ALEXSOFT
ALEXSOFT / INSIGHTS 003 Workflow automation · 2026.08.08

ALEXSOFT INSIGHTS

Does workflow automation always require new software?

Keep the tools that work and reduce repetitive transfers first. Examples from weekly reporting and Google Sheets show how connecting existing tools can automate work.

By · · Workflow automationPublished Updated Original on Naver Blog (Korean) ↗
Does workflow automation always require new software? (Original illustration in Korean.)

When people hear “workflow automation,” many first think of a new application. They assume that a software development company, in particular, would recommend one right away.

“Should we build a business system to replace Excel?”

“Do we need a separate application so several people can work with the same information?”

“Do we have to develop everything from scratch to use AI?”

Sometimes, yes.

But when automating work, I think what we can leave in place matters more than what we can build.

Building a new application changes more than people expect. Existing data must be migrated, users must learn new screens and workflows, permissions need consideration, and established business processes often change as well.

Above all, a business system is not something you pay to develop once and then forget.

As business rules and the organization change, the system needs ongoing changes. Integration changes require a response. Without the effort of incident handling, security updates, backups, data corrections, and other maintenance, it quickly becomes a burden.

A dedicated business system can be a powerful solution when it is needed. But once you include years of maintenance costs and operational responsibility alongside the initial cost, it is also one of the most expensive options.

Why I did not build a new reporting system

One organization I worked with prepared weekly activity reports. There was more than one reporting template.

Writing the reports was cumbersome, but another part took even more time.

People had to check schedules, review changes in the task management tool, and find decisions made in meetings. Then they gathered this scattered information, organized it into this week’s results and next week’s plans, and transferred it into the required templates.

“Would it be better to build a dedicated weekly reporting system?” That would have been a reasonable thought.

But a new reporting system would not make the existing templates disappear. The company’s formal reporting and approval process would remain. When the organization or approval chain changed, the same information would likely end up being maintained twice: in the new system and in the reporting templates.

So I did not build a new reporting system.

We kept the existing templates as the official record, connected AI to the information scattered across schedules, task management tools, and meeting notes, and had it read and organize that information. We then automated transferring the organized content into the correct places in the existing templates.

The point was not to replace the reporting system.

It was to reduce the searching, organizing of content and formatting, and copying and pasting that people repeated every week.

Solve it within the tools the team already uses.

I had a similar experience with issue tracking.

Instead of creating a new issue tracking application or a separate dashboard, we added statistics, filters, and guidance to the Google Sheet the team already opened every day.

When information a team needs to check is scattered across separate files and screens, people first have to work out where the latest information lives.

Every additional tool brings more costs for learning, permissions, sharing, and keeping information current.

So if a problem can be solved inside a tool the team already uses, that is where I look first.

The same applies to internal wikis, task management tools, existing systems, development materials, and shared spreadsheets. Often, it is better to leave information where it belongs and connect those places than to move everything into a new system.

Keep, connect, and build only what is needed

I think about workflow automation in three stages.

  1. First, see whether the current tools can stay as they are. If they work well and people know how to use them, there is no reason to change them.
  2. Next, look for places where people repeatedly transfer, check, and copy information. Connecting existing tools or adding a small automation can be effective here.
  3. Finally, decide whether a dedicated system is needed.

A new system is worth considering when the same information must be entered in several places, people often cannot tell which information is current, permissions and change history need strict control, or exceptions become too complex for the existing tools.

And if the work directly supports the company’s core competitive advantage, a dedicated system is something to invest in properly from the start, rather than a last resort.

This does not mean delaying a necessary system because maintenance costs money. If a dedicated system reduces data errors and repetitive work and makes core operations reliable, it can save far more over the long term.

The important question is not whether we can build a new application.

It is how much system support is needed to solve the current problem.

I do not see good workflow automation as removing every existing tool.

Keep what works, identify exactly where people repeat unnecessary tasks, and reduce those tasks. That is automation and process improvement that can be faster, safer, and useful in practice for longer.

That is why, when I hear about a business process, I do not immediately propose developing and introducing a new application. I first look carefully at the spreadsheets, shared documents, existing application screens, and workflows already in use.

A better starting point is to work out what to keep and where to connect it, before deciding what to build.

Does workflow automation always require new software? (Original illustration in Korean.)