LOGISTICS & SUPPLY CHAIN
Supply chain tech attracted $15.4 billion in VC in 2024. The hardest part isn't building the platform — it's integrating it into the customer's ERP.
Logistics software companies promise visibility, optimisation, and automation. What they often deliver is a complex integration project that requires the customer's IT team, a legacy ERP that wasn't designed for real-time data, and months of configuration before the platform delivers value. We find the gap between the demo and the deployment.
Why it’s different
Logistics software lives or dies at the integration layer — which is rarely in the demo.
Logistics and supply chain technology is one of the most integration-intensive categories in enterprise software. Every customer brings a different ERP, WMS, TMS, or carrier system. The number of legacy integrations required to deliver value scales with the customer count. And unlike CRM or HR software, logistics platforms typically sit inside time-critical operational processes where downtime is immediately visible and expensive. The pitch shows a clean dashboard. The DD needs to show what's behind it.
01
The integration layer is where the business model breaks
Gartner data shows that 62% of supply chain AI initiatives exceed their budgets by an average of 45%, largely due to unforeseen data preparation requirements and integration complexities. Logistics platforms that require custom engineering for each customer integration are not SaaS businesses; they are professional services businesses with a SaaS aspiration. We assess what fraction of a new customer deployment is automated, what fraction requires engineer time, and what that implies for the real cost to serve.
02
Legacy system dependency is a revenue risk that appears nowhere in the pipeline model
Supply chain operations run on ERP infrastructure built in the 1990s and 2000s — SAP, Oracle, JDE, Microsoft Dynamics — that was never designed for the real-time API calls that modern logistics optimisation requires. Batch file exports, EDI connectors, and SFTP-based integrations are still standard in this industry. A logistics software company that has built its product around real-time data flows from customer ERPs is implicitly betting that customers will upgrade their infrastructure. We assess the actual data pipeline architecture and its dependency on customer infrastructure quality.
03
AI optimisation claims require actual operations data — and most companies don't have enough of it yet
Route optimisation, demand forecasting, inventory positioning, and carrier selection algorithms all improve in accuracy with volume and variety of operational data. A platform with three years of data from 15 customers makes different predictions than one with six months of data from two. AI performance benchmarks in logistics need to be evaluated against real production data, not against toy datasets or single-customer pilots.
Assessment Areas
Where we focus in Logistics & Supply Chain engagements.
AI in Logistics & Supply Chain
AI in logistics is creating real efficiency gains — and a new category of implementation risk.
Supply chain AI initiatives that succeeded in 2025 shared one characteristic: domain expertise-led implementation. Companies that invested at least 15% of their AI project budgets in change management and domain knowledge transfer reported 2.8x higher adoption rates. The technology alone is not the constraint. The operational integration is.
Opportunities we verify
Demand forecasting models trained on multi-customer data. Logistics platforms with access to demand and fulfilment data across multiple customers and sectors can build forecasting models that outperform single-enterprise internal tools. The accuracy improvements compound with network size — a genuine network effect in a category where network effects are rare.
Autonomous carrier selection and rate management. AI-powered carrier selection — optimising across cost, speed, reliability, and carbon footprint in real time — creates measurable savings for shippers and stickiness for platforms. Companies with multi-carrier API access and the AI layer to optimise across them have a procurement efficiency story that is clear and quantifiable.
Predictive disruption intelligence. Supply chain platforms that can aggregate external signals — weather, port congestion, geopolitical risk, supplier financial distress — and translate them into operational alerts have a risk management value proposition that sits above tactical execution. This is where AI creates genuine strategic differentiation.
Risks we surface
62% of supply chain AI projects exceed budget due to integration complexity. This reflects the structural difficulty of integrating AI into heterogeneous legacy logistics environments. The integration architecture is either built correctly from day one, or it becomes a professional services burden that consumes the team. We assess the integration architecture specifically for this pattern.
Workforce resistance as the primary implementation failure mode. A 2024 Deloitte survey found that 72% of logistics AI implementations that failed cited workforce resistance, not technical issues, as the primary cause. For a VC-backed logistics software company, this translates directly into implementation risk, churn risk, and NPS risk.
Real-time data promises built on batch infrastructure. Real-time supply chain visibility is a compelling pitch. Batch EDI updates every four hours is the underlying reality at many logistics software companies. The gap creates customer expectation mismatches that surface in renewals. We review the actual data latency delivered in production environments.
Know what you’re backing before you commit.
X-Ray delivers a full product and tech verdict on any logistics or supply chain technology target in one business day — covering the integration layer, the AI data quality, the ERP dependency, and the operational resilience.
250+ European engagements · 100% partner repeat rate