Almost every business I work with is spending significant staff time on work that could be automated — and most of them have partially accepted that as normal. Data entry between systems, status update emails, report generation, file organization, approval routing, follow-up sequences — these are tasks that feel necessary until you examine them closely and realize a computer can do them more reliably and at a fraction of the cost.

How to Find the Automation Opportunities

The first step is a workflow audit — not the formal kind on an org chart, but a practical inventory of what people actually do with their time. I look for patterns: tasks that happen on a schedule (daily, weekly, monthly); tasks that are triggered by an event (someone fills out a form, an order comes in, a file is uploaded); tasks that involve moving data from one place to another; and tasks where the decision logic is consistent and rule-based rather than genuinely requiring human judgment.

That last category is important. Automation is appropriate for tasks where the right action can be specified in rules. It’s not the right answer for decisions that require context, judgment, or relationship management. Knowing the difference saves you from building automation that creates more problems than it solves.

The Most Common Manual Process Patterns

Across the businesses I work with, a few patterns come up repeatedly:

  • Data entry between systems: Someone exports from System A and imports to System B. Or they’re retyping the same information into two places. This is almost always replaceable with an integration.
  • Status update communications: Manually sending emails to clients or partners at each stage of a process. Email sequences triggered by status changes in your CRM or project management tool replace this entirely.
  • Report generation: Pulling data from multiple sources into a spreadsheet and formatting it for a stakeholder. Automated reporting that runs on schedule and delivers the output where it’s needed — Slack, email, a dashboard — eliminates the manual step.
  • Document routing and approvals: Emailing documents for signature or approval and tracking the thread manually. An automated workflow with clear stages, notifications, and a status view replaces the email coordination.
  • Follow-up sequences: Manual reminders and follow-ups that depend on someone remembering to send them. Rules-based sequences that trigger on time or event remove the dependency on human memory.
  • Data validation and cleanup: Manually checking incoming data for completeness or errors. Validation rules at the point of entry catch problems before they enter the system.

Tools vs. Custom Code

Not every automation project needs custom development. Many of the patterns above are solvable with well-configured existing tools — a CRM with proper automation rules, a Zapier workflow, a scheduled report in an existing analytics platform. I’m direct about this: if a $50/month tool solves the problem, that’s the recommendation.

Custom automation code makes sense when: the logic is complex enough that a no-code tool produces a fragile, hard-to-maintain configuration; the data involved requires transformation that generic tools can’t handle; reliability and error handling matter (no-code tools often fail silently); or the automation needs to integrate with systems that don’t have supported connectors.

This is the same calculation that applies across integration development — generic tools have a real use case, and custom code is the escalation when they’re not enough.

Measuring the Return

Automation ROI is usually straightforward to calculate: how many hours per week does the manual process consume, what’s the fully-loaded cost of that time, and how does that compare to the cost of automating it? A process that takes 10 hours a week at $40/hour in fully-loaded labor cost is $400/week, or roughly $20,000/year. Automation that costs $5,000 to build and $100/month to run pays back in well under a year and keeps paying after that.

The less visible cost is errors. Manual processes have error rates. Those errors have costs — rework, customer service issues, bad data in reports that leads to bad decisions. Automation applied to rule-based tasks typically improves accuracy significantly over manual execution.

Where to Start

The right starting point is identifying the highest-cost manual process in your operation — the one where people spend the most time, or where errors are most expensive, or where the dependency on a specific person is the biggest risk. Solve that first, validate the approach, then move to the next one.

I’ve applied this discipline across software projects through Salterra — the businesses that get the most value from automation are the ones who approach it as a systematic improvement to their operations rather than a one-time technical project.

If you want to map out where automation could have the most impact in your business, the software development page covers how I structure that conversation and what a realistic engagement looks like.