As IoT grows, advanced AI models are critical

Not all AI models are created equal

基于代码，hand-built AI modelssimply are not built for the requirements of a hyper-personalized customer experience. Today’s businesses need 24/7, “lights-out” evolutionary programming that can run without human intervention. Lights-out models never stop looking for opportunities to monetize customer data. An added advantage is that they remove the data scientist from the equation and put the power of AI and algorithmic optimization squarely in the hands of marketers.

但是，为什么不符合该任务的基于代码的模型呢？首先，他们根本无法跟上当今的始终与之联系的客户。当新的渠道或数据源出现时，数据科学家需要完全重新配置算法并构建新模型，从而浪费宝贵的时间和预算。

这物联网的增长means that new data sources are emerging all the time, making code-based models more cumbersome to manage. Rather than falling further behind, businesses should be using the combined power of IoT and AI to their advantage. In fact, 80% of all enterprise IoT projects will include an AI component by the end of this year, according to加特纳。

最后，基于代码的模型无法符合AI资格标准。它们建立在预测规则的基础上，这些规则不是动态的，随着时间的流逝而变得陈旧。没有在线优化引擎，基于代码的模型需要人类干预才能刷新，从而使动态的客户旅程始终保持一步。一旦他们陈旧，人类就会被要求重建模型以巨大的费用和时间重建模型。

Embrace automated AI and put the customer first

Even with these drawbacks, code-based models remain popular. It seems as though businesses are reluctant to explore the fullpotential of automated AI作为收入驾驶引擎。

这是企业采取的危险立场，因为它起源的数据和来源的复杂性只会继续扩展。随着互联网连接的传感器数量不断增加，房屋，汽车，建筑物等等，企业正在积累大量数据。至关重要的是，组织利用始终在AI模型中从该数据中提取有用的客户信息。

此外，有37％的消费者表示，他们将不再与一家未能提供个性化体验的公司开展业务，而63％的人表示，个性化是一种预期的标准服务Harris Poll survey

For this reason and others, automated AI is a business imperative. Forward-thinking and data-driven organizations have already implemented evolutionary models into business processes and have reaped the benefits of superior customer experiences. It’s time for the laggards to catch up before they miss out on additional customer touchpoints and opportunities to capture revenue.

所有的物联网议程网络贡献者都负责其帖子的内容和准确性。意见是作家的，不一定会传达物联网议程的思想。