Thursday, February 29, 2024
HomeDXWhat is a PoC (Proof of Concept)? Easy-to-understand explanation of three points...

What is a PoC (Proof of Concept)? Easy-to-understand explanation of three points for successful AI development!

With the evolution of digital technology, there has been an increase in “aggressive use of IT” to increase sales and profits. Aggressive use of IT has the potential to produce great results, but if the project fails, there is a risk that the return on investment will not be fully obtained. The process that is needed there is a PoC . This time, focusing on AI development, what is PoC? What should we do to be successful? I will explain about.

What is PoC?

PoC (Proof of Concept) refers to the process of verifying possible risks at a minimum cost in project development .

If you invest in this development without being confident that the project will succeed, you run the risk of technical problems causing the project to fail. “PoC” is the process of trying out various ideas and technologies at a minimum cost before moving on to actual development . As an example of how to put in effort, if this development is a straight right, PoC is about a jab. The first shot is weak, but you can try many shots.

 

Points for successful PoC
1. Share the recognition within the company that “PoC will proceed in a prolific, prolific style”
A PoC is an “experiment necessary for success” , so it’s okay to fail a lot. Through failure, if you understand the cause, you can use it to succeed.

A common pattern is that once the PoC fails, the company will stop investing in the project as a whole, thinking, “Why can’t we do this when other companies are successful?” However, what I want you to notice here is that the success stories rarely mention the failures that led to the success . Because I want to make it look easy and successful. If it were easy to succeed, the word PoC should not exist.

Especially in AI development, there is a risk that “you won’t know the accuracy until you make it” due to various factors such as the content of development, the amount and quality of data, the level of knowledge about AI of the person in charge, the environment of the company, etc. PoC success rate is low. Therefore, it is necessary to diversify and invest in as many projects as possible. First of all, it is very important to share the recognition within the company that “PoC is a prolific and prolific style” and that “even if you fail, you can learn.”

Specifically, the team first divides up the work and thinks about 200 issues and solution ideas as a set (for example, 20 to 30 key members from each department, 10 ideas each). Next, we discuss the necessity and feasibility, narrow down to 10 by majority vote, and quantitatively analyze these to roughly estimate the return on investment.

Finally, let’s go through the process of simultaneously conducting PoC for 5 with high return on investment, and among them, one or more will succeed, and you will get a large return while recovering the cost of the others. . If you think that it will be difficult to proceed with five projects at the same time because of the lack of human resources, time, and money, decide on a period of several months or half a year, and proceed in order from the solution with the highest quality. However, it is clear that the ideas that remained as candidates for PoC in the end can expect a large return if successful, so if the company shares the recognition that “PoC is promoted in a multi-producing and multi-death style”, there is no need to worry too much. I don’t think so.

2. Thoroughly carry out “requirements definition”
In PoC, the “quality of requirements definition” determines the success or failure of the project.

Many companies that don’t develop their own products outsource PoC development to vendors. The vendor develops based on the requirement definition of the system desired by the customer. Therefore, it is necessary to thoroughly define the requirements so that the vendor can easily develop. If the requirement definition is loose, the development will inevitably be less likely to succeed because the vendor will not know the target standards during development. Also, from the vendor’s side, they should prefer to develop “things with a clear purpose” rather than “things they don’t know what to make” .

In the case of AI development, create a requirements checklist by writing down what you should think about in detail about the following five points. (The detailed description is just an example).

Expected ROI: What problem do you want to solve? What is the purpose of the system?
Data preparation: Do we have data? how to collect
Model development: what model to use? How much accuracy do you need?
Development environment: What kind of environment do you study in? Need to relearn?
Operation/Relearning: What is the overview of the entire system? What are the API specifications?

3. “Respect” and “cooperation” between business and development are important
Proper internal cooperation between the business side and the development side is essential for project success, not just for PoC. If the business side has little knowledge of technology and only plans, and the development side has little knowledge of business and only develops, it is unclear whether the development side can realize what the business needs. As a result, there is a possibility that it will not reach the level of practical use, or that something different will be produced. First of all , it is important to respect each other and discuss what you want to do in this project and whether it is technically possible .

summary
Finally, let’s review four points that are particularly important in PoC.

PoC is a process for verifying possible risks at a minimum cost in project development.
Share the recognition within the company that “PoC will proceed in a prolific, prolific style”
Thoroughly carry out “requirements definition”
“Respect” and “cooperation” between the business side and the development side are important
For those looking for AI technology OJT and development support services
“I acquired knowledge of AI development in the course, but I stopped because there was something I didn’t understand in practice.”

“I would like to ask professional personnel with experience for development.”

For those who have such concerns, we have a “digital technical advisor” who supports the development of internal human resources with online OJT by professionals who have been involved in many developments at domestic AI startups, and an “AI implementation support” who is responsible for product development. We have prepared a service for

RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Recent Posts

Most Popular

Recent Comments