Security vendors and AI companies: product and technology contests

The landing of artificial intelligence in the security industry has promoted the further upgrading of security intelligence, and it has also brought about the external extension of this industry edge. With the deepening of the application of artificial intelligence in the security field, this industry has also attracted more and more AI enterprises, especially computer vision enterprises. As a natural algorithm manufacturer, their entry also opened security and AI enterprises in security. The artificial intelligence era about the accumulation of genes and genes.

Security video surveillance has evolved from “human eye view” to machine automatic analysis and identification and structured processing, benefiting from the rapid development of artificial intelligence, cloud computing, big data and other technologies, among which the core face recognition and vehicle feature recognition in security field Applications such as behavior recognition have gathered a large number of artificial intelligence algorithms. It is also under such industrial technology structure that AI companies have accumulated a significant influence in the security field with the advantage of leading algorithms. In the development of the industry market, the industrial roles played by traditional security companies and AI companies naturally become inseparable topics.

Breaking the limitations of industrial roles

From the perspective of business objectives, security companies have common ground with AI companies, and they all expect to provide users with smarter and more complete video solutions. Based on such a common goal, their focus is on different links. They are standing at both ends of the product and technology.

The traditional security enterprise's business focus is more on equipment R&D, manufacturing, solutions and system integration, while AI companies are more focused on chip and algorithm technology development. Security artificial intelligence from technology to products to solutions, the business between these two industry roles is bound to gradually merge. Algorithm vendors began to extend their products to improve their own products and service systems. In order to optimize the application effect of its intelligent security system, large security companies are not limited by the chip and algorithm, consolidate their market voice, and will put more effort into building their own artificial intelligence algorithm team. It is more likely that the two sides will fill the shortcomings to shape the overall competitiveness of the company.

These trends are now reflected in the industry market. It is worth mentioning that in the current market environment, enterprises with both AI and security dual deep resources are more capable of breaking the limitations of industrial roles, which is undoubtedly the lucky one of the times. Of course, they also exist in reality, and the future will surely be revealed.

Completion of short board integration development

However, the security industry is not like other manufacturing industries. It is a complex industry with multiple levels, multiple regions and multiple fields. Although many opportunities are scattered. If AI companies and traditional security companies want to survive in the era of intelligent security, they need to fill in the shortcomings and integrate development.

AI companies want to enter the security market to do deep penetration, but also need to face two major problems:

The first is the issue of industrial experience. Most of the AI ​​companies started with artificial intelligence algorithms, and their focus is on the fields of face recognition, behavior recognition, speech recognition, etc., and security is just one of the application industries in which these technologies can be cut horizontally. Most AI companies have been involved in the security field in the deep learning wave (especially face recognition) that has emerged in recent years. Compared with the long-established security companies, AI enterprises are undoubtedly the new industry. They do not have long-term experience in the security industry market, lacking a deep understanding of the security market, including real user needs, product life cycle, channel system, solution capabilities and engineering systems, which are not based on technology or capital. Can understand.

The second is the problem of technology landing. AI enterprises occupy significant advantages in chips and algorithms, but the actual security project tests the enterprise's solution capabilities. From hardware and software to systems, algorithms are only one dimension. How to transform high-quality algorithms into "productivity". Serving in practical applications is the key.

On the other hand, it is also a difficult process for traditional security companies to explore technology in AI algorithms and chips. In the industrial chain, the development from upstream to downstream may also be “following the trend”, but it will be more difficult to expand from the downstream to the upstream “upstream” because there are two major obstacles involved:

The first is that the talent structure of the company is completely different. The talent composition of security products or integration companies and algorithm companies is very different. The algorithm company's team is generally composed of technology research and development talents. The cultivation of such talents requires a strong professional environment and a long training period. In security products or integration companies, even if they use external means to quickly set up an AI R&D team, this method can't be effective quickly, even falling into a R&D pit, and after paying years of high cost of R&D, it is outdated. Or low competitiveness. In the final analysis, it is because of the problems involved in the genetics of the enterprise. The genetics of a company depends largely on its founder/founding team. If the founder/founding team does not have the thinking, organizational, coordination, and decision-making skills of algorithm development, then the following technical teams are also difficult to successfully It is very difficult to carry out work and lead to successful productization of technology.

In addition, the algorithm is a field of in-depth technical research, which requires continuous investment, whether it is high-level manpower, or research and development funds, and this continuous investment does not necessarily have a clear return, which is a test of the company's vision. And patience. Especially for small and medium-sized security companies that don't have much money and resources to pave the way, self-developed algorithms need to be carefully considered. Of course, it will be much better for large security vendors because they have stronger capital and human resources to support them. However, large companies are difficult to manage, and they have many research directions, but they are not focused, inefficient, and often dragged.

At present, some manufacturers in the industry will continue to research and invest in the day after tomorrow to obtain some technical achievements in AI. But in many cases, the effectiveness of this kind of AI is an external manifestation. From the perspective of long-term development, the AI ​​business of such enterprises is more in the form of "mixture", not pure, the core technology is still not rooted in the ground. There is not much technical output strength for practical application and long-term competitiveness. Due to the lack of natural AI genes, security vendors' self-developed algorithms may go through more difficult roads.

Cooperation and win-win

So, is it true that security vendors want to explore opportunities in the AI ​​field must be very embarrassing? In fact, traditional security vendors want to get the "boat ticket" of AI, and still can make some attempts. The most effective measure is cooperation. For security and AI companies, relying on a single technology, product or market resource can easily lead to weakness. Finding the right partner, complementing each other's strengths, can also lead to a win-win situation.

The premise of cooperation and win-win is that both parties must know where their shortcomings are. The differences in business models between different domains and the differences in the development of different companies in the same field are issues that need to be quantified in the process of selecting partners.

In the process of moving towards a refined operation, the industry will continue to generate new development opportunities. Only when the genetic advantages, technological advantages, experience advantages, channel advantages, etc. are all transformed into comprehensive service capabilities based on user needs, the advantages can be highlighted. Its value, and this is the core value of security artificial intelligence for users.

Under the absolute trend of security artificial intelligence, the future situation of the industry will undergo great changes. In the process of continuous market change, in addition to constantly consolidating its own business capabilities, enterprises must cultivate their own forward-looking landscape and die. The courage of life, open and win-win, so that we can achieve stability!