AI-supported search feature improves search results and user experience
With a new, AI-based search feature, the DTAD platform can make public tenders more efficient and easier to find, giving it a decisive competitive advantage.
DTAD GmbH operates a digital tool with which more than 5,500 active customers find new orders from the public and commercial sectors every day. The platform makes it easier for companies to access public tenders and qualified B2B leads, and helps them to optimally manage their leads in the acquisition process.
The project at a glance
- Semantic search with AI improves the relevance of search results
- Modern search experience that users expect today
- Scalable cloud architecture ensures the solution is future-proof
- Iterative development with close coordination between product management, design and development
- Sustainable development of expertise and increased ownership within the team
Background
DTAD GmbH operates a platform that collects and curates public tenders and construction projects and makes them available to its customers. While the platform has already impressed with its comprehensive collection of tender data, new digital developments are opening up exciting opportunities to make the service even more efficient.
In order to provide customers not only with extensive data, but also real added value through more targeted and relevant search results, DTAD has aimed at further improving the quality of the search feature. Their vision: an AI-supported, semantic search that understands the context and meaning behind the search queries and thus delivers even more precise results.
As such technology is not available as a ready-made solution, it was developed in close collaboration between codecentric and DTAD's product management, development and designer. The aim was to sustainably improve the user experience and to further differentiate itself from the competition in terms of technology.
Solution
As part of the project, the aim was to create a significantly more powerful search function, as known from modern search engines and AI-supported recommendation systems. This was implemented in an intuitive user interface. The web application, developed with Vue.js, follows current UX standards and offers a search experience that fits seamlessly into the expectations of today's users. The results not only appear faster, but are also significantly more relevant and easier to understand thanks to semantic processing.
The technical implementation is based on a scalable cloud architecture:
AI-supported seach
Data management with MongoDB Atlas
AWS cloud infrastructure
User-centric frontend with Vue.js
AI-supported seach
AI-supported search: OpenAI embeddings enable a semantic search that recognizes relevant results even with varying word choices or synonyms. Terms are not only compared literally, but their contextual relationship is also taken into account - similar to the way people derive meaning from a text.
Data management with MongoDB Atlas
Data management with MongoDB Atlas: In addition to the high-performance storage and retrieval of structured and unstructured data, a decisive factor was that MongoDB Atlas is able to process embeddings efficiently. Several database solutions were evaluated as part of the implementation. MongoDB Atlas proved to be the best and most cost-effective choice for the specific use case.
AWS cloud infrastructure
AWS cloud infrastructure: The search results are processed and provided by AWS Lambda, which ensures high scalability and cost efficiency.
User-centric frontend with Vue.js
User-centric frontend with Vue.js: The application offers a modern, intuitive user experience for a seamless search experience.
The entire project was implemented in an interdisciplinary team - consisting of up to eight developers, a product owner and a designer at times. Modern agile methods were used to react flexibly to new findings and user feedback.
A user-centered approach was pursued from the beginning to enable fast and targeted integration of the new search feature. Instead of rolling out the search only after development was complete, it was integrated into the feedback process at an early stage.
The new search feature was already made available to certain stakeholders in the first iteration with the help of feature toggles. This allowed real user feedback to be obtained at an early stage, which provided continuous improvement approaches. The iterative approach ensured that the new search technology was not only technically mature, but also optimally tailored to the needs of end customers.
Result
The new search feature clearly stands out from traditional full-text or keyword searches. Users no longer need to know exactly the right terms to get relevant results. Instead, they can simply enter a coherent text - for example a newspaper article about a construction project - and receive matching tenders and similar projects based on this.
By using the new search technology, DTAD GmbH has successfully set itself apart from the competition. Instead of just relying on masses of data, the AI-supported search provides real added value for customers by delivering more targeted and relevant search results.
In addition to the successful introduction of the new search feature, a key success of the project was that DTAD's development team increasingly took ownership of the product. This applies not only to technological development, but also to the product strategy and the continuous improvement process.
The iterative development process, the close collaboration between product, design and development and the build-up of expertise within the team have ensured the sustainable further development of the product. The team is now actively shaping the future of the platform and has assumed responsibility for continuous improvement and adaptation to new requirements.
Outlook
The semantic search forms the basis for further optimizations and enhancements. Potential next steps could be:
- Personalized search results that are even more tailored to the individual needs of users
- Expansion to additional data sources to further increase the coverage and quality of search results
- Fine-tuning the AI model to deliver even more precise results
With this project, DTAD has not only introduced a technological innovation, but also taken a decisive step towards a more intelligent and future-proof platform.
Kai Noffke
Director of Technology, DTAD GmbH
Together with our developers and the experts at codecentric, we have implemented a modern, AI-supported search in a very short time - a real milestone that significantly increases the added value of the DTAD platform for our customers and strengthens our market position in the long term.
Kai Noffke
Director of Technology, DTAD GmbH
Any questions about the project?
Would you like to know more about the project? Are you interested in a similar solution for your company?
Further reference projects
Find out about other successful projects that we have completed with our clients. Perhaps you will find ideas here for a use case in your own organization.
David Gontrum