When was the last time you suspended all notifications on your phone? Maybe in airplane mode or at night (and even then, only if your alarm can still work), right?

Of course, you don’t want to miss your boss’ urgent email, an alert about a family emergency in the middle of the night, or that call from the school when your child is sick. Text during dinner? Sorry kid, hold that thought. WhatsApp alert during a date? Before your tell me about your hobbies, I just need to respond to this forward from my 30 best friends (I mean, they saw I viewed the message — thanks, WhatsApp). Don’t you entrepreneurs love it when the investors pull out their devices mid-pitch? Sounds like it could be a scene from a Shaun of the Dead sequel, doesn’t it?

The mindless action of reaching for our phones in the middle of anything makes us living zombies in today’s world.

For the strong few of you who purposefully resist zombification, you have to use your finite store of self-control constantly, all day (don’t check your phone, don’t check your phone, don’t check your phone…).

Well, there are quite a few companies trying to tackle this problem. However, no one has come up with a comprehensive solution yet.

Rules are not the answer

Several companies have tried to use rules to take certain actions for you. EgoMotion, Atooma, and IFTTT all have rules you can set to automate certain actions. For example, if your device’s GPS indicates that it is moving at more than 30 mph, then send all calls to voice mail and read all texts aloud. Egomotion’s Agent is a bit more context-based but still assumes binary behavior. The problem is that we don’t live such binary lives. Humans are confusing and unpredictable and so can’t be confined by rules. For example, sometimes when my phone is moving at 35 mph I’m the passenger or taking public transportation. Sometimes I’m the driver but still need to see who is calling even if I won’t answer. Setting up more and more sub-rules becomes a mess.

Willpower is not the answer

For the willpower exercisers, Moment and RescueTime tell you how often you use your device throughout the day and on what apps. This requires additional thought, time, and training on your end to stay disciplined. How often have you said, “I uninstalled Facebook from my phone because I was spending too much time on it.” Then it’s back on your phone in a week or, for someone like me, the next day. I am good at convincing myself Facebook is work-related and generally good for my soul.

Contextual awareness alone is not the answer

There is a category of apps that are trying to be contextually aware (it was definitely a hot category last year): e.g., Google Now, Easily.do, Everything.me. These apps look at your email, calendar, etc. to get you more context, but they just create more alerts. A step in the right direction but far from the solution. This category is designed around giving you the right information at the right time, which just means more data presented to you. It’s meant to be the “right” information at the right time, but it doesn’t stop other information from distracting you. So Google Now alerts you that your flight is delayed, then you get a text from your airline with the same information, and then TripIt sends another alert. These apps are not talking to each other or to the OS, and the disconnect creates more alerts and distractions for you.

OS-level mindfulness can be the answer

One answer could be consolidation. Snowball is almost headed in the right direction. Snowball consolidates all your alerts in one place. It takes alerts from various apps and allows you to respond from within the “notification-dashboard.” A step in the right direction, but right now it’s just a consolidation of alerts in one place. For it to be effective, it has to be combined with the context or what I like to call “mindfulness” so it’s not just all the alerts in one place but also a reorganization of alerts based on priority. Literal translation of mindfulness is simply awareness.

While apps like Snowball are headed in the right direction, adding OS-level mindfulness is the answer. Your device needs to be mindful of the actions it’s asking you to take at all times. The device should be an enabler to a better, productive life instead of making you a zombie that loses focus to the device at odd hours of the day.

What would the OS-level mindfulness consist of? The following three layers:

1. Interception — Something that intercepts these alerts before you see and hear them. Currently, the only OS-level interception is if your alerts should be silent, vibrate your device, or make a custom sound. New versions of the OS should have the ability to intercept your alerts based on layers 2 and 3 below.

2. Personalization — Just like the Internet, your phone should know you. You use it 150 times a day, but still, for the most part, it does the same basic things in the same way that it does for everyone. App launchers such as Aviate and Cover were headed in that direction but again not exactly streamlining what gets delivered, with context, at the right time. In fact, they meant to do the opposite: bring as many things as possible in front of you as they happen. Instead, to combat alert overload, I want my device to control everything that gets delivered and personalize what gets delivered and when.

The personalization layer needs to understand what apps you use over time, what kinds of alerts you usually respond to (e.g., from which apps, and with what level of detail), when you respond to them (e.g., immediately after seeing them, or based on some order), and how you respond to them (e.g., whether you take a particular action, dismiss the alert, or flag it for later). For example, personalized alerts management should understand that I usually can’t respond to emails during meetings on my calendar, but I can during specific types of meetings (such as large groups with more than 20 people, or during large conference calls). And it should know that I will always respond to certain people, as well as a changing set of people — people I have chatted with on all the chat platforms today or that are in my next meeting, for example.

Google Now or other contextual apps don’t do this because they don’t incorporate your behavior on the device. So it’s not just personalization from your cloud data but also integrated with your phone book, call log, SMS, chat apps, etc.

This layer can eventually combine data from your cloud services such as Gmail, Calendar, to-dos, etc. with your behavior from your device.

3. Prioritization Engine — Lastly, the OS-mindful layer should account for when and who. The personalization engine is more about the why. The prioritization engine allows for human intervention that the OS can leverage to for training over time. This is not rule-based, but activity-based. Maybe humans can set how they want to be alerted during times of emergency versus not-so-important alerts that can wait until your meeting is over versus just a regular no-action-required alert.

The question of when to deliver alerts is the area where I’ve seen the least innovation. The default behavior is always to push notifications as soon as they are available. But in reality, I only want alerts delivered to me from apps that are relevant at that time of the day. For example, on an Android device, I never want to hear a loud alert and see a notification in middle of my day that says Google has backed up my photos. I don’t know how that alert is helpful to me and why it has to distract me from my regular in-person conversations.

For who, my device should only deliver alerts from my most frequent or least frequent contacts because it’s unusual behavior on their part to call you midday. Based on your behavior, the apps and notifications get reprioritized so you don’t see 15 random alerts after a meeting, rather just two important ones —  one maybe about your next meeting and the other a text from your boss. Again, the rule based innovations —  e.g. you can whitelist certain people whose rings you’ll hear even when Do Not Disturb mode is on, etc. — don’t work here because this is a dynamic list based on your day.

There are a lot of layers to this, but the idea is that the mindful layer processes this information before it jumps up and demands your attention. Once this layer is in place, all apps can go through this layer to deliver your most desired information, at the right time, in a proper fashion, instead of constantly causing us to jump to check our phones every time they blink or vibrate.

Mindful mobile can actually work on making sure you give attention to what is deserved at the moment  —  without the worry of making you a 21st-century zombie.

Shruti Gandhi  is an investor, entrepreneur, and engineer. She was previously with True Ventures, Samsung’s Early Stage Investment Fund and three other funds where four of her companies had successful exits in less than three years. She invests in passionate entrepreneurs that are working hard to solve problems in today’s world of work. She is a co-chair of the TiE San Francisco chapter and on the Marketing & Technology Committee of the San Francisco Ballet. Follow her on Twitter @atShruti.