The study found that 52% of logistics operators still re-enter the same shipment data across multiple systems, while 49% switch between five or more platforms to complete a single workflow. Deep Current describes this as a sign that humans remain the “integration layer” in freight operations, manually connecting systems that still do not work together properly.
The findings come at a time when logistics companies are under growing pressure to automate routine work, reduce delays and improve resilience. However, the report suggests that many AI projects are still failing to reach the execution layer of daily operations.
According to Deep Current, the problem is no longer simply visibility. Many logistics companies can now detect delays, shipment exceptions and disruptions in real time. The larger problem is what happens next: operational teams still have to interpret, validate and transfer information manually between fragmented systems.
Emails and spreadsheets still run daily operations
The report says 61% of logistics teams still depend on emails and spreadsheets for operational communication, while 57% report shipment delays caused by document errors. Only 29% of companies surveyed said they had implemented digital tools across core operational workflows, and 47% cited legacy system integration as the biggest barrier to adoption.
Deep Current argues that this is where many digitalisation efforts break down. A company may have a transport management system, warehouse management system, ERP software and visibility platform, but the data needed to run the shipment often remains scattered between documents, messages and manual updates.
Tamim Fannoush, founder and CEO of Deep Current AS, said many logistics organisations still operate with fragmented workflows, requiring a “human integration layer” across more than five systems on average per workflow.
“Even in 2026, many tech platforms and AI models still depend on this human intervention to deliver results,” Fannoush said in the company’s press release.
Document errors remain a direct cause of delays
The report places particular emphasis on document-heavy workflows, where small errors can have direct operational consequences. A missing HS code, an incorrect consignee detail or a mismatch between an invoice and shipping document can delay a shipment if the problem is not caught early. Deep Current says 57% of logistics companies surveyed reported shipment delays caused by document errors.
The company’s report argues that many AI tools are still treated as add-ons rather than being embedded into the workflow itself. In practice, this can mean that staff have to open another dashboard, copy data into another system, check documents manually and then return to the original workflow.
According to Deep Current, this creates extra steps rather than removing them. The company says the highest friction is found in data connectivity and workflow integration, where systems remain disconnected and AI operates outside execution.
“As long as AI sits outside operational execution, teams still end up doing the integration work manually,” Fannoush said. “Copy-paste workflows, repeated validation and fragmented communication continue to absorb enormous operational capacity across freight.”
From visibility to execution
The report identifies five areas that it says determine whether logistics companies can move from digital experimentation to operational AI:
- integrated digital foundations,
- decision intelligence,
- workflow embedding,
- predictive resilience and governance.
Deep Current argues that the industry has spent much of the past decade improving visibility through dashboards, alerts, control towers and tracking systems. However, the company says visibility alone does not solve operational problems if teams still have to decide manually what to do next and then carry out the update across several systems.
The report says the next phase of logistics digitalisation will depend less on whether companies have AI tools and more on whether those tools sit inside the actual workflow.
The report was published by Deep Current AS, a logistics AI provider, and its findings should therefore be read in the context of the company’s work in document automation and workflow intelligence. However, the figures point to a wider problem across freight operations: many companies may be investing in AI, but their basic workflows are still built around email chains, spreadsheets and manual data entry.
Read more: What’s possible in logistics with AI agents
