How to map a business workflow across systems: a quote-to-cash example

Why map a workflow instead of a system?#

Mapping a workflow shows how one piece of work moves through the business, and that movement is what gives operational records their meaning. A list of systems tells you a CRM, an ERP and a helpdesk exist; a workflow map tells you whether a customer complaint can be traced back to the quote and order that caused it.

That linkage is also what AI developers look for. Records that connect a request to the decisions made and the outcome that followed are more useful than large but isolated archives, because they show cause and effect in real work rather than disconnected snapshots.

Five steps to map any workflow#

Mapping a workflow takes a whiteboard, the people who run each step and access to a few real transactions. No exports are needed to start, and the first draft is usually rough on purpose.

Invite the person who handles exceptions, not only the process owner. The documented process describes how work should flow; the exceptions desk knows how it actually flows, including the workarounds that never reached a system.

• 1Step 1: name the workflow and its start and end events, such as first quote request to final payment or warranty close.
• 2Step 2: list each handoff and the system where the record lives at that point.
• 3Step 3: for each handoff, write down the ID that links the record to the previous step and the field that stores it.
• 4Step 4: trace real transactions end to end, including some from earlier years and some that went wrong.
• 5Step 5: mark every place the trail breaks, and note whether the link survives in free text, email or nowhere.

Illustrative: quote to cash at a fluid power distributor#

Illustrative: a fictional distributor of hydraulic components sells to equipment manufacturers. Inside sales manages opportunities in HubSpot, builds quotes in a quoting tool connected to the ERP and converts accepted quotes into sales orders in NetSuite.

The warehouse picks and ships from a WMS that receives orders from NetSuite and sends back tracking numbers. Invoices and payments post in NetSuite. When a customer reports a wrong part or a failed component, customer service opens a Zendesk ticket, which may lead to a return authorization and a credit memo back in NetSuite.

Each step produces a record a person could read on its own. The map shows whether those records can be read together, as one story from quote to resolution, and where that story goes silent.

The sample trace shows the chain holding from sales order to invoice across the whole NetSuite period, while quotes from before the quoting tool exist only as PDFs in email. The COO scopes a first package around orders, exceptions, returns and tickets from the years where every link holds, and parks the older quote history as a separate, later question.

The join keys that hold the chain together#

Join keys are the IDs one system stores about another, and they decide whether a quote-to-cash chain can be reconstructed. In the distributor's case, the map shows the following links and the places each one tends to fail.

Write down the field that stores each key, not just its name. A quote number that lives in a dedicated NetSuite field can be joined reliably; the same number typed into a memo line or an email subject needs text matching and will miss some records.

Customer and item masters sit under every link#

Customer and item master records are the hidden join keys beneath the whole chain. Where HubSpot, NetSuite and Zendesk each assign their own customer ID, a cross-reference table is needed, and the map should name who maintains it and how often it is wrong.

Item numbers matter just as much in distribution. If a part was superseded, renumbered or set up twice, a complaint about a failed component may not link to the order line that shipped it. Note on the map where item history was cleaned up and where duplicates still exist.

Where chains usually break#

Workflow chains usually break where people step outside the systems. Approvals given by email, quotes sent as attachments, phone calls logged as a single note and spreadsheets used to track exceptions all leave gaps that structured exports cannot fill.

System changes cause the other big breaks. An ERP migration may renumber customers and orders, an acquired branch may run its own system for years, and merged duplicate customer records can attach old history to the wrong account. Note the date of each change on the map, because records before and after it may need different join logic.

Check linkage on a sample before anyone exports#

Linkage should be checked on a sample of real transactions before export planning starts. Pick orders from several periods, including ones with returns and complaints, and try to rebuild each from quote to resolution using only the systems' own fields.

The result is a short document anyone can read: the workflow, the systems, the keys and the periods where the chain holds. That document is usually enough for a first conversation about value, and it tells IT which exports will be worth the effort.

• 1Record which links were found in a structured field.
• 2Record which links appeared only in free text or attachments.
• 3Record which links were missing entirely.
• 4Note the period and system version for each result.
• 5Summarize which stretches of history can be rebuilt reliably.

How SourceX uses workflow maps#

Workflow maps are metadata, so they can be shared during the Supply step of the SourceX five-step transaction without moving any records. In the SourceX Enterprise Data Value Framework, a chain that holds supports AI utility and data cleanliness, while a chain that breaks raises preparation cost, so a map is useful evidence early. Later, the join keys guide preparation, so personal identifiers can be removed while the links between records survive.