Scaling product development and reducing time to value for new clients
aifora was able to scale product development quickly and sustainably with the help of demand-based external support. It now takes significantly less time to onboard new customers (time to value) thanks to the automation of complex data- and machine-learning pipelines.
aifora provides a cloud-based SaaS solution that helps retail companies optimize prices and inventory across multiple channels and automate the underlying processes.
The project at a glance
- Reduced manual effort in the data science cycle
- Faster time-to-value for new customers
- Successful joint product development
- Flexible ramp-up and ramp-down to better meet demand
Initial situation
The start-up was looking for a partner to evaluate its existing tech stack and develop it in accordance with current best practices in order to secure a good rating in the upcoming series A round of financing. After a successful financing round, aifora was keen to scale up quickly and push ahead with automation to efficiently serve new customers.
aifora lacked the development resources to scale up, and these were not available in the current labor market at short notice. It was therefore planned to expand the existing development teams in the short term and to set up additional product development teams to create new features and new modules for the company's retail automation platform.
Furthermore, onboarding new customers, a process requiring extensive data integration, was challenging and increased the time to value, which is the time that passes before aifora's solution generates a measurable benefit for a customer. For this reason, the process needed to be further automated and enhanced.
Solution
Joint development team
Continuous knowledge sharing
Demand-based ramp-up/down & knowledge transfer
Automated data science processes
Joint development team
A joint development team from aifora and codecentric was initially set up for inventory optimization. codecentric provided the product owner and added expertise in data engineering, backend, frontend, agile coaching, and scrum to the team. The product development team was able to work freely and autonomously in direct and continuous consultation with end customers and users.
Continuous knowledge sharing
The mixed team had access to aifora's extensive domain and industry knowledge at all times, and with this expertise was used to refine the scope on a regular basis. This allowed the team to extend the feature set of existing modules in short feedback cycles.
The team also developed a new module for purchase quantity optimization for the aifora platform in collaboration with a pilot customer. The special challenge: the new module had to be developed to a tight schedule because the pilot customer needed to shut down its existing legacy system.
Demand-based ramp-up/down & knowledge transfer
aifora had access to codecentric's pool of experts during the entire collaboration. This allowed critical bottlenecks to be quickly resolved in key areas, such as data engineering. aifora succeeded in filling vacancies internally over time, with codecentric training and mentoring the new employees before gradually withdrawing from the development process.
Automated data science processes
A key challenge to achieving the scale aifora is aiming for is to reduce the manual effort required during the existing data science cycle. This includes, in particular, automating and scaling the processing of customer data. codecentric AG's data architects and data engineers helped design and implement solutions to automate complex data and machine learning pipelines, reducing the administrative and operational tasks involved in processing customer data on a daily basis. Integrated monitoring and alerting tools help the team maintain an overview.
Result
The highly automated data platform has reduced the time needed to onboard new customers. In particular, improvements to the data science components can now be implemented easily and iteratively. This means that for new customers, the perfect algorithm can be quickly found and a suitable model trained.
The pilot customer was able to switch off its existing legacy system on time and is now using the new aifora module to optimize its purchasing processes. What's more, the new module extends aifora's product portfolio and can be marketed to other customers.
The product development team for inventory optimization that was set up and has since become highly proficient continues to develop the modules independently.
David Krings
CTO
codecentric was particularly helpful in improving software development processes and demonstrating how software can be developed in short cycles in line with customer requirements. This support, together with technological improvements, has meant that we are now able to maintain our software better and also develop it faster ourselves.
David Krings
CTO
Any questions about the project?
Want to know more about the project? Are you interested in a similar solution for your company?
Service Lead Data & ML & AI
Other projects
Find out about further successful projects that we have completed with our customers. Perhaps you will find inspiration for a use case in your company.
Marcel Mikl
Service Lead Data & ML & AI