category of delay, a team member can create the event and map it on the spot.
"I don't want this to be a consistently technical, heavy solution where every time we want to enable something for the business, it has to go back through our dev team and people have to wait weeks," Jamie said. “Scalability is architecting it right from the start so your business has some flexibility. Why would you build anything that doesn't have flexibility in the solution?
How is AI driving business performance for LSPs right now?
The version of AI being sold at industry trade shows and the version producing real operational results for LSPs are not the same thing. The gap between them comes down to sequence: did the organization start with a defined problem, or did a vendor define it for them?
Jamie was direct about how dangerous the second path is. Vendors can only solve the problems their tools were built for. LSPs that skip the internal diagnostic and go straight to a product demo end up with tools that do not map to their actual bottlenecks.
"Where most people tend to go wrong is they let a vendor tell them what problem they need to solve versus looking internally and understanding where there's an area of opportunity," she said. “That's a really dangerous road…People subscribe to these tools and say, ‘I'm not getting any value.’ That’s because you didn't start with identifying the problem.”
The tools drawing the most attention at events like Manifest, particularly quote automation applications promising 30-second response times, require clean, unified customer master data, margin data, and integrated systems. Most LSPs do not have that combination in place.
"The reality is most people aren't ready for anything like that at scale," she said. “Ninety percent of the people in our industry, their data and their customer master data and their margins, they don't have that all tied up in a little box with a bow on it.”
SEKO's near-term AI investments reflect a more grounded approach. One tool already in deployment lives inside Outlook. It scans incoming POD emails, identifies the job number, looks it up in the operating system, and surfaces a sidebar prompt asking the operator to confirm the delivery details before pushing them through. No new login. No new workflow. The operator stays in the environment where the work already happens.
"Think of the amount of time you're taking out from swivel-chairing and data entry errors," Jamie said. “Not anything incredibly wild and crazy, but that's where you start to build.”
The near-term AI wins for LSPs are delay prediction, automated exception flagging, and route optimization. They are also the investments that build the standardized data set needed for anything more ambitious later.
What does good LSP data look like, and what KPIs matter most?
Better data means better definitions. Before any LSP can govern, standardize, or act on its data, it has to answer questions most organizations have never formally resolved: What is a customer? How does that customer trace from the CRM to the operating system to the client portal? What event codes are standardized across every carrier in the network?
Jamie described a multi-year effort at SEKO to reframe data from a by-product of operations into a product that requires the same maintenance, governance, and ongoing investment as any core system.
"Data itself is a product. It is not an outcome of a shipment or a by-product of what we do," Jamie said. “You have to treat your data the same way as you treat your operating system. You have to maintain it, put upgrades in, and continue to make it better.”
On KPIs, Jamie drew a line between the metrics that are already table stakes and the ones that are starting to separate high performers from the rest. On-time delivery rate, first-attempt success rate, route efficiency, cost per mile, and exception resolution speed are the foundation. Every LSP should have clean visibility into those numbers. Most do not.
The next layer is where the competitive advantage starts to emerge. Jamie described a future state where unstructured data from emails, calls, and support interactions gets surfaced alongside operational metrics, giving account managers a signal that a client relationship is at risk before a service failure becomes a lost contract.
"How do I know that a customer sent an email with a negative tone because they had three service delays?" she said. “How do I get that data out of my email and make it a data point in my model so I can see when my customers are at risk?”
That shift, from reporting on what happened to predicting what is about to happen, is where data maturity starts to create a real performance gap between LSPs.
What is one piece of advice for LSP leaders looking to improve last-mile performance?
The last mile is the final impression an LSP leaves with its client. Every routing decision, carrier relationship, and technology investment made upstream becomes invisible if the last touchpoint fails.
Jamie closed with the sharpest framing of the episode's core argument: the last mile has stopped being a transportation problem and started being a data problem. LSPs that recognize that shift and act on it will be positioned to scale. Those that do not will keep underperforming without understanding why.
"If you go to a party and the food's bad, you don't remember that the music was great," Jamie said. “It's the last image you have of that experience. The last mile is not an operational or a transportation problem to solve anymore. You've got to start thinking about the last mile as a data problem.
"As people are getting pressured to start enabling their business with AI, go back to square one and lay that first brick on the data and standardization side before you think about doing anything at scale."
