How Are AI Projects Quoted, Scheduled, and Accepted?

We mainly undertake AI applications, enterprise knowledge bases, adding AI functions to existing systems, as well as system go-live and privatized deployment. You can develop independent applications or integrate AI functions into existing CRM, ERP, OA, applets or internal systems.

Yes. Many of the projects we receive already have requirements documents, prototypes, and even running systems. The parts that can be used will be retained, and the parts that need to be adjusted will be dealt with in conjunction with this project. In this way, existing investment can be utilized and duplication of development can be avoided.

If you already have a requirements document, prototype or feature list, just send it directly. If the budget and launch time have been determined, you can also tell us together. It doesn’t matter if the information is incomplete for the time being. We can first understand the project based on the existing content and then see what needs to be added.

It is also an AI project. The investment will be very different between connecting to an existing system and developing a set of independent applications. We will give you a quote based on the functional scope, interfaces, data and deployment methods. External costs, such as model calls, cloud servers, or software licenses, are also listed separately.

Start once the contract, down payment and necessary information are in place. The number of functions, interface difficulty, data quality, deployment environment and acceptance requirements will all affect the construction schedule. The project schedule will list the development, joint debugging, testing, launch and acceptance times, as well as the dates that require the cooperation of the enterprise.

It is common for business requirements to change during the course of a project. Small adjustments can be processed with the current version; if the change has affected the development scope, construction period or cost, we will clearly explain the impact and discuss with the enterprise how to arrange it. If you encounter technical problems that cannot be determined temporarily, you can also do a short verification first.

Acceptance is subject to the business scenarios and test data confirmed by both parties. Mainly check whether the business process is available, whether the permissions are correct, whether the interface is stable, and whether the AI ​​output meets the agreed requirements.

We are responsible for solution, development, system integration, testing, deployment and delivery. The enterprise is responsible for confirming business rules and data permissions, providing information, interfaces, accounts and operating environments, and arranging people who can confirm requirements and acceptance results. When you need to purchase third-party services, we will indicate who purchased them and which part we are responsible for.

The delivery content includes a runnable system, the code and configuration agreed in the contract, interface descriptions, deployment materials, acceptance records and maintenance instructions. Which results belong to the company and which belong to open source or third parties will be specified in the contract. Enterprises can maintain it themselves or entrust us to continue maintenance; new functions will be quoted separately.

Corporate information is only used for current projects and will not be displayed publicly. We will explain what data the project needs to use, who will contact it, and whether external models need to be called. When sensitive content is involved, desensitized materials or company-specified environments may also be used.

No particular model is used. The choice depends on mission effectiveness, data requirements, deployment location and cost. The plan will list the model version, account ownership, calling fees, and whether the data will be sent to external services.

The system can be deployed in public cloud, the enterprise's own cloud account, private cloud or intranet, or it can be deployed in a hybrid manner. Deployment diagrams indicate the location of applications, data, models, logs, and backups. Private deployment does not necessarily have to be completely offline, and whether to connect external models or services will be noted separately.

We will not accept projects that violate laws and regulations, are illegal, bypass system permissions, or collect data without authorization. For medical and financial projects, if the company does not have the corresponding qualifications, data authorization and professional reviewers, we will not accept them.