敬业度管理 | 用人工智能来侦测组织情绪

敬业度管理是组织发展和人力资源管理领域里的一个重要话题，企业经营者需要对组织的员工敬业度水平具有可视性，分析影响敬业度因素的正向指标和负向指标，采取相应的管理行动，并且评估管理行动的效果。

“敬业度engagement”常被呆伯特拿来开涮

各家人力资源咨询公司的员工敬业度模型大同小异，常见的员工敬业度因素包括：组织价值观、沟通、身心健康、工作环境、职责明确、同事关系、奖励、薪酬激励、经理人影响、职业发展等。

组织的员工敬业度和组织文化氛围会相互影响，并且直接决定了个人绩效和组织绩效。

员工敬业度测量和信息搜集的主要手段是调研，以这种方式来搜集员工反馈，具体方式有：

-     年度敬业度调研（Engagement Survey）

-     季度/月度的脉动调研（Pulse Survey）

-     从企业内协作平台或社交工具的文字中获取信息

怡安敬业度调研挺有名，这是其几年前的移动产品

传统上，大型企业都利用年度调研方式来评估敬业度，进行分析后，采取相应管理手段；近年来，由于组织环境变动越来越大，而且就业灵活度和员工流动程度也越来高，互联网企业等新兴企业更多采用脉动调研的方式。

最近几年新兴的脉动调研平台，都将产品包装为“员工体验”反馈平台，在入职、表彰、离职、敬业度评估、绩效评估等人力资源管理事件时，搜集员工反馈。

人工智能里的语义分析、情绪分析（Sentiment Analysis）、机器学习在敬业度分析及管理大有可为，可以看到的应用场景是：

-     分发员工调研问卷、回收答案，评估组织、团队和经理人的效能

-     运用人工智能分析开放式问题的回复，透过字面意义，准确分析员工对工作环境以及领导层的感受

-     基于定性描述和开放式问题回复，洞察到改进行动建议，并且在这些行动发生后，评估效果

-     员工体验或敬业度的反馈信息，不仅可汇集企业内部数据，甚至包括社交媒体，LinkedIn，Glassdoor等职业论坛上的大数据

- 利用机器学习，积累行业特性、组织特性的敬业度知识、驱动要素分类

IBM几年前收购的人力资源管理软件Kenexa在敬业度调研领域处于领先地位，随着IBM向认知计算的转型，Kenexa也被整合到了IBM的人工智能平台Watson产品线里

IBM自身的人工智能员工敬业度管理案例是：当TA采用情绪分析工具来分析员工调研和员工社区反馈，发现一个显著的主题（theme），员工对公司配发的电脑不满意。IBM随之开始向员工提供苹果Mac电脑的选择，也就是我正在打字的这种。

人工智能用于组织文化和敬业度评估，比较有代表性的平台供应商，除了IBM外，还有一些新兴公司，值得关注，例如：

Glint：https://www.glintinc.com/

TINYpulse： https://www.tinypulse.com/

Culture Amp： https://www.cultureamp.com/

这些应用不少是搭载在IBM Watson，微软Azure这些人工智能云平台之上，借用了这些平台的人工智能分析能力。

人工智能在人力资源领域还广泛用于招聘，除了个性测评、海量简历筛选（文本信息提取）之外，甚至用于候选人自动视频面试，用AI视觉来评价候选人反应能力、测量是否撒谎。

附录：

How Google Uses People Analytics to Create a Great Workplace

Ever since the term "employee engagement'' started being widely used in the 1990s, it's been hailed as the key to high productivity and retention, profit increase and better customer satisfaction. Companies took notice of the benefits and sought to turn this abstract concept into a trackable metric. Many companies began administering annual employee-satisfaction surveys.

These snapshots help companies check in with team members and assess how happy their employees are in the workplace. While the practice now is widespread, some thought leaders in human resources (HR) have begun questioning the annual survey's accuracy and usefulness.

The problem with most feedback tools.

How do you know the answers you’re getting are valid in terms of what you’re trying to achieve? A survey by Impact Achievement Group and HRmarketer discovered 48 percent of all respondents did not believe employee surveys provide an honest and accurate assessment, compared to 31 percent who thought surveys painted a true picture.

Some argue that employees are more apt to answer survey questions positively -- creating a sense that everything seems fine (at least on the surface). Employees who give falsely high marks might fear retaliation or feel a general disinterest toward a survey that takes time out of their busy work day. Others might believe their answers won't make a difference. This sense of apathy is evidenced by the simple fact that many companies have trouble achieving high participation rates.

Here's the most revealing finding from the Achievement Group/HRmarketer study: 58 percent of respondents agreed that results did not -- or only slightly -- helped managers gain a better understanding of what behaviors or practices they could change to improve. If surveys don’t yield any actionable information, will their results make any difference in how a workplace is run? This loop perpetuates the apathy problem. If employees are conditioned to believe they won't see actual changes in their work environments, what incentive do they have to fill out a satisfaction survey?  

A recent LinkedIn post from Forbes writer Liz Ryan had some pretty harsh words about employee-satisfaction surveys: “Employee Engagement Surveys are the business equivalent of giving the prisoners in a penitentiary a survey to complete once a year and slide through the bars of their cells. The survey process cements an unequal power relationship.”

Is she right? Are engagement surveys an HR check-off box at the least -- and a tool for damage control at most? Do these surveys simply give employees the pretense of a voice within a company? Maybe it’s time to rethink employee-satisfaction surveys and how we administer them.

How Google does it differently.

At Google, surveys aren't just about checking the pulse of the workplace, they’re about constantly striving to improve it. The company’s People Operations team (formerly known as HR) uses feedback to optimize different aspects of its people processes and align them with its unique work culture. As a result, the company reports an average participation rate of 90 percent. 

Nearly every decision the company takes is data driven -- and that's representative of the culture in a majority engineer workforce, too. But the very nature of the HR field focuses on interpersonal relationships in the workplace. It can be difficult to assess based on pure input/output metrics alone. Productivity metrics are extremely important to gauge effectiveness, but they don't tell the whole story. 

For this reason, Google integrates the human aspects through use of people analytics. Mixing quantitative and qualitative data enables leaders to really dig deep into the company’s inner-culture dynamics. Google has used people analytics to improve the workplace across a number of studies. A few notable examples appear below.

Leadership Case Study: Project Oxygen.

Through surveys, company leaders learned that most Googlers are averse to hierarchy. Many employees -- cofounders Larry Page and Sergey Brin among them -- questioned the need for a management level. They thought they simply might flatten the company structure. First, though, they engaged the company's people analytics team to determine whether having managers makes a real difference.

In Project Oxygen, People Operations workers analyzed managers' performance ratings and the upward feedback gathered in employee surveys. They then compared these results with productivity metrics. The outcome? At Google, great managers lead to more engaged and productive teams.

The people analytics team took the research a step further to identify which characteristics made a great manager. The team returned to comments from surveys, performance evaluations and great-manager nominations. They then conducted double-blind interviews with the company's highest- and lowest-rated managers.

Rather than prove managers unnecessary, Project Oxygen discovered mid-level leaders were essential to create conditions in which other employees could thrive. To be truly effective, though, the company had to turn away from supervisory tactics that ran counter to what Googlers need from a manager. Micromanaging was at the top of the list. 

Google turned Project Oxygen's findings into the company's Top 8 management behaviors and now uses the list as a guidebook to identify and train leaders from within Google's ranks. The initiative's data-based roots make it easier for managers to accept these standards and move toward meeting expectations. 

Teamwork case study: Project Aristotle.

In another study, Google sought the perfect formula for creating effective teams. Leaders looked at each team's performance metrics and then at perceptions of effectiveness itself. People analytics acknowledges that abstract terms means different things to different people. Through Project Aristotle, Google learned that executives equated effectiveness with productivity. For employees, team culture was the most important measure of effectiveness. Meanwhile, team leaders ranked ownership, vision and goals as the most important influences. 

To capture these varied factors, the people analytics team examined qualitative surveys from the three major employee groups and compared the information to sales performances (stacked against quarterly quota). Combining human experience with hard data allowed Google to see which teams fared the best. Even more important, the results helped Google understand why certain teams succeeded. These findings formed the basis of Google's five essential factors to create a positive work environment.  

Translating Google's lessons to other companies.

Google's process provides HR insights into employee engagement. It also creates trust between employer and employee. Googlers feel a sense of equality because they directly shape how their company is run.

Laszlo Bock, Google's former Head of People Operations, described the strategy in a recent Harvard Business Review article. His piece listed four steps leaders can take to "move from hunches to science" in their own decision-making processes. Bock's strategy (1) asks employees to identify the company's most pressing people issues and (2) suggest ways to improve, (3) encourages the analytics team to share feedback company-wide and (4) finally empowers leaders to run experiments that test which data-supported theories work best.

It's impossible to overstate the importance of engaging the human aspect in HR efforts. Pure performance data alone isn’t enough. Qualitative results help leaders truly understand the underlying dynamics at work in their companies. Done right, employee surveys play a crucial role. Start by asking what workers want to improve and take the next steps to move forward from there.