Large parts of the globe are in some form of lockdown. Business is slowing down and people have to work from home. The first reaction could be to not invest, to do as little as possible until the lockdown is lifted and the economy is restoring itself. Once the situation starts improving companies could ramp up again, right? Or can we already take action during the lockdown in order to be better prepared than the competition? This article discusses how the lockdown can be used to turn it into a competitive advantage.
Prepare for after the lockdown
Prepared companies will be able to outcompete those that do not prepare during the lockdown. The lockdown period provides opportunities to turn a company into becoming more data driven and thus into a company that better serves its customers. Employees finally have time to follow online courses and management can strategize on how to improve business using A.I. Software can be upgraded with smart AI algorithms and online presence can become more focused. Now is the time for companies to upgrade themselves and finally work on the transition to become more data driven and serve customers better, faster and cheaper. During business as usual there is often not enough time to do this. Now is the time to catch up.
Become data driven
At Datamaister we have worked for years to optimize the strategies for companies to become more data driven. The generic strategy is explained below. Since every company has unique challenges the details can differ, but we have identified the following steps that organizations require and undertake in order to successfully implement data driven ways of working: Discovery workshop, Training & Proof of Concept (PoC) simultaneously, Deploy PoC and improve user experience, Automate large parts of the software.
To get the most out of your AI projects you need to start by finding the most promising opportunities within your business. The ideal setup is to let management (C-level management), employees with practical day-to-day knowhow and senior data scientists interact with each other and generate ideas. Discussing examples of successful data driven transitions and usage of A.I. solutions provides a common understanding and a source of inspiration. Working in short 20 minute sessions with small groups of no more than 6 persons generates ideas. These ideas are discussed and pros and cons are written down. Working in iterations results in creating better ideas and further refinement. The creation of a matrix of business value and complexity about executing the idea results in a shared knowledge and concrete business decisions. Multiplying one column with the inverse of the other results in promising projects. The following example demonstrates how to do it. Please note that the example is fictional.
|Project name||Business value|
(higher is more value)
(higher is more costs)
|BV x 1/C||Priority|
|Smarter sensors||4||8||4 x ⅛ = 0.5||5|
|Better email targeting||6||5||7 x ⅕ = 1.2||3|
|Chatbot on site||3||3||3 x ⅓ = 1||4|
|Inventory prediction||9||6||9 x ⅙ = 1.5||1|
|Automated Data Scientist software||7||5||7 x ⅕ = 1.4||2|
Training & Proof of Concept
Let’s continue with our hypothetical example. After the Discovery Workshop to find the most promising projects, the Inventory Prediction project is chosen. Now there is a dual path: 1) Train people in what it means to have Inventory Prediction software and to scope better what it can, and cannot do. At the same time: 2) A Proof of Concept is created. Within three to six weeks the first version is up and running and the users know what to do with it. This is the beginning of a journey where the employees can deepen their knowledge and skills and where the software and tools can be adapted to the needs of the users.
Deploy PoC and improve User Experience
The PoC is deployed. Now it is time to adapt the rather technical PoC to make it easy and intuitive to use. The PoC demonstrated that the technology is usable, but just like the T-Ford, there is room for improvement in the usability and in what it exactly does. The interface gets a makeover using bi-weekly iterations, and some employees start incorporating the PoC into their daily workflow. These employees will be the people who can assist and train their colleagues once the final version is ready.
Repetitive tasks are automated semi-automatically. They are automated using learning mechanisms. For example, if the software detects that certain actions are repeated quite often by a particular user, the software suggests to create a form that captures all the clicks and reduces the time to execute the actions. Or if a user wants another feature, she/he contacts the developers and typically within a week the feature is available. Another example is to use Automated Machine Learning to optimize the Machine Learning pipelines that perform Inventory Prediction. There is no need to have data scientists in house, when it is possible to automate most of the tasks of the Data Scientist. This requires some initial programming, but not the continuous use of an expensive specialist. For specific requests it should be possible to consult with a Data Scientist to adapt the Automated Machine Learning.
To turn the lockdown into an advantage, find a reliable partner with a full stack development team, A.I. specialists and experienced Data Scientists with a focus on business development who are also providing services such as training and online specialists. Focus on what you can do to improve your business using A.I. and Data Science and adapt your business strategies while gaining more knowledge.