The disconnect between business and delivery teams is becoming more pronounced these days. Business wants to gauge customer satisfaction, employee satisfaction, time-to-market, innovation, cost, revenue etc. while development teams focus on sprint velocity, tracking product burn down charts etc. These agile metrics are important to track the health of your projects. However, these measures do not say anything about the value a user or customer is getting. Customer value is one of the key areas where leadership focuses on. This gap can result in last minute surprises. For example, towards the end of the project team identifies that key requirement like Security compliance is missing. There can be scenarios where project dependencies on third parties still exist towards the project end, or application features did not work the way they were supposed to work. All these can result in extended timelines. It can bring the whole team morale down and increase project cost.The digital transformation of your agile governance can be a good answer to bridge this gap.
Digital transformation is the adoption of technology to replace manual processes with digital processes. A typical example of digital transformation is automating a workflow that involves reading PDF files and storing that into your system for further actions. You can use OCR solutions to extract the text and further apply NLP (Natural Language Processing) to understand the content and take further actions. The intent is to make the system efficient and save time to increase productivity.
AI powered with NLP and statistical models not just help in getting a good project insight, it can also help in course corrections, and increase the rate of project success. It can help companies to understand their core strengths, weaknesses, and how to position themselves in the market. This requires transformation of metrics. It includes metrics like Time to Market, Customer Satisfaction, Employee Satisfaction, Innovation etc. The transformation needs applying ML or Statistical models.
Here are my thoughts on of how it can be done:
- Sentiment Analysis to track Customer Satisfaction
Sentiment analysis is also known as opinion mining. It automatically determines the emotional tone behind customer conversations. Sentiment analysis combines Natural Language Processing (NLP) and ML techniques. In order to track customer satisfaction, you can use Sentiment analysis. For example, services like AWS Comprehend can process feedback and understand the sentiments. It takes unstructured data such as social media posts, emails, and documents as input. It analyzes the input using NLP algorithms to extract key phrases, entities, and sentiments automatically. Here is an example of how it returns the results:
‘Sentiment’: ‘POSITIVE’,
‘Sentiment Score’: {‘Positive’: 0.9353844523429871, ‘Negative’: 0.0020246051717549562, ‘Neutral’: 0.05978658050298691, ‘Mixed’: 0.0028043927159160376}
You can process each feedback and can depict the score on a chart. Figure 1 below shows client sentiments over a period.
It can help the team to make course corrections.
Another important operation of Amazon Comprehend is Detect Key Phrases. It gives a list of important phrases in the document. This can be used to get a good insight about what phrases a customer uses to describe your service. It may also help in understanding company’s strengths and weaknesses from a customer perspective.
- Statistical models like Regression to understand relationship
Regression analysis is a set of statistical procedures for estimating the relationships between a dependent variable and one or more independent variables. Applying statistical models to find relationships can be extremely useful in improving your project statistics. It will be great to know what factors impact Client Satisfaction. Here Client Satisfaction will be a dependent variable and independent variables can be Innovation, Time to Market etc. If your results show that client satisfaction is related to the Time to Market, you know what you need to work on for improving client satisfaction.
- Standard Deviation to determine variance in Sprint Velocity
The standard deviation is a measure of the amount of variation in a value set. A high standard deviation means that values are spread out over a wider range. A good agile practice recommends sprint velocity should be same with the same team size and time duration of a project. We can apply standard deviation to understand the variance and if it is above a threshold, the team can be notified.
Using ML and NLP techniques we can get insights that are important for business and development. It can help improving the project outcomes. However, you need to have a good amount of historical data to build these models.
As shown in Figure 3 below, we can display important business metrics on project dashboard and send automatic notifications to stakeholders to make the process effective.
The other thing to consider is humanizing the interaction with project stakeholders. AWS Polly can be a good help here. It allows you to create speech enabled products. The service uses deep learning techniques to synthesize natural sounding human voices. During this pandemic, where in-person interaction is missing, it can be a great value to agile teams.
As Agile teams going remote, AI and ML can be a great asset for project governance. Picking right business and agile metrics helps bridging the gap between business and delivery. All this needs to be backed up by Machine Learning and Natural Language Processing. The transformation can give deeper insights on agile projects. It can help in teams executing successful projects and company reaching new heights.
