Posted on October 29, 2019 at 8:50 AM (WAT) updated on December 9, 2019 at 6:35 PM (WAT).
Data is one of the most powerful resources that can impact the growth of an organisation as well as the effectiveness of the decision making process. The 21st century is referred to as the era of data due to the extent to which individuals and organisations have become data-concious. Proper investigation of an organisation's data can unveil a lot of important information that can drive decision-making. This is why many organisations are now starting to integrate data analytics into their workflow, to make more informed decisions and to enhance business growth.
Data analytics is the process by which professionals (called analysts) apply a set of tools and formal approaches to derive valuable insights from data. The purpose of this in a typical business setting would be to generate some sort of report which explains the current state of the business as revealed by the analytics processes and to state some observations that are worth mentioning. In this fast-paced era, business requirements are constantly changing, decision makers have new questions to ask about their data and expect a quick response in order to determine the optimum approach to an emerging business need as early as possible. This might not come as a surprise, as we have learned that data analysis is a question-driven task. This implies that we expect analysis to kick off with some basic questions which the analyst is expected to (hopefully) find answers to. It also implies that these questions need to be anticipated for their answers to be included in the results which can't always be possible. Managers need to wait for further investigation to get answers. Even for a team of skilled analysts, this process can still take a while before a suitable solution can be proposed.
At the core of data science/analytics/analysis are usually questions which trigger specific analytical tasks. For example - 'How is advertisement in Abuja contributing to overall sales?', 'What should be the expected sales margin if we were to extend campaigns to every city within 10kms radius of Ibadan?'. The job of a data analyst (or scientist) isn't just to provide tables and charts, but also to provide answers specific to certain questions. In other words, the data analyst replies to those questions through his/her results in a format that can easily be comprehended by executives with the least technical background. I could remember while working as a data scientist (intern) at Crop2Cash, my boss was most interested in simple answers to questions that required some sophisticated technical approach before a definite answer could be provided. This iterative process of questioning and answering more or less makes data science (or analysis) a sort of dialogue between executives and analysts especially with respect to a business environment.
Technology is replacing traditional methods to speed up the rate of obtaining results. If decision-making is crucial to any organisational structure (which it is), then it needs be speeded up. We have seen that making data-driven decisions is dependent on how fast data analysts can provide answers (recommendations, predictions, status report, etc.), this also implies that we need to speed up this result-generation process to enhance the speed at which crucial decisions are made. Since the interaction between these two parties is similar to a conversation (if you agreed with me on the previous paragraph), then automating conversation might be the best approach to this challenge.
Conversational AI is a mean by which a dialogue between two parties can be automated. It applies different
in natural language processing (a branch of AI) to enable automated agents (software) to mimic the response of
agent. This is done primarily to save time and efforts much of which would have been spent on repetitive tasks,
humans can then focus on tasks that require specialised expertise.
The applications of conversational agents can be seen virtually in every industry today in form of bots; Marketing, Travels, Finance, Education etc. Integrating conversational agents with data analytics might imply that there will be no more complex diagrams and tables and scrolling needed as what we need is just English. Results will be provided in simple human language plus the experience is completely interactive, anybody can relate with it (even that grumpy manager will be impressed).
Imagine a situation where you as an executive in an organisation can just enter some question into a text field like - What should we expect from sales this year? and you get an answer like - Well, the sales forecast for this year predicts a 23% increase for Nexrios and 12% increase in Tellvines. However, Ziamine wouldn't be doing so well.
Automating analysis is one step, but providing the results of this analysis interactively in a way that looks very natural and comprehensive to anybody is a huge step to bringing the benefits of data analytics to every business no matter how small.