Most businesses don’t struggle to find things to automate. They struggle to choose. Once a team starts looking, the list grows quickly: invoice entry, lead follow-ups, onboarding paperwork, stock alerts, weekly reports. Everything looks like a candidate.
The first automation you pick matters more than the ones after it. A good first project saves visible time within weeks and builds trust in the idea. A bad one, such as a messy, exception-heavy process that nobody fully understands, drags on for months and convinces the team that automation “doesn’t work here”.
This guide gives you a simple, repeatable way to decide what to automate first, using information you already have.
Why “Automate Everything” Fails
Automation multiplies whatever process it runs. If the process is clear and consistent, automation makes it faster and more reliable. If the process is unclear, automation makes the confusion faster.
Three patterns sink early automation projects:
- Starting with the most painful process: The process everyone complains about is often painful because it is full of exceptions, judgment calls and missing data. Those are the hardest things to automate.
- Starting with the most visible process: Leadership picks something high-profile, but it touches five departments and three systems, so the first project becomes a year-long integration.
- Automating a broken process: If people disagree on how a process should work, software can’t settle the argument. It will copy one version of the confusion.
The goal of your first project isn’t to fix the biggest problem. It’s to prove value quickly and safely, then use that momentum for bigger ones.
The Six Signals of a Good First Candidate
A process is ready to be automated when most of these are true:
- High Volume or Frequency: It happens many times a day or week. Automating a task done twice a year rarely pays back.
- Rule-Based: The steps can be written as “if this, then that”. People follow the same logic every time, without needing experience or judgment.
- Stable: The process hasn’t changed much in the last year and isn’t about to be redesigned.
- Digital Inputs: The information arrives in a structured, digital form, such as emails, forms, spreadsheets, or exports from existing software, not handwritten notes or phone calls.
- Costly Mistakes: Manual errors cause real problems, such as payment delays, wrong orders, compliance issues or unhappy customers.
- Measurable: You can count how often it happens and roughly how long it takes today, so you can prove the improvement later.
Tip: If you can’t explain a process to a new employee in under ten minutes, it probably isn’t ready to automate. Simplify it first.
Step 1: Build a Process Inventory
Before you can compare processes, you need them in one place. Ask each team lead to list the repetitive tasks their team does every week. Keep it simple, a shared spreadsheet is enough.
For each process, capture:
| Field | What to Record |
|---|---|
| Process name | A short, clear label, e.g. “Supplier invoice entry” |
| Owner | The one person accountable for how it works |
| Frequency | How many times per week or month it happens |
| Time per run | Average minutes a person spends on it each time |
| Systems involved | Email, ERP, CRM, spreadsheets, accounting software |
| Known issues | Common errors, delays or complaints |
Aim for 15–30 processes. You don’t need perfect numbers. Rough estimates from the people who do the work are good enough to rank them.
Step 2: Score Each Process
Next, score every process from 1 (low) to 5 (high) on each factor below. Some factors matter more than others, so each gets a weight.
| Factor | Weight | Score 5 means… | Score 1 means… |
|---|---|---|---|
| Volume | 3 | Happens dozens of times a day | Happens a few times a year |
| Rule clarity | 3 | Every step follows fixed rules | Every case needs human judgment |
| Error cost | 2 | Mistakes are expensive or risky | Mistakes are harmless |
| Stability | 2 | Unchanged for a year or more | Changes every few weeks |
| Data readiness | 2 | Inputs are digital and structured | Inputs are paper, calls or free text |
| Ease of integration | 1 | One or two systems with good access | Many systems, some with no access |
Multiply each score by its weight and add them up. The maximum is 65.
- 50 and above: Strong first candidate.
- 35–49: Good second-wave candidate, or a first candidate once a blocker is fixed.
- Below 35: Leave it for now. Often the right move is to simplify the process instead.
Here is how the scoring might look for a mid-sized distribution business. These are illustrative scores, not benchmarks:
| Process | Volume | Rules | Error Cost | Stability | Data | Integration | Total |
|---|---|---|---|---|---|---|---|
| Supplier invoice entry | 5 | 5 | 4 | 5 | 4 | 4 | 60 |
| Order status emails | 5 | 5 | 3 | 5 | 5 | 3 | 59 |
| New employee onboarding | 2 | 4 | 3 | 4 | 3 | 3 | 41 |
| Customer complaint handling | 4 | 2 | 4 | 3 | 2 | 3 | 39 |
| Annual budget planning | 1 | 1 | 5 | 2 | 2 | 2 | 26 |
Complaint handling feels urgent, but it depends on judgment and messy inputs. Invoice entry and order status emails are less dramatic, and far better first projects.
Step 3: Estimate the Time You’ll Get Back
Scores tell you which processes are suitable. A quick time estimate tells you which ones are worth it.
Use this simple formula:
Hours saved per year = runs per week × minutes per run × 52 ÷ 60
For example, if your team enters 200 supplier invoices a week and each one takes 6 minutes:
200 × 6 × 52 ÷ 60 = 1,040 hours a year, roughly half of one full-time person’s working year.
Multiply the hours by an average hourly cost to get a rough yearly value, then compare it with the cost of building and running the automation. Also note the benefits that don’t show up in hours: fewer errors, faster payments to suppliers, and staff moved to more useful work.
Tip: Don’t promise that automation will replace a role. Talk about giving time back. The people who run the process today are your best source of rules and edge cases, and they’ll only share them if they see automation as help, not a threat.
Step 4: Plot Impact Against Effort
With scores and time estimates in hand, place each process on a simple two-by-two grid. Impact is the time and error reduction. Effort is how hard it is to build, based mostly on integration and data readiness.
quadrantChart
title Automation Priority Matrix
x-axis Low Effort --> High Effort
y-axis Low Impact --> High Impact
quadrant-1 Plan as a Project
quadrant-2 Automate First
quadrant-3 Nice to Have
quadrant-4 Avoid for Now
Invoice entry: [0.25, 0.85]
Order status emails: [0.2, 0.7]
Employee onboarding: [0.45, 0.45]
Complaint handling: [0.75, 0.65]
Budget planning: [0.85, 0.2]- Automate First (high impact, low effort): Your first one or two projects come from here.
- Plan as a Project (high impact, high effort): Worth doing, but with proper scoping, a budget and a longer timeline.
- Nice to Have (low impact, low effort): Good filler work once the main projects are running.
- Avoid for Now (low impact, high effort): Revisit only if something changes.
Red Flags: Processes Not to Automate First
Even a high-scoring process should wait if any of these apply:
- No clear owner: If nobody is accountable for the process, nobody can approve the rules or handle exceptions.
- A redesign is coming: If you’re about to change your ERP, pricing model or approval structure, automate after the change, not before.
- Frequent exceptions: If more than roughly one case in five needs special handling, fix the process or split out the exceptions first.
- Regulatory grey areas: Processes with legal or compliance judgment calls need human review built in, which makes them poor first projects.
- No baseline: If you can’t measure how the process performs today, you won’t be able to show the improvement.
Common First Candidates by Department
If you’re not sure where to start your inventory, these processes often score well:
| Department | Typical First Candidates |
|---|---|
| Finance | Invoice data entry, payment reminders, bank reconciliation, expense report checks |
| Sales | Lead capture and routing, CRM updates after calls, quote generation from templates |
| Operations | Order confirmation and status updates, low-stock alerts, delivery scheduling |
| HR | Onboarding document collection, leave approvals, policy acknowledgements |
| Customer Support | Ticket tagging and routing, standard reply suggestions, satisfaction surveys |
| Management | Weekly and monthly reports pulled from several systems |
Step 5: Run a Focused Pilot
Once you’ve picked your first process, keep the pilot small and measurable:
graph LR
A["Measure the baseline"] --> B["Simplify the process"]
B --> C["Automate the core path"]
C --> D["Route exceptions to people"]
D --> E["Measure for 30–60 days"]
E --> F["Expand or adjust"]- Measure the Baseline: Record current volume, time per run and error rate before you change anything.
- Simplify First: Remove unnecessary approvals and steps. Agree on one version of the process.
- Automate the Core Path: Handle the common case end to end. Don’t try to cover every exception in version one.
- Route Exceptions to People: Anything unusual goes to a named person with all the context they need.
- Measure for 30–60 Days: Compare against the baseline. Share the results openly with the team.
- Expand or Adjust: Use what you learned to pick and scope the next process from your list.
Conclusion & Next Steps
The best first automation is rarely the most exciting one. It’s a frequent, rule-based, stable process with digital inputs, where mistakes cost money and results are easy to measure. List your processes, score them honestly, check the time you’d get back, and start with the one in the top-left corner of your matrix.
Get that first win, measure it, and let it fund and justify the next one.
If you’d like help building your process inventory or scoring your candidates, our team works with businesses on business process automation from first assessment to rollout.
