Machines excel at mundane, repetitive tasks that are easily forgotten. Humans should retain tasks that require judgment and emotional touch.
The most valuable automation occurs in steps that have three characteristics: repetitive, clear rules, and forgetting once disrupts the entire chain, such as deadline reminders, synchronizing assets when there is a new version, and classifying incoming requests. Conversely, creative approval or handling sensitive feedback should not be automated as these require human judgment. The right approach is to start with the most painful flow, keep it small, measure the results, and then scale up.
Automation does not mean automating everything. The practical question is: which steps can machines perform better than humans, and which steps lose something important if done by machines? Sinh Vũ answers directly: the most valuable steps are repetitive, governed by clear rules, and forgetting once disrupts the entire chain. These often include deadline reminders, synchronizing assets when there is a new version, and classifying and routing incoming requests.
Before establishing any automated connections, you should check that step against three questions:
This boundary is not technical; it is a boundary of value: machines handle repetitive tasks so that people have time to work on higher-value activities.
Standardizing tools and automating repetitive tasks to reduce bottlenecks is part of the design operations (DesignOps) infrastructure, not an optional add-on after the studio has grown.
Nielsen Norman Group, DesignOps Study Guide
In service O2, Sinh Vũ builds automated flows based on these steps: reminder for approval, asset synchronization, new version notifications, using tools like Slack bot, Zapier, or APIs depending on the client's infrastructure. The principle Sinh Vũ maintains is to start small with the most painful flow first, measure if there are real savings, and then expand. Do not create a complex system just to appear professional.
The actual result does not come from the number of automated connections. The result comes from choosing the right spots: which steps are most often forgotten, and which steps, when blocked, affect the most people. Automate in those areas, leaving the rest for decision-makers.
Topic: Automating the iterative process: where it’s worth the investment. Sinh Vũ guide, sinhvu.com
Select each item you find appropriate, then print or save as PDF to take with you.
If you have marked most of the signs above, this is the time to discuss in more detail. Sinh Vũ can help you review and propose a direction.
Nielsen Norman Group, DesignOps Study Guide. Sinh Vũ, O2 service profile (internal).
Each connection between two tools is a long-term maintenance task, not just a one-time setup. The more tools are connected, the harder it becomes to understand when something goes wrong. Sinh Vũ recommends starting with a single connection at the most painful point, stabilizing it before adding more. A simple system that works is still better than a complex system that no one understands how to fix.
You should not. Automating a process that is not yet standardized locks in mistakes, and fixing them later is more costly than starting over. The correct sequence is: identify the actual flow, standardize the steps first, and only then let the machine run. If you are unclear about what issues exist in the current process, Sinh Vũ can help review it before creating any automated tools.