ProArch Logo

Lewis Services Achieves 90% Efficiency Gains

With ProArch’s Dataware Data Platform

Lewis Services

About Lewis

Lewis Services is one of North America’s largest vegetation management companies. With over 4,000 staff members and 200 customers across the country, Lewis plays a critical role in the US energy infrastructure.

 

Solutions Used

 

 

Situation: Painful Invoice Processes & Time-Consuming Timesheets

Without a centralized system, Lewis’s payroll processing and customer invoicing were challenging, slow, and manual. To put it in perspective, the company processed 86,000 spreadsheets for payroll each year. This required nearly 45 full-time employees to complete, and the General Foreman had to enter the same data four times.

“The general foremen were being pulled in so many directions that it was causing inefficiencies in operations,” says Huntley Hedrick, VP of IT at Lewis.

Weekends of time-consuming work made for unhappy employees, payroll delays, slow revenue realization, and poor customer experiences.

 

Solution: Data Platform & Custom Application

Using ProArch’s Dataware Data Platform, the team integrated the HR and ERP systems into a data warehouse and then developed a custom application called “Tiempo.”

The data modeling in the background simultaneously calculates payroll and invoices based on customized rules by location, job type, pay rate, employee, and start/end time. 

 

Results: A Game-Changer All Around

  • Lewis has experienced a 90% reduction in the overall effort with a clear, streamlined process for time entry, timesheet creation, and approval.
  • Advanced reporting capabilities using data from different systems.
  • Before, Lewis experienced a high volume of invoice adjustments. Now, it is a 3% adjustment rate.
  • Customers have a better experience, and Lewis can realize revenue faster.
  • Lewis owns the tool so they can avoid future costly and complex ERP upgrades and customization.

For the general foremen, Tiempo has been transformational. Hedrick added, “What I’ve observed with the rollout of Tiempo has been greater than anything that I’ve ever seen in my career with how accepted it’s been in such a short amount of time.”

AI Predictive Maintenance Case Study: Power Plant Equipment Health

About

A large‑scale power generation facility set out to move from reactive maintenance to AI‑driven equipment health monitoring. This case study documents how real‑time sensor data and AI models were used to identify performance anomalies early—laying the foundation for reduced unplanned downtime across critical assets.

Services

Challenge
  • Operational data exported manually from AVEVA PI for offline analysis
  • Delays between data collection and actionable insight
  • No real-time visibility into asset health
  • Limited scalability across assets
  • Early detection of equipment issues was nearly impossible

The organization aimed to move from reactive, spreadsheet-based analysis to proactive, AI-driven predictive insights to improve operational efficiency, optimize energy output, and reduce unplanned downtime.

Solution

The engagement began with an ImpactNOW AI Proof of Value to move the plant from offline analysis to real‑time equipment intelligence. Using Microsoft Fabric Real‑Time Intelligence, near real‑time OT data from AVEVA PI was ingested alongside manufacturer engineering specifications and historical operating data.

AI models continuously compare balance‑of‑plant equipment performance against original design curves to detect multi‑signal drift, surface anomalies earlier, and validate findings against historical maintenance events.

Dashboards provide clear visibility into equipment health and performance risk—establishing a scalable foundation for expanding AI‑driven monitoring across additional assets and facilities.

Focused equipment scope

  • Monitored 3 critical pumps and 4 associated sensors, selected based on operational importance and failure risk
  • Established a repeatable pattern for starting small and expanding across additional assets

Real‑time data and AI analytics

  • Replaced manual data exports with direct ingestion from AVEVA PI into Microsoft Fabric
  • Implemented real‑time detection of pressure, vibration, and flow deviations
  • Evaluated multi‑signal behavior to identify early indicators of performance drift

Performance vs. design analysis

  • Compared live pump performance against manufacturer engineering design curves
  • Calibrated performance curves using historical operating data specific to plant conditions
  • Prioritized anomalies based on deviation magnitude and operational risk

Model validation and trust

  • Validated AI outputs by cross‑referencing detected anomalies with historical maintenance logs
  • Demonstrated alignment between detected drift patterns and known past events
Result

Observed Operational Impact

The engagement delivered measurable operational value by shifting the plant from offline analysis to near real‑time equipment performance monitoring using AI.

  • AI‑driven drift and anomaly detection validated against historical maintenance events
  • Near real‑time visibility into pump performance versus manufacturer design curves
  • Significantly reduced reliance on manual data exports and delayed analysis
  • Improved confidence in identifying performance degradation before it resulted in efficiency loss or unplanned events

Positioning the Plant for Predictive Maintenance at Scale

Building on the initial ImpactNOW engagement, the solution has progressed into a durable, AI‑enabled equipment intelligence capability—supporting both immediate operational decisions and longer‑term predictive maintenance objectives.

Working with ProArch, the organization is now positioned to achieve:

  • Up to 50% reduction in unplanned downtime through early drift detection and predictive alerts
  • Up to 25% reduction in maintenance costs by transitioning from schedule‑based to condition‑based maintenance
  • A clear path to expanding AI-driven insights across assets and systems
What This Case Study Demonstrates
  • How AI can detect early equipment performance drift in power generation
  • Why performance vs. design curves matter more than static thresholds
  • How real-time monitoring reduces hidden efficiency loss and outage risk
  • How plants can start with one asset and scale equipment intelligence across the fleet

 Read the full story 

Industry

  • Power Generation

Challenge

  • Manual AVEVA PI data analysis limited visibility into equipment health, allowing performance drift to quietly increase efficiency loss and outage risk.

Solution

  • Conducted an AI proof of value to establish a centralized equipment intelligence foundation using near real‑time OT data, engineering design specifications, and AI‑driven drift detection.
  • Equipment performance is continuously compared against manufacturer design curves to identify drift early, prioritize anomalies, and surface root‑cause insights.

Result

  • Near real‑time visibility into equipment performance versus design intent, enabling earlier intervention before efficiency loss or forced outages occurred. 

Fast Track Unlocking Value from Your Data

Connect With Us