注册 / 登录

Use Conversational AI to Realize Insights As A Service with Big Data

分会场:  人工智能/AI驱动/AI实践

 

案例来源 :

案例讲师

Mary Hu

Microsoft Data scientist

Mary Hu: Data Scientist at Microsoft Azure Core Data Team. 3+ year experience in data science, worked in A/B experimentation, NLP, Conversational AI, etc.
Sophia Peng: Data Scientists at Microsoft Azure Core Data Team. Speicializes in deep learning, contextual bandit, statistical modeling, etc. PhD from UCLA

扫描二维码分享案例

 

案例简述

 

We have built a Conversational AI that understands the business metrics a users asks, fetches the data from different databases, join multiple sources together if needed, generates meaningful visualizations to help the user navigate insights from the data, and all of these processes happen on the fly instantaneously. This largely reduces the time we spent on finding the right data sources, learning the join keys, understanding the exact formulas needed to calculate a KPI, etc. All of this was realized using different types of Microsoft products and can be applied to any company, or any different domains.
We have been building a platform where any team can onboard their own data into the system, define the business scenarios they have, and the platform will take care of the rest: generating synthetic data to train the NLP model, generating scripts for visualization, getting connected to the UI/application that the user wants to use to ask questions.

 

案例目标

 

The goal is to reduce the inefficiency on generating insights to solve business problems using big data, which can involve steps like finding the right people, right data source, right logic to apply, right visualization to present the insights, etc.. We solved it in a Conversational AI approach, while leveraging a wide variety of Microsoft products: “Hadoop-like database”, Azure Cognitive Services, distributed statistical language, Microsoft Teams, etc.

 

成功(或教训)要点

 

Different domains can have completely different business scenarios, and balancing between a generic solution and a deep-customized solution is a design trade-off.

 

案例ROI分析

 

Reduced the time needed to answer account manager type business questions by 90%.


 

案例启示

 

When well-established business processes encounters big data, the inefficiency will increase as the data volume and complexity increase. AI can help reduce the inefficiency generate new insights.

 

领取大会PPT

我要参会

大会全套演讲PPT

立即领取

大会即将开幕,点击抢票!

我要参会