AI in Your ERP Transformation: Where It Actually Improves the Project

Sep 16

Written By W Jones

Artificial intelligence is quickly becoming part of the ERP transformation conversation. Leaders across government, higher education, manufacturing, aerospace and defense, and grant-funded organizations are asking the same question:

Where can AI create meaningful improvement without adding another layer of risk?

The answer is not everywhere: and it is not by simply adding an AI tool to an existing project.

AI does not replace sound planning, reliable data, clear business processes, or accountable leadership. It cannot turn an unstable ERP program into a successful one by itself. However, when applied to the right problems, AI can help organizations see issues earlier, reduce repetitive work, improve decision-making, and increase the likelihood that transformation benefits continue after go-live.

The most practical approach is to treat AI as an accelerator and an additional set of analytical capabilities: not as a substitute for ERP fundamentals.

Start with the reality: people, process, and data still determine success

ERP transformation projects are often described as technology initiatives. In practice, the greatest risks usually involve people, processes, and data.

A new ERP environment may expose unclear ownership, inconsistent policies, duplicate records, disconnected workflows, and undocumented business rules. Those issues existed before the transformation, but the project makes them more visible because they must be addressed at scale.

AI will not automatically fix a broken approval process or determine the correct interpretation of an organizational policy. It will not make incomplete grant records reliable or resolve competing definitions of the same financial term.

What it can do is help teams find these problems faster.

Think of AI as a powerful searchlight in a large warehouse. The searchlight can reveal misplaced inventory, empty shelves, and blocked aisles far more quickly than a person walking through every row. But people still decide what belongs where, which items should be removed, and how the warehouse should operate going forward.

That distinction matters. AI improves ERP transformation when it helps leaders and subject-matter experts make better decisions: not when it is treated as the decision-maker.

1. Faster discovery of process problems

Most organizations understand their processes through a combination of interviews, workshops, reports, institutional knowledge, and informal workarounds. Those methods remain important, but they do not always reveal how work actually moves through the organization.

AI can analyze large volumes of transaction and process data to identify patterns that are difficult to see manually. It may help surface:

Repeated approval delays
High-volume exception paths
Transactions that require unnecessary rework
Differences in how departments perform the same activity
Unusual activity that deserves human review
Bottlenecks between systems, teams, or approval points

For a public-sector organization, this could help clarify where procurement, budget, or grant-related activity tends to slow down. In higher education, it may reveal variations across departments in purchasing, research administration, or student-related processes. In manufacturing and aerospace and defense, it may highlight handoff delays, production exceptions, or data gaps that affect planning.

The benefit is not simply a faster analysis. Better discovery helps leaders focus business process transformation on the areas most likely to improve service, control, cycle time, and user experience.

2. Better data readiness before migration

“Follow the data” is one of the most useful principles in an ERP transformation. Data reveals how the organization actually operates, where definitions differ, and which processes create downstream consequences.

Legacy data is rarely clean. It may contain duplicates, incomplete fields, inconsistent codes, outdated records, or conflicting values across systems. Moving that data into a new environment without understanding its condition can transfer old problems into new workflows.

AI can assist by reviewing large datasets and flagging:

Potential duplicate people, suppliers, customers, or records
Missing or incomplete information
Inconsistent naming and coding conventions
Values that fall outside expected patterns
Relationships that do not align with business rules
Records that may require specialized review

This does not mean AI should make final decisions about what to keep, change, or remove. Data owners and business leaders must remain accountable for those decisions. AI can, however, reduce the time required to identify where human attention is most needed.

For grant-funded organizations, higher education institutions, and government agencies, stronger data readiness can support more reliable reporting, better operational visibility, and greater confidence in financial and program information.

3. More effective testing and validation

ERP testing is one of the most important: and most demanding: parts of a transformation. Testing must reflect real business activity, including routine transactions, unusual scenarios, approvals, integrations, reporting, and high-volume periods.

AI can support this work by helping teams:

Generate candidate test scenarios from documented processes
Identify gaps in test coverage
Prioritize scenarios based on business importance and risk
Compare expected and actual results
Detect unusual patterns in test logs
Highlight areas requiring additional validation

The value is not in generating a large number of test cases for its own sake. More testing is useful only when it reflects the organization’s actual priorities and operating conditions.

Human reviewers still need to confirm that scenarios are relevant, results are interpreted correctly, and critical processes receive appropriate attention. This is especially important when testing financial transactions, grant activity, procurement, payroll, production, or other processes where an incorrect result can affect operations and public trust.

AI can make testing more targeted and responsive. It cannot replace business ownership of the results.

4. Preserving knowledge when people move on

ERP transformation projects often depend on a small number of experienced employees who understand how policies, processes, systems, and exceptions fit together. When those employees retire, change roles, or leave the organization, important knowledge can disappear with them.

AI can help capture and organize business rules, process decisions, frequently asked questions, and explanations of why certain choices were made. It can also make approved information easier for employees to find and use.

This can strengthen knowledge continuity during the transformation and after go-live. It can support:

Faster onboarding for new team members
More consistent answers to routine questions
Better continuity between project phases
Clearer explanations of process changes
More useful ERP training and user support

The quality of the result depends on the quality of the source material. AI-generated summaries should be reviewed, approved, and maintained. Unverified content can create confusion just as quickly as missing documentation.

5. Improving day-to-day operations after go-live

The transformation does not end when the new ERP environment becomes operational. Teams still need to learn new processes, complete routine tasks, find information, and resolve questions.

AI can assist with everyday work by helping users:

Find approved process guidance
Summarize routine information
Navigate common questions
Identify missing information in a request
Prepare standard reports or explanations
Receive more timely ERP training support

Used appropriately, this assistance can allow employees to spend less time searching for answers and more time on analysis, service, planning, and problem-solving.

The goal is not to remove human judgment from the organization. The goal is to reduce avoidable friction so people can focus on work that requires experience, context, and accountability.

6. Giving leaders earlier warning signs

ERP programs generate large amounts of information about scope, schedule, risks, decisions, testing, changes, and readiness. Leaders often receive this information through periodic reporting, which can make it difficult to see emerging patterns.

AI can help analyze program information and identify signals that deserve attention, such as:

A growing concentration of unresolved issues
Repeated delays in a critical workstream
Increasing change activity in a sensitive area
Test failures clustered around a particular process
Readiness concerns that are not improving over time
Risks that may affect cost, timing, or adoption

This type of analysis can help leadership address issues while they are still manageable. It is not a replacement for governance or executive judgment. It is a way to improve visibility and focus conversations on the conditions most likely to affect the project.

Where AI does not help: and where caution is essential

AI can produce confident answers that are incomplete, inaccurate, or based on misunderstood context. It can also repeat errors found in the data it analyzes.

That creates several important boundaries:

Garbage in, garbage out: Poor source data can produce poor recommendations.
Human review is required: AI-generated outputs should be checked before they influence decisions.
Sensitive information requires care: Data privacy and security obligations must be considered before information is shared with any AI-enabled capability.
Accountability cannot be delegated: Leaders and process owners remain responsible for decisions involving financial, grant, operational, or compliance information.
Results must be traceable: Organizations need to understand what information informed a decision and who approved the outcome.

For government organizations and higher education institutions, public accountability makes these boundaries especially important. Government IT compliance is not strengthened simply because an AI system was used. Stronger outcomes depend on clear ownership, appropriate controls, documented decisions, and disciplined review.

AI adoption needs governance, too

The same discipline that supports a healthy ERP transformation should guide AI adoption.

Organizations should establish clear answers to practical questions:

Who owns each AI use case?
What information may be used?
When is human approval required?
How will outputs be reviewed?
What decisions must remain with designated leaders?
How will changes and exceptions be documented?
How will the organization know whether AI is creating value or risk?

These questions are not meant to slow innovation. They help ensure that innovation supports the organization’s mission instead of creating a new source of uncertainty.

How CD&A helps leaders determine where AI fits

CD&A Consulting Services Inc. provides an independent perspective for organizations evaluating ERP transformation, IT transformation, business process transformation, and related readiness concerns.

As the Independent Set of Eyes, CD&A helps leaders understand whether their data, processes, governance, and organizational readiness support practical AI use: or whether AI would add risk before the foundation is ready.

Our approach is vendor-neutral. We do not sell software or promote a specific platform. The focus is on helping government, higher education, grant-funded organizations, manufacturing companies, and aerospace and defense leaders make informed decisions that improve transformation outcomes.

If your organization is planning an ERP transformation or already managing one, now is the right time to ask:

Where can AI improve the project, and where does the project need stronger fundamentals first?

Contact CD&A Consulting Services Inc. to discuss your transformation goals, readiness concerns, and the practical role AI may play in achieving better results.

Related perspectives from CD&A

W Jones https://www.cdaconsultinginc.com