Context

A privately held global market leader in construction technology. Products shipped to 100+ export-controlled countries across 14 regional market groups with 40+ product localizations. Hardware-first culture. FCC, CE, UKCA, and CCC certifications required per market. Approximately 250 employees. 95% of design, manufacturing, and repair at a single US headquarters.

The company had a fledgling digital platform, but it was not cloud-native and could not serve global regulatory requirements or scale as a growth engine. There was no AI capability, no product operating model, no UX function, and no business intelligence connecting product telemetry to business outcomes. The core product line had not seen a new flagship in over a decade. A two-year engineering change order backlog. $7M in accumulated technical debt. A market leader that had stopped investing in the future of its own product.

I joined as Head of Product on the executive team, reporting directly to the CEO with portfolio P&L authority over the full product line. The mandate was a turnaround: kill misaligned investments, redirect capital, ship new products, build a scalable digital revenue platform, and establish the product organization from scratch. The vision was clear: simplify the product experience for a workforce that was getting less experienced every year, and embed intelligence into the device so the product makes the expert decisions the operator no longer can.

Challenge

The company’s field operators work underground. They use frequency-based locating equipment to track drilling operations in real time. The environments are GPS-denied, often with intermittent or zero cellular connectivity. Extreme conditions: vibration, temperature swings, dust, water exposure.

The existing workflow required operators to interpret raw frequency data manually. An experienced operator reads signal strength, depth estimates, and interference patterns from a display and makes real-time decisions about drill path and utility avoidance. This is a skill that takes years to develop. A misread can mean a utility strike, project shutdown, or injury.

Skilled labor shortages were compressing the talent pool. Fewer experienced operators entering the market. More complex jobs. The same analog interpretation process that had been in place for decades.

A cloud-based AI solution was not viable. The underground construction environments where the product creates the most value are the environments with the least reliable connectivity. Sending sensor data to the cloud for processing and waiting for a response introduces latency that is unacceptable for real-time field decisions. The AI had to live on the device itself.

Approach

I drove the product strategy for an edge-first AI platform. The core architecture decision: an on-board neural network running on the device’s embedded processor, within the power and thermal constraints of a handheld battery-powered field instrument. No cloud dependency for core operations.

The data pipeline. The neural network processes 8,000+ sensor data points per workflow execution. These are frequency scanning inputs: signal strength across multiple frequencies, signal quality indicators, interference patterns, and environmental noise readings. The neural network evaluates this data in real time and generates autonomous frequency scanning decisions that previously required a human expert. The device updates its intelligence model periodically from the cloud when connectivity is available. Between updates, it acts independently.

The architecture decision. I evaluated three approaches with my engineering and product teams: cloud-only (rejected for latency and connectivity reasons), hybrid with cloud fallback (rejected because fallback to manual mode in the hardest environments defeats the purpose), and edge-only with periodic cloud sync (selected). The edge-only approach meant the device needed sufficient processing power to run inference on-board. This drove hardware design decisions about processor selection, memory allocation, and battery life that rippled through the entire product architecture. The AI was a new product architecture built from the ground up as part of the first all-new platform in nearly two decades.

The field trials. I championed and led the company’s first-ever customer field trials. The company had shipped products for over 30 years and had never tested a product with customers in real operating conditions before production commitment. I had to argue for this approach against a culture that trusted engineering simulation and internal test benches. The argument was simple: the only way to validate an AI product for underground environments is to put it underground. Simulation cannot replicate the soil variability, interference conditions, and operator behavior that determine whether the AI works.

The field trials were not just product validation. They were the fastest path to market signal. We tested with real operators in real ground conditions across multiple job sites. Within weeks, we had direct feedback on what worked, what confused operators, and where the AI needed calibration. The trial data validated the AI’s frequency scanning accuracy and provided the baseline for the 40% support call deflection projection. The support deflection metric represents the proportion of frequency interpretation decisions that the AI can make autonomously, eliminating the call an inexperienced operator would otherwise make to technical support for guidance. That market signal converted directly to production confidence.

Building the team. The AI platform was one output of the 20+ person direct product organization I built from scratch, with over 40 additional team members matrixed through GTM. I founded UX Design (promoting the lead from within), Business Intelligence, and Product Operations as new functions and completely reset the existing Product Management function.

The UX team’s work on the setup experience was a prerequisite for the AI platform and the clearest proof of the simplicity thesis. The legacy setup workflow took 60 minutes per unit. Once the UX strategy and vision were in place, our engineers ran with it. The creativity that came out of the engineering team when they had a clear UX framework to build against was remarkable. They redesigned the workflow to 15 minutes. A 68% reduction. That outcome was not a UX project or an engineering project. It was what happens when a product organization gives talented engineers permission and direction to solve for simplicity. If the operator struggles with basic setup, they will never reach the AI-powered workflow. The AI is only as valuable as the product’s ability to get the operator to the point where the AI can help them. Simplicity at the UX layer is what makes intelligence at the AI layer accessible.

Outcome

40% projected support call deflection. The AI automates the frequency interpretation that previously required a human expert. For an industry facing structural skilled labor shortages, this is an access improvement. Less experienced operators can now perform work that previously required years of field training.

8,000+ automated decisions per workflow. Each workflow execution processes thousands of sensor data points through the neural network without human intervention or cloud dependency.

$15.9M in identified lifecycle cost savings. I commissioned a production data cost model to quantify the total financial impact of the portfolio decisions: the $40M program kill, the VR product sunset, the partnership termination, and the new product investments. The model validated the investment case for the AI platform and informed the capital allocation sequence. At a bootstrapped company with no outside funding, every dollar redirected from a killed program has to earn a higher return on the replacement. The $15.9M was the proof that the replacement portfolio was worth the organizational disruption.

68% time-to-value reduction. The UX-driven redesign of the setup experience (60 minutes to 15 minutes) was measured by timing the complete setup workflow before and after the redesign across multiple device types and operator experience levels.

The platform launched at the industry’s largest trade show one month after my departure. Unveiled by the company’s second-generation leadership. The product operating model, the team, and the roadmap governance I established are still in place. The new flagship product. Neural network on board. 8x more frequencies than any competitor. Wireless calibration. Global OTA firmware updates. The first all-new platform in nearly two decades.

Technical standards and compliance context

The company’s products require market-specific certifications: FCC Part 15 (United States), CE marking (European Union), UKCA (United Kingdom), and CCC (China), among others. Each certification affects hardware design, firmware configuration, RF emissions characteristics, and packaging. The AI platform had to meet these standards across all 14 market groups, which constrained processor selection, antenna design, and power management decisions. The digital cloud platform I defined for data sync and model updates runs on Azure with GDPR and CCPA compliance for international data handling.

Lesson

Edge-first is the only viable architecture for industrial AI in field environments. The temptation is to build a cloud AI product and assume connectivity. In construction, manufacturing, and logistics, connectivity is the variable, not the constant. The intelligence must live where the action happens. Cloud handles model updates, fleet analytics, and pattern recognition at scale. The edge handles decisions.

The second lesson: simplicity enables intelligence. The 68% time-to-value reduction was not a UX project. It was a prerequisite for the AI platform. If the operator cannot get through setup, the neural network never fires. Every layer of complexity between the user and the intelligence layer is a layer where adoption dies. The product leader’s job is to remove those layers before investing in the intelligence.

The third lesson: field trials are market signal, not just product validation. The data from real operators in real ground conditions told us more in weeks than internal simulation told us in months. For AI products in physical environments, the field is the only test bench that matters. I had to fight for the company’s first field trials. The data they produced justified the entire investment and gave the organization confidence to commit to production.

The fourth lesson: the operating model matters more than the product. The AI platform will eventually be superseded by better AI. The product organization and operating model I established to produce it will ship the next product too. Organizations outlast products. Build accordingly.

Technologies and standards referenced

  • On-device neural network inference (edge AI)
  • FCC Part 15 (US wireless certification)
  • CE Marking (EU product conformity)
  • UKCA (UK product conformity)
  • CCC (China Compulsory Certification)
  • GDPR and CCPA (international data privacy)
  • Microsoft Azure (cloud platform for OTA updates and analytics)
  • RF emissions and antenna design constraints
  • OTA firmware update architecture

About the author

Product executive. 15+ years building industrial AI platforms, B2B SaaS products, and connected smart device ecosystems in regulated industries across 100+ countries. Three portfolio turnarounds. Three org builds. Three times the methodology transferred, only the industries changed.

Nick builds at the hardware-software-data intersection. Industrial AI. Edge-to-cloud platforms. Workflow automation systems making 8,000+ decisions per workflow with zero cloud dependency. The career pattern: enter complex regulated environments, find the kill decisions others avoid, and redirect capital from legacy programs to products that ship and outlast him. The acquiring company kept his product. Threw away their own.

Most recently Head of Product at Digital Control Incorporated. Global product portfolio. Turnaround-to-growth. Previously at Zonar Systems, a subsidiary of $44B annual revenue Continental AG, leading a $70M connected device platform across three continents, and at Rehrig Pacific Company building an innovation function from scratch.

Leading global products and global teams as a Chief Product Officer, Head of Product, VP of Product for B2B and B2B2C companies for digital transformation and product growth leadership.

More about Nick →