What the NIRO AI Navigator Is.
The NIRO AI Navigator is a tool for specific AI projects in industrial companies. It organizes the questions that need to be answered before, during, and after a pilot project, and records the answers for each project in one place.
A guide with a working document
Nine chapters for orientation and a fact sheet summarizing the results of a project. Not an AI textbook, not a maturity model, and not a recommendation for a specific product.
Business units, IT, and management working together
For anyone who is responsible for, evaluates, or approves an AI project. Technical terms are explained where they are needed; the most important ones are also listed in the glossary in Chapter 09.
A Project Ready for Decision-Making
A detailed project profile with objectives, key metrics, baseline values, data access, roles, checkpoints, pilot decision, and operational responsibility. Available as a handout for download or printing.
How a Project Progresses Through the Navigator
An anonymized example from the program, condensed into three steps. It shows the types of answers the Navigator requires before technology is selected.
Describe the Initial Situation
Customer specifications are read manually and transferred to an internal requirements document. This process is time-consuming, and omitted requirements lead to follow-up inquiries.
Assess Impact and Feasibility
Processing time per document and the correction rate are measured in the current process. Only then is a comparison made to determine whether an AI-driven workflow offers an advantage over a simpler solution.
Decide on the pilot and implementation
The pilot is limited to a single document type. The success threshold, critical errors, control points, fallback procedures, and decision date are established before the start.
The NIRO KI Navigator was developed through collaboration with companies in the NIRO network. Details on how the program unfolded and what came of it can be found at the bottom of this page.
Where are you in your AI project right now?
There’s no set starting point. Choose the entry point that best fits your current situation. The NIRO AI Navigator does not prescribe a fixed order. It guides you to the questions that should be addressed next.
The NIRO KI Navigator is not a substitute for a technical, legal, or operational assessment of specific cases. It is a document based on experience and insights gained from real-world projects within the NIRO network.
You’re not sure where to start
Identify typical weaknesses in the process, use real-world examples, and describe an initial candidate without giving away the solution.
Find your bearings → Getting Started 02You already have a specific idea
Check the impact, task, data access, and the simplest adequate solution before selecting a technology.
Evaluate the idea → Introduction 03You are already working on a project
Make sure to prepare the pilot, roles, checkpoints, approvals, and the subsequent operation in a binding manner.
Ensure implementation →Four Recurring Tasks in AI Projects
Regardless of how you get started, four tasks crop up time and again in these projects. The chapter references indicate where they are addressed in the NIRO AI Navigator.
Identify and Define the Use Case
- Identify waiting times, errors, rework, search effort, or hard-to-access knowledge
- Describe the task, affected role, frequency, and current procedure
- Identify the desired improvement and the measurable baseline
- Only then compare AI with simpler alternatives
Providing Structured Support for a Project
- Use a consistent format for each use case
- Highlight open issues across the six dimensions
- Continuously document actions, observations, and obstacles
- Make the next steps and responsibilities binding
Systematically Capture Data and Processes
- Verify data access, quality, and level of detail on a use-case-specific basis
- Identify relevant process steps and information flows
- Consider data types and decision problems together
- If necessary, use direct process observation as a data collection method
Clarify Roles, Risks, and Approvals
- Identify subject matter experts and relevant roles
- Define the division of tasks between people and the system
- Review data protection, employee participation, and internal approvals in advance
- Consider training and subsequent operational responsibility
All Chapters at a Glance
This page is divided into three sections. Chapters 01 through 03 provide an initial overview. A complete run-through with a specific project takes about 45 minutes and can be paused at any time.
One sheet per project: the project profile.
The project profile brings together all the details about an AI use case in one place: objectives, data, roles, decisions, and open questions. Use it to structure your project and discuss the next steps with your team.
- Open it online and fill it out as you go. Open the profile in your workspace and enter your results directly into the designated fields.
- Don’t tryto answer everything on the first day. Start with the problem, the affected role, the desired outcome, and the current situation.
- Empty fields provide information. They indicate areas that need further clarification and will be filled in as the project progresses.
- Section I is updated on an ongoing basis. Use it to record what has been done, what goal was pursued, and what obstacles arose.
- Save your results. You can download or print the handout. Your entries will be lost if you reload or close the page.
Your entries will not be saved or transmitted to NIRO or any other parties. They apply only to the currently open page.
- A Project Definition: Company, Title, Type, Affected Dimensions
- B Objective andImpact from Section 04
- C Suitability andApproach from Section 04
- D Data from Section 05
- E Roles andParticipants from Section 06
- F Division of Tasks Between Humans andSystems from Section 07
- G Pilot, Section 07
- H Operation andFurther Development from Section 08
- I Documentation — the ongoingsection: ongoing, one line per measure
- J Reflection: What worked, what got in the way, what new questions have arisen
- K NextSteps—three steps and the next follow-up
This is what a detailed profile looks like
Three anonymized examples from the program show just how concise a useful profile can be. A fully fleshed-out example follows below.
—
—Three steps, six dimensions.
The process follows three steps: clarifying the subject matter, assessing feasibility, and embedding the application into the process. In addition, there are six dimensions through which each project is examined. The third step does not end with implementation but transitions into ongoing operations.
The Three Steps
Select a step to see which questions are answered there.
The Six Dimensions
The six dimensions are not a maturity model. Each project is examined from all six perspectives; the depth to which a dimension needs to be addressed depends on the use case and its current phase. Select a field to view the guiding questions.
Early Warning Signs
These statements indicate that a project has not yet been sufficiently clarified. Select a signal to see the appropriate response.
Practical examples from three areas.
The examples draw on the projects covered in the program. They serve as inspiration and illustrate typical scenarios, but do not constitute a recommendation for any specific technology. This chapter does not include any profile fields; the exercises begin in Chapter 04.
First classify the problem, then select the approach.
Not every problem is an AI problem. The suitability check evaluates the nature of the task, the impact matrix identifies the objectives, and problem classification leads to the appropriate solution family. This section covers sections B and C of the profile.
Quick Feasibility Check
The first two questions determine whether an improvement is worthwhile. The next two questions indicate whether a meaningful test is already feasible. Even an attractive project may not yet be ready for a pilot.
What impact is intended?
A project is only viable if it has an impact on a measurable target. Select an impact category to view the typical key performance indicators.
What function should the solution perform?
First, the desired task is described. Then, standard software, rules, classical methods, machine learning, optimization, and generative AI are compared as possible solution approaches.
Rules, automation, AI workflow, or agent?
These terms are often used to describe successive stages of maturity. In fact, they refer to different forms of control. The key factor is who determines the next relevant step during execution and which actions the system is permitted to trigger.
| Criterion | Rule / Standard Function | Automation | AI-powered workflow | AI agent |
|---|---|---|---|---|
| Who determines the procedure? | The function or rule is completely predetermined | The sequence of steps is completely predetermined | The sequence is predetermined; individual steps use AI for assessment or generation | The goal and boundaries are predefined; the system selects individual next actions at runtime |
| Scope for action | None outside the rule | None outside the defined process | Substantive flexibility within the AI step, but no freely chosen process path | Limited flexibility in the selection and order of permitted actions |
| System Permissions | Read or perform a clearly defined function | Defined actions in connected systems | Defined actions based on AI results and fixed checkpoints | Use multiple tools or systems within explicit permissions |
| Suitable when … | the business logic can be clearly defined | a stable, recurring process is to be automated | unstructured information is processed within a known workflow | the path to the goal varies depending on the situation and cannot be fully predetermined |
| Example | Approval Once a Specified Threshold Is Reached | Create an order according to defined rules | Classify customer emails and generate a draft response for review | Investigate an issue, gather information from multiple sources, and propose the next steps |
| Need for Oversight | Check Rules and Exceptions | Monitor processes, error cases, and interfaces | Define AI results, uncertainty, and human checkpoints | Additionally, ensure action permissions, scope, logging, termination, and fallback options |
The more selection and action rights a system has at runtime, the more important it is to implement restrictions, logging, monitoring, and fallback options.
Realistically assess the data for the specific use case.
First, determine the type of data, then its usability. Both apply to the individual use case, not to the company as a whole. This section completes Block D of the profile.
What types of data are involved?
Different decision-making problems require different types of data. Select all types that are relevant to the project. Multiple selections are allowed.
How useful is the data?
Five criteria, each with four levels. For each row, select the level that best describes the current status . The evaluation distinguishes between accessibility, suitability of content, and prerequisites for stable operation . The decisive factor is the weakest criterion, not the average.
Clarify roles and build trust.
AI projects rarely fail because of the model. They fail because no one is in charge or because the results aren’t used in everyday life. This aspect determines whether the application is actually used in everyday life.
Five Responsibilities That Must Be Clarified
A role describes a responsibility, not necessarily a specific position. A single person may hold multiple roles , or these roles may be filled by external support. Before the pilot begins, no key responsibility should remain unclear.
Management
- Vision
- Goals
- Budget
- Prioritization
Department
- Process knowledge
- Problem definition
- Key Metrics
- Validation
IT
- Data connectivity
- Infrastructure
- Integration
- Permissions
Data / Analytics
- Processes
- Tools
- Interpretation
- Model maintenance
Project Coordination
- Project Management
- Foundations of all disciplines
- Translation between roles
Technical quality alone does not guarantee adoption
Acceptance arises from the interplay of recognizable benefits, comprehensibility, trust, and effective integration into the workflow. A lack of transparency can make usage particularly difficult, especially when making decisions with significant consequences.
The results are applied in real-world workflows and improve the desired metrics.
The department can validate a result and explain any deviations.
Comprehensible presentation, visible uncertainty, and explainable influencing factors.
Implementation is a process of change
Implementations often do not proceed in a linear fashion. Uncertainty, questions, criticism, and refinements should be factored into the plan, without dismissing valid objections as mere resistance .
Get involved early
Involve users and affected stakeholders while the process and control points are still malleable.
Highlight benefits and limitations
Avoid abstract AI promises; instead, explain specific tasks, expected improvements, and known error cases.
Use feedback
Evaluate criticism, workarounds, and corrections as indications of process, quality, or trust issues.
Provide Ongoing Training
Don’t just provide one-time training; instead, update knowledge as new features, error patterns, and responsibilities emerge.
From the data source to the business decision.
A working prototype is not yet an application. Value is only created when the result is integrated into data flows, workflows, responsibilities, and control points . This section covers sections F and G of the profile.
End-to-End Process
Select a step to view the related design questions.
What is the level of decision-making for the application?
The level of decision-making determines how much control, documentation, and coordination are required. It is one of the first decisions to be made, not one of the last.
Preliminary Review of Legal and Governance Issues
The preliminary assessment does not determine whether a project is permissible. It identifies the risks, specialized agencies, and in-depth assessments that may be relevant prior to a pilot or full-scale implementation.
This classification is intended to provide structure and does not replace a case-by-case review of legal, data protection, or employee participation matters.
Minimum requirements for a reliable pilot
Decision Point Before the Pilot
Not a weighted score, but a transparent decision: Are the problem and benefits relevant, is the next learning step justifiable, and are the risks manageable?
- The baseline and desired improvement have been identified.
- The effort required to conduct a robust test has been roughly estimated.
- Risks, critical errors, and a fallback option are described.
- The owner, scope of the pilot, and decision date have been determined.
After the pilot, the real work begins.
Most guidelines end with the implementation phase. But that’s when the longest phase begins for the company: The application must keep running even when no one is watching it, when source systems change, and when the person who built it has moved on to other tasks.
Handover to Production
The transition from project mode to routine operation is a distinct step and not a side effect of implementation. Without designated responsibility, the application will fail at the first sign of trouble.
Is the application actually being used?
An often-overlooked metric. A technically flawless solution can be bypassed in everyday use without this being immediately apparent in the model or system quality.
Result Quality and Impact
Two separate questions: Is the result getting worse? And is the target value changing relative to the baseline? Without the second measurement, the benefit remains merely an assertion.
Changes to Sources and Dependencies
An ERP update, a modified form, new master data logic, a different model, an adjusted prompt, or a changed platform can affect the application without anyone noticing.
Traceability and Versioning
Continues the preliminary review from Section 07 in the operational setting. Without a history, it is impossible to reconstruct, after an error, which status led to which decision.
Ongoing Costs
Project costs are one-time expenses, while operating costs are ongoing: licenses, usage fees, computing power, database maintenance, and the time spent on testing and support.
People and Knowledge
Many applications depend on a single person. If that person leaves the company or the service provider changes, the solution can no longer be maintained, even if it continues to run.
Review: Continue, Adjust, or Shut Down
A fixed date with pre-agreed criteria. Shutdown is a legitimate outcome: an application with no impact wastes attention and undermines trust in future projects.
From the First Use Case to the Second
Key metrics from the first project rarely remain. What remains are the reusable elements: clarified data access, defined roles, validated approval workflows, and an evaluation metric.
Frequently asked questions from practical implementation.
The answers summarize insights from the workshops and projects and must be applied to the respective use case.
A Brief Explanation of Terms
Only the terms that are actually needed to work with the Navigator.
- Baseline
- The current process or a simple benchmark against which the pilot is evaluated.
- Granularity
- The level of detail in the data, such as order, item, machine, second, or month.
- AI Workflow
- A largely predefined process in which one or more steps utilize AI for evaluation or generation.
- AI Agent
- A system that selects individual next actions at runtime within defined goals, rights, and constraints.
- Fallback
- The agreed-upon fallback to a secure or previous workflow if the application fails or malfunctions.
- Owner
- A specifically named person who assumes professional responsibility and makes decisions.
From the initial idea to the operational decision.
This anonymized example takes the case from the overview through all six stages. It is intentionally concrete, without presenting fictitious project results or key metrics.
Initial Situation
Customer specifications are read manually and transferred to an internal requirements document. This process is time-consuming; omitted requirements lead to follow-up inquiries.
Impact and Baseline
The goal is to reduce processing time and ensure a more traceable transfer of data. Prior to the pilot, processing time per document and the correction rate in the current process will be measured.
Solution Approach
An AI-driven workflow extracts requirements according to a defined sequence. Subject-matter experts remain responsible for the transfer; an autonomous agent is not required.
Data Status
Representative specifications and the internal target format are available. Terminology, document variants, usage rights, and access to legacy documents are clarified in advance.
Pilot and Decision
The pilot is limited to one document type. Success thresholds, critical omissions, a human checkpoint, a fallback plan, and a review date are defined before the start.
Operation
For a rollout, operational responsibility, versioning, utilization rate, ongoing costs, and triggers for reassessment are documented.
The profile form to fill out.
The work area is initially hidden to keep the Navigator page clear. The field labels correspond to the instructions in sections 02 through 08. Clicking on a field label in the flow text takes you directly to the corresponding field. Your entries apply only to the currently open page. They will be lost when you reload or close the page.
Ready to get started?
Open the complete online profile or start by downloading the comprehensive case study as a starting point.
AI Project Profile
One line per action. The “Goal” column is crucial: it explains why this step comes next. Observations and obstacles also belong here. These are the things you’ll need to report back to the group later.
| Date | Context / Measure | Objective of this measure | Challenges / Observations | Dimension | Status | Row |
|---|
From the NIRO program for practical application.
The NIRO AI Navigator draws on the experiences gained from a collaborative development and learning process within the NIRO network. It makes the key questions from the completed corporate projects available for use in future AI projects.
Orientation Before Technology
The starting point was the question of how companies can identify suitable AI projects, assess them realistically, and integrate them with benefits, data, organization, and operations from the very beginning.
Developed collaboratively
Together with companies in the NIRO network, we addressed and evaluated specific starting points, application ideas, obstacles, and potential solutions.
Two Working Documents
The NIRO KI Navigator serves as a guide and a tool for review. The profile document records the results, decisions, and outstanding issues of a specific project as an ongoing working document.
The content is based on practical experience gained through the program. It does not constitute a representative impact study, nor does it represent a technical, legal, or operational review of specific cases.
The NIRO KI Navigator improves with your feedback.
The content is derived from the projects of the participating companies. Your feedback on obstacles, solutions, and any questions that haven’t been addressed will help refine the current status and further advance the use of the NIRO AI Navigator within companies.
Obstacles and Barriers
Where did things get stuck? Data access, interfaces, approvals, acceptance, costs—categorized into one of the six dimensions.
Solutions That Worked
Even small ones: a phrase that convinced management, a way to get an approval, a usable data extract.
Consultations and Stakeholders
Who needed to be involved, in what order, and with how much advance notice? This kind of experience is hard to find in books but invaluable in everyday practice.
What’s Missing from the Guide
What questions arise in your company that aren’t answered here? This helps identify where the NIRO AI Navigator has its limitations.