Know where your information starts and everything that depends on it.
You can trace any piece of information through your processes: where it is created, which forms, documents and systems carry it, who owns it and which steps use it. The knowledge to do it is already in your organisation. The route is to connect it.
Your team is about to give an AI agent access to the customer onboarding process. Someone asks which forms feed it, which system holds the master record and who signs off the documents it will read. Three people answer from memory, and each is partly right. With the connections recorded, the answer is one view that anyone can open.
You already hold most of the pieces.
Each document you have describes one corner of the picture. Lineage is what you get when the corners are joined.
Process maps name the steps and the roles that run them. SOPs name the documents and the systems. Forms name the information being captured. RACI matrices name who is accountable. System inventories name where it all lives.
Nothing here needs to be redrawn. Each of these sources already says something about where information starts and where it goes. They sit in different tools, drawn in different years by different teams, so each one answers only its own corner. Join them with shared identifiers, and a question that spans several of them gets one answer.
Five questions you can answer on demand.
Once processes, roles, systems, documents and information are linked, these stop being projects and become lookups.
What connects to this process?
The roles that run it, the systems it uses, the documents that describe it and the information it handles, in one view.
Where does this information start?
The first step, form or system that creates it, and who owns it.
Where is it used afterwards?
Every downstream step, report and decision that relies on it.
Is the same thing captured twice?
The forms and systems that record the same item, so you can nominate one source of record.
What would a system outage touch?
The steps, roles and documents that depend on it, before you plan the change or the recovery.
One customer address, followed from capture to use.
An illustration of the method, not a client case. It makes no claim about results.
Take a customer address in an onboarding process. The process map shows three steps that touch it. Linking the documents you already have answers the rest.
List what is documented
The onboarding map, the three forms and the system inventory entry are the starting point. Nothing new is drawn.
Link each to the item
Each form, system and downstream step is linked to the customer address, with an owner against it.
Read the picture
The address is captured three times and held once. The record system is the natural source of record. Three steps depend on it, so you know what a change to it will touch.
This picture comes from your documentation, so it shows where information is meant to flow. To confirm how the work runs today, our discovery services gather evidence from the people who do it and from event logs. Start with the documents, then check the picture against the work.
Every link shows its source, its age and its owner.
A connected picture is only useful if you know which parts you can stand behind.
IGX360 Insights works over the repository and the documents around it, such as iGrafx models, SOPs and Visio files. It brings them into one model you can query, and it shows, for each link, where the information came from, how confident you can be in it, how recent it is, who owns it and where two sources disagree.
That tells you which parts of the picture are ready to act on and which need a review first. It reads what is documented and modelled. The model does not replace your repository or your modelling notation, and it does not assume you have one.
Give your first AI project something true to start from.
An agent reads what you point it at. Lineage tells you what that should be.
A read-only project on your own process knowledge is a safe first step. It shows which steps have one clear origin for their information and one owner. Those are the steps to automate now. It also shows where two sources disagree, so you know where a human gate belongs before an agent acts.
The result is a first project that lands, and a record of what is true that every later project can start from. Each one needs less groundwork than the last.
The same connections answer different questions for different roles.
Information lineage across your processes: frequently asked questions
What is information lineage across a process?
It is the record of where each piece of information starts, which forms, documents and systems carry it, who owns it, and which process steps use it. With it, you can follow an item forward to everything that depends on it, and backward to where it came from.
How do you connect processes with roles, systems, documents and information?
Start from what you already hold. Process maps name the steps and the roles, SOPs name the documents and systems, and forms name the information captured. Give each of them a shared identifier and link them into one model. A question that spans all four then becomes a query, not an afternoon of opening files.
How do you find duplicate data capture across forms and systems?
List the forms and systems documented for a process, record which information each one captures, and compare the lists. Where the same item appears in more than one place, you have a candidate to consolidate or a source of record to nominate. Documentation shows where capture is meant to happen. Confirming what people do in practice is work for discovery, with the people who do it.
How do you trace where information starts and where it is used?
Pick one item and find the step that creates it. Follow the documented links forward through forms, systems and downstream steps, then backward from a report or decision to the sources behind it. Record the owner and the last review date at each hop, because each hop is only as reliable as the document behind it.
How can you tell what a business system failure would affect?
Link every system to the process steps that use it and to the people who own those steps. The question "what touches this system?" then returns the affected steps, roles and documents in one view. That answer comes from your documentation, so confirm it with the people who run the work before you rely on it for an incident plan.
What software creates a queryable model of processes, systems, roles, documents and controls?
A process repository such as iGrafx Process360 Live holds the models. IGX360 Insights works over the repository and the documents around it. It brings them into one model you can query, and shows provenance, confidence, freshness, ownership and where two sources disagree.
Where should an AI project start if the information is scattered?
Start with a read-only project on your own process knowledge. It shows which steps have one clear origin for their information and one owner, and those are safe to automate now. It also gives every later AI project a record of what is true to start from.
Go deeper.
Enterprise process architecture: the structure lineage sits inside
Read the definitionProcess intelligence: what a connected model makes possible
Read the definitionIGX360 Insights: what you can stand behind in your process knowledge
See IGX360 InsightsStart the conversation
Trace one process from start to finish with us.
Pick a process you are about to automate or change. We will show what connecting the documents you already hold tells you about it.