Skip to main content

Glossary

This section defines a single English vocabulary for the methodology. It is needed so that business, the AI office, delivery teams, architecture, security, and governance bodies share the same understanding of what is happening to an initiative at every stage.

The documentation uses terminology aligned with the AI Conveyor platform: business funnel, product delivery track, initiative card, stage gate, required fields, transition rules, tasks, analytics, and the AI assistant.

Business problem​

A business problem is a pain, constraint, or improvement opportunity inside a process, product, or function of the company that can be linked to measurable indicators.

Examples:

  • a high volume of repetitive manual operations;
  • long processing times for requests, documents, or tickets;
  • dependence of a process on a few key experts;
  • weak transparency over status, decisions, and accountability;
  • a high share of errors and rework;
  • lack of forecasting and early warnings;
  • difficulty finding knowledge inside the organization;
  • the need to analyze data volumes that are hard to process manually.

A business problem is the starting point of an initiative. Without a clear problem an initiative quickly turns into a technology experiment with no owner and no impact.

Idea​

An idea is a first hypothesis about how AI can help solve a business problem.

An idea typically contains:

  • a short description of the problem;
  • the proposed approach;
  • the initiator;
  • the unit or business process involved;
  • a first hypothesis about the expected impact.

In the platform an idea can be created manually, generated in batches based on the company profile, or captured via the AI assistant, which builds a brief, checks similar initiatives, and helps select a product.

Initiative​

An initiative is a formalized unit of management in the AI conveyor.

An initiative records:

  • the business problem and a description of the current process;
  • the expected impact and baseline indicators;
  • the initiator, owner, and project lead;
  • the chosen product or a set of products;
  • the current stage of the business funnel;
  • the stages of product delivery;
  • tasks, artifacts, reviews, and the history of decisions.

All initiatives form a portfolio. The portfolio is not just a registry — it is the basis for deciding which ideas to pursue, which to stop, where teams are overloaded, and where impact is expected.

Use case​

A use case is a structured description of an initiative: which process is changing, which task is being solved, what data is needed, which product fits, and which metrics should improve.

A use case typically includes:

  • a description of the current process;
  • the problem or inefficiency;
  • target indicators;
  • the proposed solution;
  • the expected impact on metrics;
  • data requirements;
  • security, architecture, and operational constraints;
  • the product through which the solution will be delivered.

A use case is used to evaluate an initiative and to pass the stage gate before delivery.

AI product​

An AI product is a reusable solution, platform, or delivery loop through which initiatives of one class are implemented.

Examples of AI products:

  • a corporate large language model;
  • a knowledge assistant with search over internal documents;
  • a machine learning platform;
  • process automation;
  • meeting transcription and analysis;
  • a code agent for software development.

Whenever possible, initiatives should be delivered through existing products rather than starting from scratch each time. This lowers cost, accelerates delivery, and makes portfolio management predictable.

Framework​

The framework is the set of components that build the AI operating model: operating model, AI product portfolio, AI initiative portfolio, playbooks, artifacts, and related sections.

It answers what the system is made of — not the product and not the adoption methodology.

Methodology​

Methodology is how to implement the framework in the organization: stages, governance, rituals, moving from pilots to a managed loop.

Described on Methodology. The AI conveyor product runs the framework in the platform.

AI conveyor​

AI Conveyor is the product (platform) to run the framework: lifecycle from idea registration to confirmed impact.

Do not confuse it with the framework’s management loop or with methodology.

The canonical lifecycle along the business funnel:

New
↓
Assessment
↓
Delivery
↓
Awaiting impact
↓
On support / Closed / Rejected

In the platform these stages map to the states NEW, ASSESSMENT, DELIVERY, AWAITING_EFFECT, ON_SUPPORT, CLOSED, and REJECTED.

In parallel each product can have its own delivery track: for example, scoping, prototype, validation, and rollout. The business funnel shows the path to value; the delivery track shows how the work is being delivered inside the chosen product.

Stage gate​

A stage gate is the rule that governs moving between stages. It answers the question: can the initiative move forward, does it need rework, or should it be stopped.

A stage gate may verify:

  • that required fields are filled;
  • that a product has been selected;
  • that the similar-initiatives check has been passed;
  • that there is no negative security verdict;
  • that the required artifacts exist;
  • that data is ready;
  • that owners and accountabilities are assigned;
  • that impact has been confirmed.

In the platform stage gates are configured through transition rules, required fields, and stage-based field visibility.

Artifact​

An artifact is a document, record, or structured work product that confirms the initiative is ready for the next stage.

Examples of artifacts:

  • the initiative card;
  • the initiative brief;
  • the similar-initiatives report;
  • the financial model;
  • the security verdict;
  • the architecture review;
  • the test program and methodology;
  • the delivery report;
  • the impact report.

The set of artifacts depends on the maturity of the organization and the product. What matters is not the number of documents but whether they support better decisions and lower risk.

Impact​

Impact is a measurable change in business indicators after a solution is adopted.

Impact may be expressed as:

  • cost reduction;
  • revenue growth;
  • shorter process time;
  • lower operational risk;
  • improved quality or accuracy;
  • less manual work;
  • faster decision-making.

Impact must have a baseline value, a target value, a calculation method, an owner, and a verification date. Otherwise the initiative remains a "successful launch" rather than a proven business outcome.

AI Operating Model​

The AI Operating Model is the target system in the company: how the organization selects, launches, delivers, and scales AI initiatives.

Built from the framework and adopted through methodology. AI conveyor runs the framework in the product; the model is not the conveyor.

Framework components include:

  • roles and accountabilities;
  • the business funnel and delivery tracks;
  • decision-making rules;
  • stage gates;
  • portfolio analytics;
  • management of AI products.

The operating model exists so that AI stops being a collection of disconnected experiments and becomes a managed mechanism for creating value. Adoption follows methodology; day-to-day execution can use AI conveyor.