Repetition
How often the same operational action appears and how much manual capacity it currently consumes.
Find where automation is worth investigating.
Estimate the manual load behind a repetitive process and check whether its inputs, outputs, approval points, and exceptions are clear enough for a scoped automation.
No email required. No invented ROI. The result is a directional starting point, not a delivery estimate.AUTOMATION CHECK
How suitable is this process for automation?
A quick estimate of how repetitive and structured the process is, how much manual work it currently takes, and where human review still matters.
TRY A REAL EXAMPLE
Not sure what counts as one repetition? Load a realistic example - every value stays editable.
Swipe to see more →
AI FIT ADVISORDescribe the process in your own words
Paste a short description of the recurring work and what you would most like to stop doing by hand. The AI structures it into the calculator below and suggests the closest match from Jovan's documented work - clearly separating what is delivered from what is only feasible.
HOW THIS COULD WORKAI suggested the starting values from your description. Check them below and change anything that does not match how the process actually works.Directional advice from verified portfolio context - not a quote, not a savings promise. The final assessment always happens in a conversation.
AssessmentAdd your process details to see a directional assessment.
- / 100Automation Fit
Current manual workload20 h / month
This estimates current workload, not guaranteed time saved.AUTOMATION FIT
What does this assessment mean?
What counts as one repetition?
One inquiry, email, order, invoice, document, booking, or another case that your team handles in a similar way.
What counts as a tool or system?
The applications or data sources involved in the process. For example: email + Excel + CRM = 3.
What does Automation Fit mean?
It measures how structurally suitable the process is for a controlled automation, based only on your answers - repetition, digital inputs, a clear result, review, consequence of mistakes, approval, and the number of tools. It is not an AI-generated probability and not an ROI estimate.
Does 22 h/month mean I will save 22 hours?
No. It estimates how much manual work the process currently contains. Actual savings depend on what is automated, exceptions, review requirements, and implementation quality.
WHAT THE TOOL CHECKS
Automation readiness is more than task volume.
A repetitive task can still be a poor automation candidate when the input is inconsistent, ownership is unclear, or exceptions cannot be reviewed.
Digital inputs
Whether the required information already exists in inboxes, forms, spreadsheets, CRM, ERP, or another accessible source.
Defined output
Whether a useful result can be described as a structured record, draft, score, alert, recommendation, or synchronized status.
Approval boundary
Where a person should review, correct, approve, or stop the workflow before a consequential action.
Exception handling
Whether missing data, duplicates, failed deliveries, and unusual cases can be surfaced instead of silently ignored.
System handoffs
How data needs to move between the tools the team already uses without creating another disconnected dashboard.
INTERPRET THE RESULT
What a strong fit usually looks like.
Promising workflow signals
- The work repeats with recognizable inputs
- The result can be reviewed before delivery
- Source data is digital and accessible
- Exceptions have a clear owner
- The first version can be limited to one bottleneck
Clarify before automating
- The process changes every time
- No one owns the final decision
- Important data exists only in people's memory
- Success cannot be described or checked
- The proposal depends on replacing every system at once
HOW I WOULD APPROACH IT
From workflow map to a reviewable first version.
- 01
Map
Identify the trigger, inputs, current decisions, system handoffs, and the actual bottleneck.
- 02
Constrain
Define what the system may prepare, what it must never do alone, and where human approval belongs.
- 03
Prototype
Build the smallest useful path with structured outputs, visible states, and realistic test cases.
- 04
Review
Evaluate edge cases, recovery behavior, data quality, and whether the workflow is genuinely easier to operate.
COMMON QUESTIONS
Before you automate a workflow.
Can this assess an AI agent idea?
Yes. It is useful for agentic workflows, research and scoring systems, CRM or ERP processes, inbox automation, and decision-support tools.
What if the process is not ready?
That is still useful information. The next step may be standardizing inputs, clarifying ownership, or reducing the scope before any automation is built.
Does every workflow need AI?
No. Some problems are better solved with a deterministic automation, a cleaner form, a database rule, or a simpler operating procedure.
NEXT STEP
Discuss the workflowHave a real workflow in mind?
Send the current process, tools, approximate volume, and approval boundary. I can help identify a sensible first scope.
