At Prophecy Labs, we developed a phased approach that allows us to take a unique approach to data science projects. Our approach empowers us to keep an eye for only building things that add business value for our clients. Working in phases let’s us gradually build your optimal solution with success or fail fast without wasting precious time or money. This allows us to stay agile, provide a transparent approach, and to deliver the projects on time while meeting or exceeding the expectations. 

Three phases form the cornerstone of our project strategy: Explore, Build, and Automate. Each of the phases is separately scoped, billed and executed.

Before starting any project, we set up a Discovery Session with you as our prospective client. The goal of this session is to get to know each other and find out how we can help you with your business problems. This introduction allows us to get a clear grasp of your organization’s problem and brainstorm possible solutions together. If we decide to kick off a project together, we provide you with a clear step-by-step plan to bring tangible business value.

The first phase in our project cycle is the Explore phase. The goal is to test your use case’s viability by answering the question: “Is it worth investing time and money into a technological solution?”. This allows us to offer quick advice without wasting your precious time or money. The explore phase provides a minimal solution to your problem with a Proof of Concept (PoC). We compare different approaches without spending too much time, identify potential problems and find efficient solutions. We also test the software, hardware and infrastructure that is going to be used and make sure you gain essential experience, skills and confidence working with AI and ML. The PoC aims to convince your company’s business stakeholders to invest in the AI solution.

The second phase in our project cycle is the Build phase. The goal is to develop and deploy a fully functional Minimal Viable Product (MVP). This solution provides your organization with tangible business value in the form of actionable insights, forecasts, optimized processes or automated tasks. Our MVP satisfies all your project requirements with a focus on stability, testing, good code quality, proper infrastructure and scalability if needed.

The third phase in our project cycle is the Automate phase. The goal is to make our solution operate autonomously with minimal human supervision needed. We integrate the solution in the current infrastructure such that it is fully scalable and robust. This automated solution typically entails the end of the project.

After finishing our project, your solution of course needs to remain up and running over time. The world is constantly changing, so it is important to track the model quality and predictions over time. Maintenance is needed to evaluate and monitor the performance of the deployed model. This allows to react to feedback from deployed models and improve them further over time. Our sister company NannyML specializes in monitoring AI solutions.