Smart Manufacturing Is Nothing Without Connected Data: Why Most Factories Aren’t Ready

Smart Manufacturing Readiness: What to Fix First

To assess smart manufacturing readiness, choose a production problem and check whether your processes, data, people and systems can support the proposed improvement. Identify missing prerequisites, give each gap an owner and test a limited change against agreed measures before expanding. You do not need every machine connected or an AI model in place to begin.

Perhaps you want planners to spend less time chasing updates. Operators may be entering the same information twice, or quality staff may struggle to retrieve inspection records. These are useful starting points because the problem and the people affected are identifiable.

The assessment should help you decide what to do next: improve a process, repair an existing connection, introduce a tool or investigate further.

Start with a decision that needs improving

“Build a smart factory” gives a project team little guidance about where to spend its time.

A more useful objective describes a decision or task that is currently difficult. For example, a planner needs to identify which jobs can start tomorrow without calling production, stores and quality separately.

Follow that task through the working day. What information does the planner need? Where does it come from? Who confirms it? What happens when it is missing?

Record the current effort and consequences. That could include time spent collecting updates, schedule changes caused by incorrect information or work released before a prerequisite was satisfied.

Keep the scope manageable. One production area and a recurring job type can provide enough detail to establish whether an approach is worth extending.

Check whether the process is understood

Before automating a task, ask the people doing it to explain the normal route and its exceptions.

Two departments may interpret “ready” differently. Production may mean that a machine is available, while quality expects material release and approved instructions to be confirmed first. Automating the status change would preserve that disagreement.

Agree the conditions that matter to the proposed improvement:

• What starts the activity?

• Which inputs must be available?

• Who makes or approves the decision?

• What indicates completion?

• Which exceptions require review?

The process does not have to cover every imaginable situation before work begins. It does need a defined boundary and a clear response when something falls outside it.

For additive manufacturing, test the relationship between builds and parts. Completing a shared build does not establish that every part has finished its individual post-processing and inspection route.

Assess the information needed for that task

A factory does not need all its data in one system to make progress. It needs dependable information for the activity being improved.

For each important input, establish its source, meaning, identity and update expectations.

Question	What to establish
Where does it come from?	The authoritative record and its owner
What does it mean?	Agreed units, status definitions and calculation rules
What does it belong to?	The relevant job, part, build, material or resource
How current must it be?	An update frequency appropriate to the decision
What if it is missing?	How the gap is recognised and who responds

Consider material availability. A quantity in stock may include material awaiting review, already reserved for another job or unsuitable for the proposed process. A planning improvement needs the definition that supports scheduling, rather than whichever stock figure is easiest to retrieve.

Check a sample against physical work. Can the team explain a discrepancy between the system and the shop floor? Resolving that question may be more valuable than adding another dashboard.

Test whether people can use the proposed process

Readiness includes the ability to work with the change during an ordinary shift.

Ask operators, planners and quality staff to complete the relevant task using the intended devices and information. Observe where they hesitate, leave the system or ask someone else for help.

A required update may be straightforward at an office desk but awkward beside a machine. A review process may work during the day while depending on an approver unavailable overnight.

Check that people have:

• Suitable access and equipment.

• Time to perform the task.

• Training relevant to their role.

• A clear way to report problems.

• Support during the hours they work.

Include the effort of maintaining the information. Saving a planner time by adding a much larger recording burden for operators may shift the problem elsewhere.

Ask what people would still keep in a separate spreadsheet. Their answer can reveal a missing function, a usability problem or information they do not yet trust.

Examine connections and failure handling

A successful data transfer is only one part of integration readiness.

Follow a representative update from its source to the place where someone uses it. Confirm that the receiving record contains the correct value and retains its association with the relevant work.

Then test a delay, a duplicate message and an unavailable connection. Who sees the failure? Can normal work continue? How are records reconciled when the connection returns?

Also test access using the intended roles. A person needs enough permission to complete their work, with restrictions appropriate to the information and actions involved.

These checks help establish the support burden as well as technical fit. A connection that requires frequent manual repair needs an owner, available expertise and time allocated to maintaining it.

Decide how much automation the task needs

Some improvements need a clearer view of the work. Others benefit from a notification, a rule or a recommendation. Direct control of equipment introduces a different level of responsibility.

Choose the approach that suits the task.

Approach	Example	What to establish first
Visibility	Show jobs waiting for inspection	Reliable job status and a usable view
Notification	Alert a reviewer when evidence is missing	A meaningful trigger and named recipient
Rule-based automation	Create a task when a defined condition is met	Agreed rules, exceptions and error handling
Predictive analysis	Estimate the likelihood of a particular failure	Suitable data, a defined target and tested model performance
Direct process control	Adjust an equipment parameter automatically	Appropriate engineering controls, limits and validation

AI can support a specific forecasting or analysis problem. Its suitability depends on the data available and the consequences of an incorrect result.

For example, an anomaly alert may warrant investigation. It does not by itself establish that a part is defective. Likewise, a failed inspection needs the agreed review and disposition process before replacement or rework is authorised.

A useful first project may simply make those decisions easier to perform and record.

Turn the assessment into an owned action list

The review should produce a manageable set of actions. Record each gap, its consequence, the work needed and the person responsible.

An illustrative action list might look like this:

Finding	Consequence	First action	Typical owner
“Ready” has different meanings across teams	Unsuitable jobs enter the schedule	Agree prerequisites and status definitions	Production process owner
Material release status is unavailable to planning	Planners chase confirmation	Confirm the source and provide an appropriate view	Materials and planning leads
Operators record progress at the end of a shift	Other teams act on old information	Test a practical way to capture updates during work	Production supervisor
Failed transfers have no recipient	Errors remain unresolved	Assign monitoring and recovery responsibilities	Integration owner
An automated action lacks an exception route	Unusual work follows the wrong path	Define review conditions and authorised actions	Process and quality owners

This is an internal planning exercise, rather than a certification scheme.

Prioritise gaps by their operational consequence and the improvement they prevent. An unclear release rule may need attention before an advanced forecasting project. Missing historical data may require a period of collection before a model can be evaluated.

Sequence dependent work. There is little value in testing automatic task creation while the team still disagrees about the condition that should trigger it.

Test one improvement before widening the scope

Choose a project with a clear boundary, an accountable owner and an outcome you can observe.

Agree acceptance criteria before configuration starts. For a planning view, those might include identifying eligible jobs, recognising stale information and finding the owner of a blocked item.

Include routine work and relevant exceptions. Record what passes, what needs a workaround and what remains untested.

Our guide to running an additive manufacturing workflow software pilot explains how to select representative jobs and turn evaluation questions into practical tests.

Compare the result with the baseline using consistent definitions. Measure the effort of collecting and maintaining information alongside the time saved by the people using it.

If the change works, identify what must be repeated or adapted before extending it to another area.

Include ongoing ownership and cost

A project can work technically and still be difficult to sustain.

Establish who maintains the configuration, reviews access, investigates errors and trains new users. Confirm how changes to production routes, connected systems or reporting requirements will be handled.

Include those responsibilities in the investment assessment. Subscription charges are only part of the cost; setup, integrations, training and ongoing support may also require budget.

For AM teams evaluating software, our guide to comparing the full cost of an additive MES helps make those assumptions explicit.

Give the benefit an owner too. Someone should confirm whether the change has reduced the identified problem and decide what to adjust if it has not.

Where Authentise fits

Where the gap involves coordinating additive production, Authentise FlowsAM connects scheduling, machines, materials and production records. That provides a relevant starting point for examining how planning and production information could be brought together. FlowsAM by Authentise

Use the readiness assessment to shape that discussion. Bring a representative job, the decision you want to improve and the information required to make it.

Ask the team to demonstrate the proposed workflow, including missing information and exceptions. Agree which capabilities are available in the proposed configuration and which require integration or additional work.

Choose the next project you can support

Smart manufacturing readiness is useful when it leads to a concrete decision.

Select one production problem, establish what the improvement depends on and assign responsibility for the gaps. Test the change with the people who will use it, then assess whether it has delivered the intended result.

If your priority is better coordination across AM planning and production, explore FlowsAM using that specific problem as the basis for the conversation.