4 years Arinti – 4 lessons to lead an AI project to success

Data science for business is a hot topic. No wonder, given the massive amounts of data being produced every day. Its popularity has steadily grown over the years, and companies actively started implementing data science techniques to increase their business and improve customer satisfaction.

If you’re looking into putting your data to good use, know that an AI project always comes with a learning curve. And for some companies that learning curve might be quite steep.

So, before you jump in head-first, we are sharing the four most important lessons we learned the past four years to turn an AI project into a massive success.

Be open to change – and communicate transparently about it

Embrace change and ensure your company is ready to follow your lead. For people to put their confidence in technology and Artificial Intelligence solutions, you have to be exceptionally transparent. Create a culture of innovation, technology and openness to change.

Implementing AI in your company always comes with some degree of change management. It goes without saying that transparency should be your top priority when communicating to employees.

Be meticulous about explaining what AI does, how it operates, how it can be trained, and where it collects its information. Indicate that AI decision-making is always based on the input: your datasets and databases. Not only tell your employees about AI, but also show them how it can be applied to steadily expand business.

Case – Unilever

At Unilever, Arinti implemented AI using 10 years of data. An impactful change in their way of working. Robin, Unilever project lead, made it his duty to spread AI awareness within the company.

Empower everyone – inside and outside your company

Involve the people! Especially those who are carrying out the project and those who will utilize the final technology. AI empowerment is not something you drop and expect your employees to pick up. You need to nurture a culture of innovation.

Creating this empowerment bottom-up might take a longer time, and require a vigorous effort on your part. However, in the long-run, you’ll experience much less resistance.

Our number one tip? Get the IT department involved from the get-go. They are an incredibly valuable stakeholder in the not-so-distant future. After all, they hold the key to all important data you might need.

Don’t forget about your external stakeholders, either. What better source of information than your ideal client? So organize workshops with your customers and gather their feedback, insights, and input.

Don’t forget about the hidden costs – those out of scope

In 99% of all Artificial Intelligence projects – or any project for that matter – you will pile on extra costs. These so-called opportunity costs mostly consist of the time and effort that different internal stakeholders invested.

We’ll give you a few examples.

• To implement Data Engineering and AI, we involve your IT department to gain access to the different systems and datasets.
• On top of that, we need experts who know all the ins and outs of your business and industry.
• Since AI success doesn’t happen overnight, you’ll want to cultivate ambassadorship within the company. A person or a team takes on this task and invests their time into defending and explaining to peers, managers and stakeholders how AI will help grow the business.
• And, when moving into a later stage, employees put their time into learning and working with AI.

You see, time is your most precious resource and an essential part of your AI business investment. One extra piece of advice: don’t underestimate the amount of time that’s needed.

Decide on your goals – before you put the wheels in motion

If we had to boil it down to just one lesson, this would be it. Without an identified goal, the chances of an AI project tapping into its full potential are pretty much non-existent.

No goal = no AI project.

Allow us to elaborate. Leaving your goal open-ended eats away at your success rate. Without a goal, an AI project most likely won’t be viable or go into production. And if it does, it’s a risky endeavour.

In the initial project phase, you sit down and think ‘what could this project potentially mean to our business?’. For example, increasing efficiency or reducing time. Attaching a value to this goal allows you to calculate your ROI. However, during this phase that value would be a guesstimation at best.

So don’t get too lost in the details. Focus on defining a goal, and the details will follow.

Case – Partena Professional

Are you ready to start your own data science project? Are you on the lookout for ways to grow your business and improve customer satisfaction? Get in touch and let’s explore the possibilities together.