The Next
Era of ERP.

Connected systems can still produce conflicting answers.

A system integration can successfully transfer records while the business remains unable to agree on what those records mean. That is why a connected architecture and a trusted information environment are not the same thing.

Gartner's 2026 ERP Hype Cycle highlights connected data as a foundation for more flexible enterprise applications. But integration by itself only transports information. A manufacturer's order promising accuracy can still fail when item identifiers differ between ERP and warehouse systems, lead times are maintained inconsistently or inventory status means different things to planning and sales.

Four problems that get called “data quality”

These problems require different remedies. Rebuilding an API does not resolve competing definitions of inventory availability. A master-data cleansing exercise will not repair delayed synchronization. A new dashboard may simply put inconsistent inputs into a more attractive format.

Build a decision-first data diagnostic

Instead of demanding an enterprise-wide cleanup, choose two or three decisions with material financial or service consequences: available-to-promise, margin by customer, inventory exposure or supplier reliability. For each, trace the lineage from source event to final report and quantify the workarounds introduced along the way.

Gartner's 2026 data-management research also notes the growing importance of semantic and context layers as AI ambitions increase. That reinforces the point: machines need consistent business meaning, not merely accessible data fields.

Project drift begins when decisions age without owners.

ERP implementation risk is often visible months before a missed go-live date. The leading indicator is frequently not software configuration progress. It is the number, age and consequence of unresolved business decisions.

Consider a distributor whose sales organization wants special fulfillment rules for strategic customers. Operations wants standardized picking and wave planning. Finance requires stronger controls before release. Each requirement is understandable. Without timely executive decisions, the implementation team may code exceptions, revisit process designs and repeat testing as priorities change.

Four controls worth demanding

Distinguish productive change from project drift

Not every changing requirement is evidence of failure. A new regulation, acquisition or discovered operational constraint may justify change. The difference is governance. Productive changes have a clear benefit, named decision-maker, measured effect on the plan and a deliberate acceptance or rejection. Drift accumulates where requests enter without that discipline.

Ask the project team to identify the five most consequential unresolved decisions and display their age alongside the critical-path milestones they affect. Then ask how many accepted changes have required process redesign or repeated testing. Those two views are often more actionable than a conventional color-coded status report.

Do not confuse go-live with business success

A technically successful cutover can still leave customer service impaired, manual work elevated or financial controls weakened. Value realization needs a baseline and a post-launch owner. Otherwise the organization celebrates the date, then discovers it has not achieved the reason for the investment.

Before another ERP or AI investment, test whether the data can carry it.

Enterprise Data Management (EDM) should be evaluated through the operating decisions it supports. A practical diagnostic finds where records, ownership and business rules are creating measurable friction—before the company spends money automating the wrong process.

From diagnosis to action: The Connected Data article identified where systems disagree. This Ultra Perspective shows how to prioritize those defects, assign ownership and verify that remediation changes a real business outcome.

For manufacturers and distributors, a few data domains recur in consequential decisions: item masters, bills of materials, supplier records, customers, units of measure, locations and inventory statuses. Problems here do not stay inside a database. They affect procurement, production planning, order promising, pricing and reporting.

The four-part EDM diagnostic

What the results might reveal

Illustrative example: A manufacturing company profiles its purchased-item records and finds missing preferred-supplier data. The IT team could populate the field through a one-time import. But a sustainable solution also needs a policy for when a supplier is considered preferred, a business owner who approves the designation and a workflow that keeps the value current. Otherwise the same defect returns.

The point is not to create a perfectly populated database. It is to reduce operational uncertainty. Before prioritizing a data defect, ask which decision it influences, how often it causes a workaround, who feels the consequence and what remediation will change.

Prioritize by business exposure, not record counts

A high-volume defect affecting an infrequently used legacy field may deserve less attention than a small group of inconsistent units of measure affecting production purchasing. A useful priority model weighs business impact, frequency, downstream reach, risk and cost to correct. Those factors should be discussed with operations and finance, not assigned by technology teams alone.

Ultra's EDM positioning appropriately connects data quality, governance, migration and ongoing support. The more compelling executive conversation is the sequence: diagnose first, make ownership explicit, address high-exposure breaks and verify benefits before scaling automation or migration.

Four questions that turn a technology discussion into a decision.

A leadership team can use these four prompts to test whether an ERP, data or AI initiative is grounded in business value. Each question should produce an owner and a next action, not simply agreement around the table.

A 15-minute executive exercise

Ask the CFO, COO and CIO to answer the four questions independently before the next steering meeting. Compare where their answers differ. The gaps often reveal the decisions that matter more than another vendor demonstration or technology roadmap slide.