The sheer volume of tender documentation made the process heavy
Bidprocesses for large construction projects are long and multi-stage. A single tender request contains dozens of documents and hundreds of pages that define the project’s requirements, objectives, schedules and evaluation criteria. Finding the essential information took experts a lot of time. Significant risks also need to be identified early, before more time goes into the bid.
AI supports the bid process step by step
We built an agent-based solution around three stages of the process. In the first stage, AI quickly forms an overall picture of a new tender request and surfaces the project’s requirements, objectives and risks. The team uses this information to decide whether to pursue the competition.
After the bid decision, AI supports building the bid itself and takes the requirements and objectives identified in the documentation into account when forming the content. At the end of the process, the solution supports quality assurance and checks that the bid meets the minimum requirements. It also simulates how the bid might be evaluated against the tender’s scoring criteria.
We divided the work among several AI agents
We built the whole from several AI agents in the Azure environment, each with its own task. A single function, such as risk analysis, emerges from the collaboration of several agents. The central source of information is the documentation supplied with the tender request. From it, the solution produces summaries and analyses for different purposes. Users can refine the perspective and iterate the response in the user interface.
Different bid formats and procurement types require different things to be considered, so we built the solution to process the documentation according to what each tender request requires.
The next step is using accumulated knowledge
In the future, the solution can be extended to use internal historical data as well as external sources. This will allow the knowledge built up over the years available to bid teams and help raise the quality of their bids.
Learn more about AI agents
