With more than three decades of high-growth experience managing complex, multi-phase billion-dollar engineering projects with rigorous compliance and documentation retention requirements, ATCS was experiencing growing pains around its information management platforms. This was becoming increasingly concerning as the company began to take advantage of more advanced BI and AI tools.

Like many firms of their size and tenure, ATCS’s operational data has grown and scaled, but across disconnected systems. Financial transactions lived in Deltek. Payroll ran through ADP. Images and videos captured at every construction phase sat in separate storage—technically preserved, minimally accessible. The result was a ritual of manual data extraction and spreadsheet reconciliation.

ATCS partnered with dbSeer to address this at the foundation: a Databricks lakehouse deployed on AWS cloud, designed to connect these systems, eliminate data inconsistencies grown from entering data in multiple platforms, and position the organization to begin leveraging advanced analytics and AI on governed, trustworthy data.

The Challenge: When Your Own Teams Can’t See Their Own Data

The problems ATCS faced is not unusual for a firm that had grown steadily for three decades—but they were becoming increasingly costly.

Department leaders and project managers, particularly newer hires, could not independently pull the data relevant to their work. Information was centralized, filtered, or available only through manual reporting cycles. The operational consequence was a team making decisions on delayed signals rather than current reality.

Perhaps the most insidious problem was what ATCS described as “dueling data”: the same project information calculated or stored differently across systems. As they put it:

“Dueling data is not just bad because we’ve got multiple entries for the same information. It affects our ability to strategize and streamline ownership of data.”

Tamir Sherif, Executive Vice President ATCS

Without a unified foundation, ATCS can’t identify portfolio-wide patterns, predict which projects were at risk, or prepare for modern capabilities like AI. The data foundation had to come first.

The Solution: A Unified Lakehouse Architecture, Built for ATCS’s Reality

dbSeer designed and implemented a Databricks lakehouse deployed within ATCS’s AWS environment. The Databricks workspace operates as a secure extension of their existing infrastructure.

Medallion Architecture: Three Layers, One Source of Truth

The architecture follows Databricks’ medallion pattern, progressively refining data through three layers.

The bronze layer ingests raw data from source systems exactly as received—financial transactions, timesheets and project information from Deltek, payroll data from ADP, and critically, the images and videos from construction sites. For the first time, this multimedia archive is ingested into the platform and made searchable in connection with project records.

The silver layer is where transformation happens. Data cleansing logic addresses quality issues that accumulate in ERP systems running for decades. Enrichment processes merge Deltek project hours with ADP payroll costs, creating unified views that weren’t possible when these systems operated in silos. Critically, this layer establishes the consistent business logic that ensures everyone—across every project and every team—calculates metrics the same way. No more dueling data.

The gold layer delivers role-specific views designed for consumption. Project managers see portfolio status with daily updates. Executives see organization-wide backlog, pipeline, and financial performance. Finance teams see revenue recognition and vendor payment status. These gold-layer views feed directly into Power BI dashboards that replace the manual Excel consolidation ritual.

Unity Catalog: Governance That Enables AI

Unity Catalog is live within ATCS’s Databricks environment and serving as the governance layer for the platform — managing access controls that ensure project managers see their projects while executives see across the organization.

The next step is populating the catalog with business definitions. When each table and column is documented with the meaning behind the data, the platform becomes fully self-describing, and that’s where the AI story begins. Unity Catalog is specifically designed to support this: once business definitions are in place, the metadata layer gives machine learning and generative AI models the context they need to deliver accurate, trustworthy results. The models will know not just what the data is, but what it means.

This is what dbSeer means by “smart foundations” — building the architecture today so that when the organization is ready to act on it, the capability is already there. ATCS isn’t waiting for AI readiness to become urgent before building toward it. The foundation is now in place.

Proof-of-Concept First

As a Databricks Consulting Partner and AWS Advanced Partner, dbSeer brought expertise in both the technical platform and the operational realities of engineering firms managing complex, resource-intensive projects. dbSeer’s proof-of-concept–first methodology aligned closely with ATCS’s need to validate solutions before full deployment, ensuring the architecture addressed real business requirements rather than theoretical ideals.

“At the outset we were surprised at how familiar dbSeer was with our business challenges, their ability to share many best-practices and the very business centric approach they took to developing solutions.”

Tamir Sherif, Executive Vice President ATCS

The Outcome: From Snapshots to Real-Time Operational Intelligence

The impact of the Databricks lakehouse was felt immediately across ATCS’s operations.

Operations teams now have direct access to their own information. Where department leaders previously depended on centralized reporting cycles to understand their own domains, they now have independent, direct access to the data relevant to their every day—in a format that is now more intuitive to workflows. This shift alone represents a fundamental change in how the organization operates day-to-day.

Data inconsistencies are surfacing and being resolved.The unified platform has already begun identifying disconnects between ERP systems that previously went undetected. ATCS can now proactively address inconsistencies rather than discovering them downstream: protecting both project integrity and financial reporting accuracy.

Visualization capabilities have transformed entirely. Moving to Power BI dashboards connected to a live lakehouse has, in their words, allowed them to “play at a dramatically different level.”

“We’re better able to identify and resolve disconnects between our ERP systems. It’s going to help us better streamline our workflow, so we don’t have dueling data.”

Tamir Sherif, Executive Vice President ATCS

What’s Next: Moving Up the Data + AI Maturity Curve

Databricks Data & AI Maturity Curve

With the foundational platform in place, ATCS and dbSeer are already thinking about what becomes possible next. Their near-term focus is predictive analytics. This kind of predictive capability is now within reach precisely because the data foundation is in place.

Longer-term possibilities include Salesforce integration to connect sales pipeline data with project execution, generative AI search across the image and document archive, and broader predictive capabilities to identify at-risk projects before problems materialize.

Perhaps most notable is how ATCS is approaching AI readiness from the inside out. Leadership is actively building data literacy across their teams—through structured internal learning, cross-functional collaboration, and deliberate skill development—with the goal of ensuring that as AI capabilities are introduced, the organization is prepared to use them effectively. The Lakehouse foundation makes this possible; the culture work will make it sustainable.

The Databricks Lakehouse and Unity Catalog foundation ensures these future capabilities can be delivered incrementally, each building on governed, well-documented data. That’s the promise of a smart foundation: not just solving today’s problems but making tomorrow’s ambitions achievable.

Ready to Build Your Smart Foundation?

If your team is still consolidating spreadsheets instead of making decisions, it’s time to address your data foundations. dbSeer’s foundation-first approach identifies where your data lives today and designs a practical path to unified, AI-ready analytics—without ripping out the systems you already have.