Imagine you are excited to jump the train of artificial intelligence (A.I.) and data science. You read about it a lot, all your friends are doing it and your company has set up a solid data base structure over the years. You are wondering, if you could use all the data you created in the past or the data you create now, to make your business processes faster, better or cheaper. Overall who doesn’t want more profit, happier customers and more satisfied employees?
Sadly your company does not have any employees who have done these kinds of A.I. projects in the past. If you are honest to yourself, you might not even know, what kind of skills a person or a team would need to successfully set up, run and finish such a project? Surely, you would need a data scientist for a data science projects, but what about a data engineer? And do you really know the difference between MLOps and DevOps? And overall, would an A.I. or data science project be handled like a normal waterfall project or rather like a software development project using SCRUM? Or is a mix of both? Would your inhouse project manager be able to run these kind of projects or run away from them after a month?
Questions arise and you feel that you might need outside help. So you start searching the web or your LinkedIn contacts to find companies, who would do these kind of projects for you. Since you have never done any of these projects before, you really don’t know which criteria to use to do an accurate vendor selection. If we are honest to ourselves, reasons for choosing an outsourcing partner might include
Website looks nice and clean
- Top search result on Google
- Came back to me quickly, was friendly on the phone
- Speaks my language or at least fluent English
- Big-name references on their website
- Referred to by another colleague
All these reasons might be valid, but do they truly include the very specific needs of your company? Are these reasons valid predictors for a successful project in your company?
As you see, there are many unknown variables which one cannot answer at first, but at the same time, these variables determine the later success or failure of your A.I. project. As for all new things, your personal risk for reputation damage is high, if things might go wrong.
In the next section we will dive deeper into the problems which can occur, when running outsourced projects and in part three we will look into solutions, how we can solve these problems and use the amazing opportunities outsourcing offers us today.