How to start with AI?

We explain some clear steps to undertake before starting an AI project.

In the last couple of years Artificial Intelligence has drastically moved its applications to more and more industries. Regardless of a company’s size or type, its executives typically look for ways to help its organization in operating as efficiently as possible.

Many executives hear about Artificial Intelligence and the massive growth it can deliver within an organization. When considering an AI-based project, the questions which executives often ask are: 1. Where to start, 2. Do we have enough data, and 3. Who should we bring around the table to decide on possible AI use cases.

Knowing the Areas of the Business that you Want to Improve

If you are a business that has a clear plan of the challenges that it is facing and the objectives that you want to achieve, your organization may be ready to implement AI. There are almost unlimited number of business challenges that machine learning and predictive analytics can solve for your business. Harvard Business Review estimated that AI will add $13 trillion to the global economy over the next decade. Hence why you should have a clear idea of what are the most painful challenges to you, and if AI can help you solve them. Even if you are a small business, operating B2B or B2C, there are many areas where AI can help you gain competitive advantage and increase market share. AI can help and reshape businesses in areas like organization and management to process automation. You can read 5 ways AI is reshaping businesses here.

Have an Overall View of The Data

Data is an important aspect in most AI projects. Many AI techniques are based on having a lot of data which is used to train the algorithm, allowing the model to operate over new data.

In order to understand if and how AI can help you, it is useful to have an overview of the data sources that are held by your company. Having a list of the data sources, what they contain, regarding which timeframe, is important to understand what you can do with it. The more detailed the view, the more you will be able to have a qualitative feedback from an AI expert such as the Kantify team.

Having sufficient data

The data needed to create a custom-built AI tool may differ depending on the nature and complexity of the model. We have divided different AI use-cases in three categories, depending on the amount of data needed for the model to perform well:

  • more than 1.000 data examples for: Customer clustering, Sentiment analysis and Image classification;
  • around 10.000 data examples for: Hyper-personalization, Anomaly Detection, Churn Detection and Lifetime value prediction;
  • more than 100.000 data examples for: Predicting Complex Customer Behaviour, Predicting market trends and Extracting data from unstructured sources.

This is just a high level view, that differs depending on the use case. So contact us if you have a specific use case in mind so we can provide you a detailed assessment. For some companies, a lack of data can lead to difficulty in implementing an AI-based solution. For that reason, we help companies organize, process and examine their data, also point to external datasets that can obtain insights.

Know Which Stakeholders Can Support You

Like any new project, AI has to be considered in terms of change management. You are going to introduce a new technology and need the right stakeholders to support you and commit the necessary support, or budget, or IT resources, or time. This is why you should know what are your key stakeholders so they can be onboarded when time comes. Again, this is something in which we also have experience, so ask us about it !

It is essential that businesses invest in implementing AI-based solution in the area that is most needed. Identifying the overall business goals and objectives in the short, medium and long run is a good point to begin with, and the next step is assessing technologies that can help achieve those goals. Depending on the company’s challenges, different use cases can be applied. You can read more about different use cases for AI within a company here, and/or book a 15min slot with one of our team members to discuss your questions (on English, French or Dutch).


How much data is needed to build an AI solution
We explain different AI use cases and the amount of data each of these AI algorithms needs in order to be trained and work properly.Read More
Who should be your AI project manager
We share some lessons we learned from running AI projects, on how to appoint the best person in charge of your new AI project managementRead More
Short overview of Machine Learning techniques
In this article we cover four major types of machine learning techniques, we explain their applications and help you differentiate themRead More
What are the main errors to avoid when starting in AI as a company?
We have listed some errors to avoid before planning AI implementation in your businessRead More
Difference between Data Science, Machine Learning and Artificial Intelligence
We help in understanding the difference between DS, ML and AI and when to use each of themRead More
AI trends in 2020 : our selection
What can we expect from AI in 2020Read More
The state of AI in 2019: 3 discoveries you should know about
A collection of some of the great AI breakthroughs in 2019Read More
Data sources that can be used in Artificial Intelligence
We explain the different types of data and data sources companies can leverage to implement Artificial Intelligence and improve the decision making processRead More
What is the ROI on AI?
We have gathered the facts about the ROI of AI to help decision makers clarify what can be its contribution to their business profitabilityRead More