How GMB saves 50% time with Altura AI

We spoke to Marloes van der Liende, Tender Coordinator at GMB about their process with Altura AI. They have been using it since August 2023.

The process before Altura AI

GMB, a company specialised in water safety, water quality, bio-energy and industrial construction & infrastructure, faced a challenge with their old bid management process. We spoke to Marloes van der Liende, Tender Coordinator of GMB about their experience with Altura.

GMB used Sharepoint to store old proposals. The bid management team had these 4 challenges when searching for texts and information: 

CTRL+F did not suffice

‍Searching for the correct information from previous tenders in their own database and manually opening various folders took a lot of time. Also, GMB could only search 1 document at a time with the CTRL+F function.

Difficulty finding information

Because GMB could only search for exact words and not related words, information was difficult to find. Especially because colleagues used different terms. An example is 'connecting', where others would use 'collaborating'. 

Manual tasks

Uploading plans and assigning tags or metadata to texts and old proposals was labour intensive. It was often forgotten or postponed because it was too much work. If texts or plans were not registered properly, information could not be found. And a lot of information was lost.

Unnecessarily disturbing colleagues

“Where is information X located?”

The tender writers were constantly being interrupted.

"When colleagues incorrectly entered texts or old proposals into our knowledge base or forgot to add metadata, it was impossible to find text based on keywords, which was very frustrating."

- Marloes van der Liende, Tendercoördinator  GMB

‍Solution: Alturas AI bid content library

GMB chose Altura AI to solve this. The solution for them was the ProposalSearch tool - a bid content library where you can quickly and easily find high-quality text pieces or information from old proposals. You simply filter on performance or search for semantically matching search terms.

Results with Altura AI

GMB has been working with Altura AI and the bid content library for 2 months now. The results are clear:

Improved discoverability, saving time

The team no longer has to guess which word other colleagues had used. They can now intuitively search for the right piece in different documents because AI matches the information that belongs together. As a result, GMB saves 50% of their time searching for texts or old proposals.

Forgotten texts are now used again

Because GMB had to search for relevant information from memory, useful texts from years ago were forgotten. Text from a proposal from 4 years ago was successfully used for their most recent registration. 

Documentation is less prone to errors‍

Because they no longer have to add tags or metadata to texts or documents, the GMB team no longer makes mistakes and nothing is lost.

Automatically generate texts

Because the team can now ask specific questions in their bid content library, text is generated to answer specific questions. They can immediately communicate these answers between colleagues for further input. 

Less busy writing texts over and over again

Before using Altura AI, tender writers often needed input from technical colleagues about, for example, their motivations and team roles. Because they are often indifferent about providing this, it was regularly postponed. 

Now the bid content library can generate the piece of text so they no longer have to wait for input. This piece is sent to just be reviewed by these colleagues, saving them 25% of their time. 

Because technicians now only have to focus on the technical plan, which they like more and are better at, GMB uses them more effectively within the tender.

Marloes says the following about this:

"Findability has greatly improved: we can now search based on what we think, without using exact words from the do

Altura AI's Proposal Search saves GMB time significantly and helps them write better text proposals.