Most automation projects fail before they start because they begin in the wrong place. A new platform is chosen by the IT team, some example workflows are developed, and then the team is surprised that the process is not being adopted. Instead, the best approach is to start with the employees who experience issues regularly, and not with the individuals who are in charge of the technology infrastructure.
Countless tasks are repeated in every department. The finance team rewrites invoice information. HR is on the lookout for onboarding paperwork. After each meeting, sales manually updates the CRM entries. These are not issues related to IT. They are problems of the department, and they would be most effectively solved at the department level.
Audit Before You Automate
First, map out what is being done. Where are high-volume, rule-based, and time-consuming processes being executed? Invoice processing? Employee onboarding? Report generation? Data transfers to and from your CRM and ERP? Those are the processes you should go after first.
Your audit doesn’t have to be formal. Just ask your team leads to roughly account for where their team spends two or more hours on the same task each week. The answers will likely give you a list of what you are already suspecting.
Give Non-Technical Staff the Tools to Build
The low-code movement has brought a significant change to what can be done by business users. The solutions in this space empower heads of department to create and deploy workflows requiring no coding skills. The worldwide market for low-code development technologies is projected to reach 26.9 billion U.S. dollars in 2023 (Gartner), which indicates how fast organizations are adopting it.
It matters, this notion that some liken to citizen developers, because it takes IT out of the critical path. If a finance manager can construct her approval workflow, and she doesn’t need to raise a development ticket, it happens faster. IT can focus on governance and security, rather than lining up simple automations.
Microsoft Power Automate is a practical example of this in action, it interconnects Microsoft 365 applications with hundreds of third-party apps using out-of-the-box cloud connectors, and is explicitly built for people who understand their workflow but do not have coding knowledge.
Connect Your Legacy Systems Without a Full Migration
Many systems used within an organization do not have an API available. Pre-existing accounting software, legacy ERP systems, or on-premise databases are unable to communicate with new modern cloud software.
Robotic process automation can address this issue. RPA robots perform their functions at the presentation layer. They can read the screen, replicate copy-and-pasting data, and click on icons. Therefore, a process can be automated and a connection made between the legacy system and a new cloud system without changing either system. It’s not a long-term solution, but it’s a viable short-term stopgap for migration.
Centralize How Your Teams Get Notified
One source of lost productivity that often fails to register is alert fragmentation. We’ve all witnessed a scene where someone is monitoring their email, phone pinging with notifications, a dashboard refreshing every minute, and a specific inbox within an app open. It’s exhausting. Things get missed. Context gets lost.
The alerts for everything that’s important can’t always be contained in one system. But if you can automate key alerts to flow into one hub, typically Microsoft Teams or Slack, which many organizations already use, you’ve reduced the noise. When a sales order crosses a key threshold, a support ticket is escalated, or an overdue approval blocks progress, the right person gets a single, contextual notification in one place. They don’t need to check five tools to stay informed.
This sort of workflow orchestration doesn’t just make lives easier; it creates a natural audit trail. Every alert that is triggered is logged. This turns out to be a godsend when you need to review how a process ran.
Validate Data as it Moves Between Systems
Automation that moves bad data faster is worse than no automation at all. Data validation steps need to be built into any workflow that transfers information between systems.
That means building logic into the workflow itself: check that a field isn’t empty before passing it downstream, confirm that a number falls within an expected range, flag duplicate entries before they’re written to the destination system. These checks run silently in the background and catch the kinds of formatting errors and missing values that would otherwise take hours to trace back to their source.
Garbage in, garbage out applies directly here. A clean automated workflow is only as reliable as the data quality rules built into it.
Scale Without Scaling Headcount
One of the most compelling reasons to implement process automation isn’t the speed, it’s the fact that volume doesn’t matter anymore. A process that is managing 50 invoice approvals a week will manage 5,000 exactly the same way. There is no increase in staff time proportionally, no bottleneck at quarter end when volume increases.
This aspect of process automation means that it is a strategic decision, not just a tactical one around efficiency. Organizations that implement it at a departmental level (where people know how the process works and feel the pain) end up with automation that actually mirrors the real process.
So, start with one team, one broken process, and one working solution, and then add more.
You must be logged in to post a comment.