The developed health assistant has the functions of intelligent analysis of health data inside and outside the hospital, health reminder, etc. The advantages of AI health assistant management group compared with conventional management group in terms of comprehensive compliance rate, metabolic index level, hypoglycemia incidence rate was further studied.

1. The investigators have developed an AI health assistant suitable for diabetes patients, which has the functions of automatically uploading blood pressure, blood glucose data, intelligent reminder, automatic analysis of reports inside and outside the hospital, intelligent question and answer, etc. it is simple to operate, highly interactive, and maximizes the management level of diabetes patients outside the hospital.
2. Through AI personal health assistant, 196 diabetes patients were managed to further improving the comprehensive compliance rate of metabolic indicators such as blood glucose and blood pressure of diabetes patients, improving the patients' self-management ability.

Inclusion Criteria:

The diagnostic criteria of T2DM patients are the diagnostic criteria of the 2017 China guidelines for the prevention and treatment of type 2 diabetes HbA1c 7.5-13% Age: 18-65y BMI 18.5-30kg/m2 The course of disease is less than 5 years Have good cognitive ability and can correctly use health assistants

Exclusion Criteria:

Acute complications of diabetes (diabetes ketosis, etc.) Diabetes complicating pregnancy or preparing for pregnancy HbA1c < 7.5% or > 13% Age < 18y or age > 65y Pre pregnancy BMI < 18.5 or > 30kg / m2 Severe liver and kidney dysfunction (ALT greater than 2.5 times the upper limit of normal, EGFR less than 45 ml / min / 1.73m2) Using drugs that may affect blood glucose are being used (including steroids, hydroxyprogesterone hexanoate, anti AIDS drugs, etc.)