domeet webmaster

Hello everyone, in recent years I have been studying computational advertising on the Internet. The market lacks a systematic introduction to this field, so I have also written a field of computational advertising Book and was fortunate to be recognized by the internet industry.

Today I want to discuss with you a case: Will artificial intelligence help marketing or brand services in the new era?

It is generally believed in the society that artificial intelligence is the deep learning of algorithms by machines. But I can tell you exactly that this concept is wrong.

Artificial intelligence was proposed by Taishan Beidou in the computer field. The purpose is to use computers to replace people and solve some intelligent problems, such as replacing human listening, speaking, reading, writing, and communication. Artificial intelligence can do this with or without machine learning.

That said, machine learning is a method and tool for artificial intelligence, not artificial intelligence itself.

I. 2 key consciousness of marketing and customer service

In the narrow sense, the field of artificial intelligence mainly includes “perceived intelligence” and “cognitive intelligence”: cognitive intelligence replaces human features, and cognitive intelligence replaces human thinking and communication.

Boston stated in a book that artificial intelligence can provide customers with personalized services, advertising and interaction in terms of corporate marketing services.

All existing advertising marketing is non-interactive with the user, and it’s over to the other party like a flyer. But the data shows that brands that can create personalized experiences can increase revenue by 6% to 10%.

Everyone knows the word personality. Today’s advertising is personalized, with thousands of people. But using artificial intelligence technology can allow advertising marketing to interact with users on a personalized basis.

I see a lot of people now, when it comes to artificial intelligence marketing or intelligent services, they always say this set of words:

“How many active devices do we have “,” How much data is collected through the SDK “,” How many tags have been constructed “,” Identify the target population “,” Precise delivery “, etc.

These words cannot be verified or falsified. I believe they use the technology of machine learning, but it cannot be said that they use artificial intelligence itself. The topic we are discussing today is to use artificial intelligence itself to promote marketing services.

This process is quite difficult, because most people’s understanding of artificial intelligence itself is still in its infancy.

All advertising marketing today has a feature that takes the initiative. But it is mainly one-way publicity, and it doesn’t matter if the advertisement is sent out. This kind of advertising marketing is badly lacking in communication and interaction.

When developing marketing, a very useful technique is: targeted interaction. In this way, using data and machine learning, you can flexibly respond to different situations and put advertisements in thousands of people.

Look at customer service again. In contrast to marketing, customer service is where users come directly to find problems, but do not do research on users, but instead solve problems on the basis of user problems.

These two tasks are essentially one thing, and the insurance salesperson does the best of both.

The insurance salesperson does not distinguish the marketing and service (customer service) process. First, he will communicate with you and become friends with you. The purpose is to obtain your information. They are very patient in this process.

In marketing, we are passive in obtaining user information, while insurance salespersons are active, just like customer service, and choose sales based on the needs collected.

In terms of after-sales service, insurance salespeople are also very concerned. We often send small gifts to users during the holidays, ask what is happening at home, collect feedback, and continuously track and update user information. The ultimate goal is to do secondary marketing.

This seamless integration of marketing and services has two points worth learning: one is service awareness and the other is customer information awareness.

In the future, we may be able to imitate the approach of an insurance salesperson and build a virtual marketing process. Of course, this will take time.

Second, we must have more imagination of “agent”

Existing artificial intelligence research is divided into four categories: perceptual intelligence, computational intelligence, motion intelligence, and cognitive intelligence.

The problem of perceptual intelligence is the problem that people want to solve for a long time. Perception includes eyes, nose, ears, speech recognition, etc. These have evolved for the longest time.Animals have evolved with these senses.

For the problem of computational intelligence, artificial intelligence can find a fast solution path with extremely high computing power. This is very difficult. The most typical example is Go, but relatively speaking, artificial intelligence has already solved the Go problem earlier. . Today, machines can kill humans in all game-like problems.

Motion intelligence has also evolved the longest, and it has begun to evolve from the primitive period.

Cognition is a unique ability of human beings and a source of reasoning, thinking and communication. Our communication, dialogue, and understanding are all cognitive intelligence.

Artificial intelligence can surpass humans in game issues, and humans will evolve in three aspects: perception, motion and cognition.

We have also made a lot of discussions on artificial intelligence research in the field of marketing services ——

I have experienced the interactive experience of calling a robot on the phone. It feels natural to chat with it. If it is not a professional, it is impossible to tell that it is a robot. This is a big step in the field of interaction.

Outbound robots have a good effect in the marketing field, which is equivalent to an insurance salesperson who is making suitable product recommendations for you. This concept is “precision marketing”.

However, the robot customer service is the opposite of an outbound robot. Customer service is service and outbound marketing. However, the logic of the two is similar, and both use technology to solve the problem of marketing services.

In summary, the direction of brand marketing and service in the era of artificial intelligence is to build perception and cognitive capabilities, to achieve a service agent for the entire life cycle of each customer, to record and analyze user data, and to provide continuous and interactive services. .

This agent must have a brain (interactive things), available products, a personalized model (similar to the service awareness of an insurance salesperson, and serve customers in specific situations), and customer information.

For example, Grandpa Whitebeard KFC can completely become an agent in the era of artificial intelligence. All marketing and service channels use this image to interact with customers.

Now machines can already actively identify themselves. Of course, there is a regulatory boundary here. For example, the screen of Focus Media, with reasonable authorization, can actively interact with people.

Although this scenario has not yet been realized, we can imagine it.

III. Weaknesses of artificial intelligence

Soon, machines will be closer to humans in the field of perceptual intelligence.It is therefore available in active cooperation scenarios.

But don’t think that artificial intelligence has reached an inflection point. Its fundamental problems have not been solved yet, and the ability of artificial intelligence in cognitive intelligence is much worse.

The problem of cognitive intelligence is very difficult. The inspiration of everything depends on cognitive intelligence.

Our conceptual reasoning is not learned from school education, just like a college student and an illiterate to communicate basic concepts without any obstacles. More than 80% of human concepts and logic were learned in a very wonderful way before the age of six.

Now, all artificial intelligence is weak artificial intelligence.

For example, this common sense—I drank a bottle of water, so the bottle became lighter—the machine ca n’t learn, and for example, some news, when people hear it, it is fake news, but the machine does not know.

Because of this, we need to have very complicated and wonderful common sense, and the machine cannot find a corpus that can learn these common sense. Therefore, it is impossible for the machine to catch up with people in understanding and inferring concepts.

This also means that our imagination is to focus on marketing and services through human-computer interaction. If it is in a broad field, it will be difficult.

However, in some vertical fields, such as medicine, we feel very professional, but it is very simple for machines. This is because the professional knowledge books are available on the machine. As long as there is a corpus, the machine can learn.

Therefore, cognitive intelligence is currently the biggest challenge for artificial intelligence. This cannot be solved simply by functional improvements or resource input. This is a theoretical gap.

* The article is an independent opinion of the author and does not represent the standpoint of Noteman.