Most of the time your team loses in dashboards and admin panels does not come from one big task. It comes from routine work split into many small actions: log in, find the section, open a form, fill field after field, switch to another app for a number, come back, save. At HBS we have counted 14 manual steps to publish one article by hand, and most teams repeat tasks like that every day. Research puts a size on it: in a 2022 Harvard Business Review study, workers toggled between apps about 1,200 times a day and spent just under four hours a week getting back on track. Below is what the research says, a simple way to measure the clicks and minutes behind one task in your team, and the fixes that work.
What is the 14-click problem?
We use the name for any routine task that is simple to describe but slow to do. Publishing an article is a good example because everyone understands the result: a new post, live on the site, with an image and the right search fields. These are the separate actions it takes by hand, as we list them on our own product page:
- Log in
- Open articles
- New article
- Title
- Write content
- Upload image
- Image description
- Choose category
- Search title
- Search description
- Page link
- Publish date
- Preview
- Publish
None of these steps is hard. The problem is that they add up, they repeat, and each one asks for a little attention. The same pattern appears across the business: changing a price in the online store, pulling last week's figures from the analytics dashboard, updating a client's status in the CRM, or checking stock in the ERP and preparing a purchase order. Each task looks too small to fix, so nobody fixes it, and together they fill the week.
How much time do teams lose switching between apps and tabs?
A useful measurement comes from a 2022 Harvard Business Review study of 20 teams, 137 users, across three Fortune 500 companies, observed for up to five weeks. Workers toggled roughly 1,200 times each day, which added up to just under four hours each week spent reorienting themselves after toggling, roughly 9% of their time at work.
The number of tools behind those switches keeps growing. A Gartner survey of 4,861 full-time employees in the US, the UK, India and China, run from September to November 2022, found that the average desk worker uses 11 applications, compared with six in 2019. 40% use more than the average, and 5% use 26 or more.
These are global studies, mostly of large companies, and we could not find a comparable published survey for Egypt or the Gulf. But the setup is familiar to any regional team that runs a website, a CRM, an ERP, a dashboard and a chat app side by side, often in two languages.
What is "work about work" and how much of the day does it take?
"Work about work" is time spent coordinating work instead of doing the skilled job someone was hired for: chasing updates, looking for files, moving information from one place to another. Asana's Anatomy of Work Global Index 2023, the latest edition of the study, surveyed more than 9,615 knowledge workers in the United States, the UK, Australia, France, Germany and Japan. Work about work still took the majority of their day: 58%.
The same study links the number of apps to things falling through the cracks. Knowledge workers used 8.8 apps a day on average, and 10 at director level and above. Among people using 6 to 15 apps, 15% reported missing messages and actions, rising to 25% for those using 16 apps or more.
How much time goes into searching for information?
Dashboards are supposed to make information easy to reach, yet finding the right number still takes work. The same Gartner survey found that 47% of digital workers struggle to find the information or data they need to do their jobs well. An older but widely cited McKinsey Global Institute report from 2012 estimated that the average interaction worker spends 28 percent of the workweek managing email and nearly 20 percent looking for internal information or tracking down colleagues who can help.
Inside a dashboard, searching looks like this: open the analytics tool, pick the date range, add a filter, compare with last month, export, paste into a slide, then ask a colleague why the number looks odd. The data was there all along. The time went into getting to it. If this is mostly a management problem in your company, our guide on moving from monthly reports to instant answers covers that side.
Why do small interruptions cost more than they look?
Each switch between screens is a small interruption, and interruptions cost more than they seem. The American Psychological Association quotes the psychologist David Meyer, who has said that even brief mental blocks created by shifting between tasks can cost as much as 40 percent of someone's productive time. That is an upper estimate, not an average, but it explains why a day of small switches feels tiring even when no single task was difficult.
There is a quality cost too. Every number copied from one system into another, and every field filled from memory, is a chance for a typo, a wrong category or an outdated figure. The more steps a task has, the more places an error can enter.
How do you measure the clicks and minutes behind one routine task?
You do not need software for this. You need one task, one colleague who does it often, a stopwatch and an honest count.
- Pick one routine task that happens often: a weekly report, publishing an article, adding a product or preparing a purchase order.
- Write down the trigger and the finished result, so everyone agrees where the task starts and ends.
- Watch it done the normal way and count every click, field, screen and app change, without suggesting shortcuts.
- Time three separate runs and keep the middle value.
- Count how often it happens across the team in a normal week, using the calendar, the task list or the system's own history.
- Note the waiting time: approvals, missing images, or a developer who has to make the change.
- Multiply the minutes per run by the runs per week to get the weekly cost of one task.
| What to record | How to record it | Example: publishing one article |
|---|---|---|
| Task name | One short line | Publish one article in English and Arabic |
| Trigger | What starts the task | The final draft is approved |
| Finished result | What "done" looks like | Live on the site with an image and search fields |
| Clicks, fields and steps | Count while watching | Our own count for this task is 14 steps |
| Screens and apps used | List each one | Admin panel, image folder, draft document |
| Minutes per run | Middle value of three timed runs | From your stopwatch |
| Runs per week | Count across the whole team | From your publishing calendar |
| Waiting time | Where the task stops | Waiting for an image or a translation |
| Error points | Where mistakes usually happen | Wrong category, missing image description |
| Owner | Who usually does it | Content editor |
Repeat the exercise for your five most common tasks and you will have an honest list of where the week goes. The same baseline is what you need later to judge any change, as we explain in our guide to measuring AI return on investment.
How can you cut the clicks before buying anything new?
Start with the fixes that cost little. Many teams remove much of the friction without a new system.
- Use fewer tools. List every app the team opens in a week and retire the ones that duplicate each other. Two client lists mean twice the searching and twice the errors.
- Save views and shortcuts. Saved filters, default dashboards and bookmarks to the exact admin page cut the first steps of many tasks.
- Use templates. Article templates, product templates and prefilled fields leave fewer empty fields to fill.
- Batch similar work. Updating ten products in one sitting costs less attention than updating one product at a time between other tasks.
- Automate fixed rules. When a task always follows the same rule, such as creating a CRM record when a contact form is submitted, a simple automation removes it.
- Allow plain-language requests for tasks that change every time, where fixed rules break. We compare the options in AI agents vs chatbots vs automation.
| Fix | What it removes | Limit |
|---|---|---|
| Fewer tools | App switching and duplicate data | Needs agreement across departments |
| Saved views and shortcuts | Searching for the right screen | Each person has to set them up |
| Templates | Repeated typing and empty fields | Someone still clicks through every step |
| Rule-based automation | The whole task, when the rule never changes | Breaks when requests vary |
| Plain-language requests | Hunting through screens and fields | Needs clear permissions and approvals |
How does Amrak handle the 14-click problem?
Amrak (أمرك, "your command") is an AI assistant from HBS Group that works inside the systems a company already has: the website and its content, dashboards and analytics, the CRM, the ERP, the calendar, the client portal and design tools. Staff type a request in plain Arabic or English and Amrak carries it out. It is built on MCP, an open standard for connecting AI assistants to company systems (see what MCP is), and HBS Tech builds a custom connector for each client's own sections and rules.
Applied to the 14 steps above, the whole list becomes one sentence: someone on your team writes the request, then checks the result. The same goes for other tasks that usually mean clicking through screens:
- "Publish a post on a set date with an image" instead of filling every field by hand.
- "Explain the traffic drop and show the top-performing pages" instead of building the view yourself.
- "Summarise yesterday's work" instead of opening each tool in turn.
- "Flag low stock and draft a purchase order" instead of moving between the ERP and a spreadsheet.
Every request follows five steps: Amrak understands it, checks the person's permissions, prepares the full change, asks for approval if the action is sensitive, then completes the task and logs it. We walk through an example in how Amrak works. Messages to clients, deleting anything and changing prices always wait for approval, each person reaches only what their permissions allow, and every action is recorded with who asked, when and what changed, as covered in our guide to safe AI actions. There is no data migration or system replacement, and Amrak can be switched off at any time.
Amrak does not remove the need to check the work. It removes the hunting. For ideas, see our ready-to-use commands by department and our case study on how HBS runs its own work on Amrak.
If your team spends its days clicking through dashboards and admin panels, start by measuring one task with the table above. Then book a free discovery session on the Amrak page, and our team will help you map which routine tasks can become one sentence.
