What tasks should be automated and how to do it
Miscellaneous / / February 16, 2022
From calendars to full-fledged studies.
The more actively artificial intelligence develops, the more opportunities people have. For example, they can automate a variety of tasks. Which ones to start with, say developers Matthew Mottolla and Matthew Cotney.
In the book “Remote Economics. How cloud technologies and artificial intelligence are changing work” the authors not only analyze influence of AI, but also advise companies on how to better organize the activities of employees in new conditions. With the permission of the Alpina PRO publishing house, Lifehacker publishes an excerpt from the 13th chapter.
1. Calendar management
If there is one activity that almost all entrepreneurs, freelancers, team members, and managers dread, it's meeting scheduling. Internal, external, formal, informal conversations over a cup of coffee. Today's emphasis is on "productivity" (read, maximum workload), and corporate staff spend a lot of part of their day in meetings, so arranging the meeting time of two people, even on the same team, becomes difficult task. Throw in a few more people, time zones, and loose freelance associations, and you can easily spend a few hours a week just juggling your schedule. It's as fun as a sharp stone in a shoe.
Welcome to the world of automation! Turns out, planning appointments are ideal for computer delegation because it is repetitive, limited, and learnable. There are several approaches and types of instruments in the current market.
These are simple productivity tools with a bit of artificial intelligence built into them. Calendly, Doodle and similar products access your calendar and then provide a link and a website so that other people can choose the time that works for them. The app then downloads their responses and offers several different options based on the best accessibility for all participants.
Sometimes the simplest solutions are the best. You don't always have to focus on artificial intelligence.
However, these tools place the burden of choice on those in charge and simply shift the work from you to them. For meetings with higher-level managers or VIPs, you may need a personal touch. Use a tool with a short name X.ai.
This smart tool also syncs with your calendar, but instead of forcing people to choose time on the website, you simply add the X.ai agent to the meeting request email address. Let’s say you receive an email from a sales manager, a potential partner − freelancer or even an old school friend. You just copy X.ai and it sends an appointment request email to those mentioned in the email thread. It works just like the other tools but looks much more natural.
So what do I recommend? First, stop and ask yourself the question: “Do I really need this meeting?” significant growth productivity can be achieved simply by realizing the fact that you do not need to attend every meeting, you do not need to assign new meetings and many cases can be fully managed using email, instant messengers or voice messages. Leave meetings only to discuss the most important decisions or to develop deeper relationships with your network of contacts.
2. Responses to letters
Next on our list of productivity killers is email. As with meetings, think about an alternative first. Email is the worst way to communicate quickly for most projects. Think about the cognitive effort you have to put in every time an email arrives. It is important? This is urgent? This is for me? Do I need to answer? What kind of client needs it? In what context? Are they crazy there?
You can dig into e-mail hours, you just have to lose control, and at this time the real value of your work disappears. It's like a vacuum sucking your soul (wow, a deep metaphor turned out).
Try to translate the operational communication of your team or organization to Microsoft Teams, Slack, Trello or similar application for operational business communication. Set up channels for active projects and groups and exchange emails there. I know it sounds like a pipe dream, but I worked at a startup where we almost exclusively used Slack. I received about 20 emails a day, mostly from clients. This. It was. Awesome!
If your workplace still needs email, don't despair. There is still hope.
As before, let's start with simple solutions. Microsoft Outlook, Google, Apple Mail and other email tools allow you to create rules, which flag messages, move them to a folder or delete them completely, and solve many other tasks. When it comes to simple workflows, this approach can work wonders. For example, you can set up a rule to flag emails that require a specific response (for example, about a service issue). clients) based on a pre-made standard template.
If we talk about more complex email management, you can use all sorts of intelligent automation tools of varying complexity. Entrepreneur Ajay Goal launched a product called wordzen, which brings together email users and email editors.
Early on in the development of the Internet, he correctly guessed two trends: email would become a powerful tool personal and professional productivity, and Gmail will become the de facto standard for small businesses and private persons. Ajay has created tools that can be easily built into Gmail so that non-technical users can automate tasks like emailing themselves. marketing suggestions and help with writing letters.
AI has also begun to infiltrate directly into Gmail. There's an auto-suggestion feature that automatically suggests the next few words based on what you're typing. This is a powerful and fairly simple use case. AI. There are literally billions of emails going through Google that can teach AI how to use common expressions. And he continues to improve his skill, reaching such a degree of perfection that even I am impressed. This feature reduced email response times by about 10-20% - an instant productivity boost.
Dictation is also an underrated but handy feature for replying to emails and text messages.
I recently bought a pair of Air Pods (three years after they came out - I'm a little slow on the adoption of new technologies, like a shoemaker who goes without boots). These headphones combined with my iPhone, which has both work and personal email, allow me to reply to e-mails while driving, and I can deal with an overflowing inbox on the way home to make time for the kids, not work. It's priceless.
But can AI help even more? Couldn't he write a full answer instead of a simple hint for the next word? Promising results have already been obtained in this direction. Let's take, for example, chatbots. The primitive early versions, which were just terrible, have been replaced by much more sophisticated technological solutions based on deep learning. They still need a fair amount of training to learn meaningful responses, but now they are able to continuously learn from real data, rather than being hard-coded with a narrow set of questions and answers.
These chatbots, usually running on one of the biggest cloud platforms - Amazon AWS, Microsoft Azure, IBM Watson or Google Cloud - support many call centers and services today clients. For example, Bank of America has "Erica" (pun intended - AmErica?), a virtual personal assistant in mobile application that understands natural speech and helps you do everything from accessing an account and transferring money to receiving financial advice.
Other, more automated technologies such as Roboresponse and reply.ai, designed to operate in a more open and free world than a call center building. They are increasingly built into your favorite work tools so you don't have to move to another platform.
3. Trip Management
One of the most common enterprise tasks, at least until seamless telepresence will become commonplace, booking tickets and rooms, managing travel and processing reports on expenses. These are necessary but terribly unproductive processes. They take literally hours of work time per week, and with the trend towards a reduction in the number of administrative assistants, most of this work falls on the shoulders of leader.
Several companies are working on a solution to the problem of travel automation. For example, the Mezi service, which was acquired by American Express, helps manage your flights, hotels and restaurants. Hopper similar and even helps you save a lot by predicting the best time to book a flight or hotel. Service Pana takes over management of corporate travel and helps respond to unforeseen events such as job interviews and events.
Google Flights not only helps to find and buy tickets, but also predicts flight delays better than the airlines themselves. Incredibly comfortable. And the application Utrip even takes into account your personal preferences and social media likes to recommend tailor-made trips.
4. Researching
No wonder most of us are knowledge workers these days, the pinnacle of 21st century evolution. But our future will depend on how quickly and thoughtfully we can collect and process information. Otherwise, a younger and more technically savvy generation will start stepping on our heels. Sorry grandpa. Make your way to the future.
Let's start with the simplest research task - a survey. In many cases, we want to interview our colleagues to get the information we know they have, but we don't want to bother with individual requests. A good way is simple survey forms. Many modern platforms offer the ability to organize surveys in one form or another, from on-premise Microsoft SharePoint forms to Slack and online social networking. Online services such as Survey Monkey, Zoho and mailchimp, take it a step further and allow you to add robust conditional logic.
If your research needs are more extensive and vague, now at your disposal numerous automation tools, even if you ignore the obvious powerful AI that supports Google and Bing. Most content platforms have application programming interfaces (APIs) that can be used by developers, and increasingly there are both open source tools and commercial products that are available any of us. For instance, newsapi.org consolidates all news events into one searchable feed.
The New York Times also has an interface for developers to access content, both new and archived. Most thematic content platforms also have their own APIs, including educational materials, scientific and historical archives, as well as codes of laws and government documents.
In addition, there are tools that allow you to conduct research for specific professions. For example, ROSS is an artificial intelligence that helps lawyers do research that used to take them hours of painstakingly searching and reviewing documents. Scientists have tools from companies like Chemical Abstracts and Elsevier at their disposal.
I first immersed myself in the world of business while working in startup, who was building a machine learning system for pharmaceutical chemists in the late 1990s. Nothing compares to the old school. But what was cutting-edge then is now commonplace, and many companies are naturally adding AI features to their software.
One example of this is Tony Tripp, whose company Patinformatics has taken the concept of “collaboration with machines” to the next level. It combines open research tools with proprietary systems to provide scientists with detailed reports on the areas they are researching or hoping to get. patent. When Tony presents his findings to the scientists, they are usually amazed at how much they didn't know.
After all, scientific information, like the Internet, takes up a lot of space.
“Even tech-savvy people can't know everything,” Tony says. Some scientists think they understand what's going on in the scientific world because they attend several conferences and subscribe to several journals. But information grows so fast, eh technologies develop so rapidly that no one can keep up with them. Organizations like mine help capture the breadth and depth of what is happening in the technology fields.”
I also happened to be a witness to quite impressive - and far from all, of course - examples of the automation of intellectual work. Take, for example, the story of David Wishanoff, assistant professor of religious studies at the University of Oklahoma.
Professor Wishanoff's research interests are in how people treat those who practice other religions than their own. His interest in this problem arose when he was still a small boy. He was born in Tunisia, North Africa, in a family of Protestant missionaries, and from childhood he wanted to know more about those who held different belief systems than his family and felt a desire to understand their. “That was the topic of my thesis,” says Wishanoff, “how can I better listen to people? Part of it is my moral duty to listen to other people. We're awfully bad at listening."
A few years ago, Wishanoff met Dave King, founder of data science software company Exaptive. Exaptive finds interdisciplinary connections in complex datasets, generating unexpected insights for data scientists, researchers, and statisticians.
Wishanoff immediately realized that his research had great potential for automation. He used to scour bookstores and buy books on topics such as the interpretation of the Qur'an. But a person physically cannot read so fast as to let through huge stacks of necessary books he has found.
Exaptive's software tools allowed Wishanoff to discover surprising connections in his literature. For example, he found that some medieval thinkers sought to solve contemporary problems of Muslim society in ways that he could not imagine.
It was exciting to watch new research topics emerge from the sea of text. “Now I can focus on the work I always wanted to do as a scholar, and I'm starting to see new intellectual currents in the interpretation of the Qur'an,” Wishanoff says.
“Programs work more efficiently than humans. I immediately see which books will be most useful to me. I also see how the machine finds interesting problems that I would not even think about. Now I can explore them. I'm getting more information about what's going on in religion."
5. Information sample
Getting information is half the battle. You have to be able to put it to good use. Also in student For years, I looked down on "cheaters" who read pamphlets with brief retellings instead of reading a book. It seemed to me that this was wrong and that one should read the original books in their entirety in order to catch all the nuances.
I was a very diligent nerd. Looking back, I realize that I was wrong. Of course, in some cases it is necessary to read the whole book. But now that time has become my most valuable asset, I'm afraid to spend it on anything but the most important things. The pamphlets with the retelling are not so bad.
Today, in order to consume information, I need a business version of a summary pamphlet.
I don't want to wade through three pages of someone's reasoning. I want to understand the essence, make a decision and move on. Now I cringe nervously when I receive e-mails from colleagues on several pages, from which it is impossible to understand what needs to be done, when and who will do it. In addition to coaching in the field communicationsI need someone or something to put together a summary of my world.
And here again AI comes to the rescue. You've probably already seen examples of brief summaries in search results. This snippet of text under the heading of the search results (it's called a "snippet" sounds pretty cool, yes?) is created based on the search query, what other users found useful, and other data. But automation can do much more.
The tool called SMMRY can take an entire document or web page and convert it into a readable annotation. There are many other similar tools. This area is under active development, so in the coming years we can expect the emergence of new applications, including those more deeply integrated into Microsoft Office and Google Chrome.
Other tools help you find references to "entities" (people, places, companies) and concepts in documents, which can be useful for quickly transcribing content content, linking to related content, or even creating a "digital fingerprint" for comparison with others documents.
One such tool, Thomson Reuters' Open Calais, is used to generate news content. If you've ever seen an article with the name of a company, its ticker, and a link to its home page, thank a product like Open Calais.
Such tools for finding objects and concepts can be quite sophisticated and complex. I had to create analysis tools for companies that could decipher quotations from court documents and extract the relevant provisions from a legal agreement.
One of my favorite examples is Vin Vomero, the founder of Foxy AI. I met Vin while talking at a meeting in Boston about how to integrate IBM Watson into custom applications. He was a great guy who worked on bringing AI to real estate valuation. Shortly after our meeting, he launched a comprehensive property valuation service based on available photographs. Another reason to talk about the fact that machines take people's jobs (to be honest, I don't really feel sorry for realtors and real estate appraisers).
Via Foxy AI Vin created a new application area which he called visual real estate research. “We are building computer vision applications to extract information from photographs and then use them in other applications such as real estate appraisals. Our house2vec deep neural network was trained on millions of images of residential buildings for two weeks. Thanks to this training, our neural network has identified and learned visual characteristics that correlate with value and clearly reflect intermediate quality gradations.”
While it almost sounds like science fiction, it's not hard at all if you use the right AI tools.
Using deep learning, Foxy AI transforms photographs of properties that were previously considered unstructured data into structured data, turning pixels into a numerical representation of objects and their properties, contained in the image. Foxy AI then uses this information to improve the accuracy of existing estimation methods. “These numerical representations reflect features of the image that are associated with value. In other words, we turn the image into information about the quality and condition of the property.”
Real estate valuation is no surprise, but Win says Foxy AI is more accurate. "Catalogue rating Zillow - the most common approach, but it does not take into account the quality and condition of the object. We conducted several experiments of our own, during which we collected valuations for a number of properties for sale. We found a new house, predicted its cost, and then waited for it to go on sale. We then compared the selling price with the Zillow forecast. Our price forecast has often been closer to the actual selling price and we continue to improve the system.”
You cannot use Foxy AI directly, but if you buy or sell home, you may well be using the products of this technology without even knowing it.
6. Complex task management
It's one thing to send out a calendar of appointments or answer emails. But what about the rest of the tasks that you perform in your daily life, switching between various applications and contexts? Can they be automated too?
Turn on the march of superheroes. Products such as Zapier, IFTTT and Coda.io. These tools bring together various third-party applications, workflows, and logic conditions to solve just about any problem you can think of.
As these products incorporate machine learning into their platforms, it becomes easier to integrate them into business processes.
You don't need to be a developer for these tasks. But you need some skill with conditional logic and procedures. If you've ever built a decision tree just for fun, then you'll be fine. And even if this is not your forte, you can always hire the right specialist in the cloud.
Voice interactive systems such as Alexa, Siri and Google are also increasingly becoming platforms for automated integration of various tasks and information sources, allowing you to change the structure and logic of processes, and all this in voice management. Just don't forget to say "please" and "thank you." Your children hear you.
If you work from home or your company has adopted hybrid model, "Remote Economics" is the best guide to increase productivity and efficiency. Modern technology can do more than you think.
Buy a bookRead also🧐
- 5 tricks for maximum productivity
- What is the difference between productivity and efficiency and what is more important
- 15 task management systems that will make life easier for your team