Glossary
The terms you hear in a steering committee, a software-vendor demo or an AI workshop, explained in plain language with what they actually mean for a decision-maker. Written by PlanAxion's partners, an independent IT advisory firm in Montreal.
Why a glossary
“Scope,” “stabilization,” “pilot,” “standard”: in an ERP or AI project, finance, IT and the implementation partner rarely mean exactly the same thing. This glossary sets the definitions we use in our engagements so decisions are made from a shared understanding.
Category 1
The terms that come up from the first software demo through go-live, including the ones you will see in an implementation partner's contract.
An ERP is a single system that manages an organization's core processes: finance, purchasing, sales, inventory, production and payroll. Its value is a shared data foundation across departments. Its risk is reproducing poorly understood processes inside a system that is expensive to change.
ERP selection is the process from defining requirements to recommending software and an implementation partner. A strong RFP compares solutions against the same business scenarios instead of letting each vendor run an open-ended demo. Many of the decisions that shape the project are made here.
Implementation is the phase from signed contract to go-live: design, configuration, data migration, testing, training and cutover. For an organization with 300 employees or more, a 12-to-24-month timeline is common. Duration depends on the number of modules, integrations and the quality of the data.
The implementation partner is the firm that configures and deploys the software for the client. It is paid to deliver the solution, either by time and materials or fixed fee. It is essential, but it does not represent the client when scope and budget trade-offs are made.
The software vendor is the company that develops and sells the ERP, such as SAP, Oracle or Microsoft. Its business depends on licences and subscriptions. Its sales team has an incentive to sell more modules, which is not always the same as what the organization actually needs.
An independent ERP consultant advises the client without selling software licences or implementation hours. The consultant frames requirements, compares options and protects scope and budget during the project. The goal is a decision the client can defend. That is PlanAxion's role.
S/4HANA is SAP's current-generation ERP platform, available in cloud and on-premise deployment models. Moving from SAP ECC is a transformation in itself, with public, private or hybrid architecture choices that can shape the organization for years. Standard ECC maintenance ends in 2027.
Oracle Fusion Cloud ERP is Oracle's cloud suite for finance, procurement, projects, HR and other enterprise functions. It is progressively replacing Oracle E-Business Suite. Moving to the cloud generally means adopting more standard processes instead of rebuilding every customization from the legacy system.
EBS is Oracle's long-established ERP suite, typically deployed in the client's environment. Many Quebec organizations still use it. The question is not only whether to migrate, but how much value and useful life remain in the current investment.
Fit-to-standard means adapting the organization's processes to the capabilities delivered by the software. Customization means changing the software to preserve existing processes. Every customization adds cost during implementation and again during future upgrades.
Data migration is the transfer, cleansing and validation of data from the old system to the new one: customers, vendors, products, open balances and history. It is a frequent source of go-live delays because data-quality problems are often discovered late.
Go-live is the point when the organization stops using the old system and starts operating in the new one. It should be prepared with a detailed cutover plan, explicit go/no-go criteria and a rollback plan. You do not go live because a status report is green; you go live because the readiness criteria are met.
Stabilization is the period, often four to twelve weeks, after go-live. Project teams remain mobilized to fix issues, adjust processes and support users. A project is not truly finished on cutover day; it is finished when the new operating model is stable.
A legacy application is an older system, often written in COBOL, PowerBuilder, VB6 or Delphi, that still supports critical business processes. Important business rules may live in the code and in the knowledge of a few people. The risk grows as that knowledge becomes harder to retain.
Category 2
The terms that determine whether a project stays under control, and the ones you start hearing more often when it does not.
Project scoping defines the objectives, scope, responsibilities, constraints and success criteria before the full budget is committed. A focused scoping phase can prevent months of disagreement over what the project was supposed to deliver.
Scope is everything the project is expected to deliver: processes, modules, sites, entities and integrations. Anything that is not clearly included can later become a change request. Precise scope is one of the first tools for controlling budget.
Scope creep is the gradual addition of requests without measuring their impact on cost, resources or delivery dates. A warning sign: scope keeps growing while the budget and schedule stay unchanged in the status report.
Project governance is the set of roles, forums, decision rights and escalation rules that determine who decides, based on what information and with what impact. Good governance does not create more meetings; it makes important decisions visible early enough to act.
The steering committee brings together client decision-makers, and often the implementation partner, to resolve issues involving scope, budget, risks and dates. It is useful when it makes and records decisions. It is not useful when it only listens to a green status report.
The decision log records every required decision, its owner, due date, impact and status. It is one of the clearest differences between a project that is actively governed and one that moves forward by default. PlanAxion maintains it on the client side.
A dependency is a condition one team must satisfy before another can move forward: validated data, a delivered interface or a decision that has been made. Dependencies between business teams, IT and the implementation partner are where delays often hide outside each team's individual plan.
The critical path is the sequence of tasks where any delay pushes back the project's final delivery date. Knowing it helps focus attention on the few dependencies that actually affect timing instead of the many tasks that do not.
A project diagnostic is an independent assessment, usually completed over two or three weeks, of the project's real situation: gaps, causes, dependencies and blocked decisions. It comes before a recovery plan. Without a diagnostic, teams often treat symptoms instead of causes.
Project recovery is the intervention used to regain control of a struggling initiative: diagnose the situation, stabilize critical risks, put decisions back in the right hands and build a realistic relaunch plan. It becomes necessary when the project can no longer produce clear enough decisions to protect scope, budget or schedule.
A Project Management Office structures how an organization governs its portfolio of projects: methods, prioritization, tracking and executive visibility. It can be internal or delivered as an external PMO service, part-time or full-time.
Change management is the set of actions that prepares teams to work differently: impact analysis, manager engagement, communication and training. It starts during project scoping, not three weeks before go-live.
Adoption measures whether teams actually use the new system or process as intended, without workarounds. A delivered system with weak adoption has not created its expected value. Adoption can be measured through usage, work still happening outside the system and support demand.
An early warning signal is an observable fact in day-to-day project work that points to a problem before it appears in the budget or schedule. Example: the same issue returns in three committees without a decision. PlanAxion uses these signals as a starting point for project diagnostics.
Category 3
The terms leadership teams hear when someone asks “what should we do with AI?”, and the ones that determine whether a pilot can move into production.
An AI use case is a specific business problem experienced by a clearly identified team where AI can produce a measurable result, such as triaging requests, matching payments or extracting information from documents. A strong use case can be tested within a defined scope; “deploy an AI assistant everywhere” is not a use case.
An AI pilot is a limited test of a use case with a defined scope, timeline and success measure set before work begins. Its purpose is to decide, with evidence, whether to scale, adjust or stop. A pilot without a success metric is a demo.
An AI roadmap is the prioritized list of selected use cases, including their sequence, dependencies such as data, systems and teams, and the next decisions to make. It should be the output of prioritization work, not the starting point.
An AI workshop is a short, structured engagement, four weeks at PlanAxion, that evaluates business ideas against four criteria: problem, value, data and effort. The output is a prioritized set of decisions: launch, prepare, reframe or defer. It is a management decision process, not a training session.
Data quality measures whether the data required for a use case is complete, accurate, current and accessible. AI pilots often fail when the business value is clear but the history is incomplete or captured inconsistently. Data quality should be checked before launch, not after.
Data governance defines who is responsible for each data set, how it is defined, where it is stored and who can access it. Without it, two reports can answer the same question with different numbers, and AI models cannot be trusted consistently.
Generative AI refers to models that can produce text, code, images or other content from natural-language instructions. In organizations, practical uses include drafting, summarizing files, reading documents and assisting with legacy-code modernization, with human validation where needed.
Agentic AI refers to systems that can chain multiple actions to complete a broader task, for example reading an invoice, matching it to a purchase order and proposing an accounting entry. Major ERP vendors are increasingly adding these capabilities to finance modules. The key decision is control: which actions can the agent take on its own, and which require human approval.
Process automation replaces repetitive manual work, such as data entry, transfers, approvals and reminders, with workflows executed by software, with or without AI. A strong candidate is frequent, predictable, time-consuming work whose data already exists in the current systems.
Cash application is the process of matching incoming customer payments to open accounts-receivable invoices. It becomes difficult when payments are partial, grouped or poorly documented. It is a strong AI and automation use case in B2B distribution because much of the required data already exists in the ERP.
AI-assisted application modernization uses generative AI tools to read, document and translate legacy application code into modern technology stacks. AI can accelerate analysis and code generation; validation, architecture and final decisions remain human responsibilities at each stage.
Quebec Law 25 governs the protection of personal information in private-sector organizations, including consent, privacy impact assessments, accountability and privacy incidents. ERP or AI initiatives that process customer or employee data need to account for it in the architecture and governance.
The ROI of an AI project can be measured through hours recovered, errors avoided, cycle time reduced or revenue protected, compared with the cost of the pilot and scaling. The success measure should be defined before the pilot, not added afterward as justification.
Process optimization simplifies an existing way of working before adding technology: remove steps that create no value, clarify responsibilities and reduce workarounds. Automating a heavy process without optimizing it first can simply hard-code the problem.
Useful pages
For topics that require scoping, selection or independent client-side support, these PlanAxion pages go deeper.
Beyond the definitions
ERP selection, a project drifting off course, or your first AI use case: describe the situation in a few lines. A PlanAxion partner will reply with an initial perspective within one business day.
Describe my situation