We are now in a stage that thanks to serverless applications and the shear amount of available cloud services, anything can be deployed in weeks. But even the most accurate system needs testing, more trusting than testing.

The way we see a team receiving a data science product is not as a complete and finished package, but as a service co-produced by developers and consultants alongside the client for the client. Which means: we develop and test, the client co-develops and tests. A joint force to produce a flawless operation.

Here, we outline four steps that we see necessary to introduce the newly developed system within the client’s own ecosystem.


The very first step towards seamless adoption is co-developing. During development, your innovation partner must be user-centric and bring all the stakeholders along in the process. As a result, putting the product through continuous performance testing and human checks within the client’s team will be much easier.


As soon as the first and simplest iteration of the new product is available, it’s time to test it alongside the existing client’s ecosystem. Start working in real time, as if it was already completely adopted, but without removing the current workflow. This parallel testing will allow to iron out any issues with the current state of the product, but more importantly will check its accuracy in the real world and get further feedback on the user experience.


A chain manager won’t trust something that decreases his/her operational costs by 70% immediately after so many years of experience in the role. You have to build the trust. As any start-up trying to gather clients, the product needs to be adopted by its clients: end corporate users.

Even if such levels of savings are feasible, you should approach this change in working methodology in a sustainable way for your team members. Progressively incorporate the product proving that these savings are feasible but more importantly: create a need rather than just trust on the new tool.


Our product is now wanted and trusted, the team loves it and they can’t get enough of it. It’s time to unleash the full potential of our tool and remove the limits to the predictions. The client is now ready to use our product in full, in full maturity. In addition, if built in blocks, maintenance and future development will be faster and safer.

These four steps will take slightly longer to deliver the desired outcomes. However, the product will be more reliable and used by your team in the long run. Building for longevity instead of quarter financial reports will pay off with better operations and happier and empowered teams.

Leave a reply