Daugherty Enhances Chatbot with AWS to Improve Knowledge Sharing of Bayer’s Data Assets

Every time an organization expands, it experiences compartmentalization — departments become efficient using their own processes and platforms, whereas other internal teams are not as familiar with individual departmental knowledge. As a result, some of the metadata (for example, architectural frameworks) are not well understood, nor is the meaning of certain terms.

During a merger and acquisition, an organization experiences even greater need to ensure the spread of departmental knowledge. Here is an example of a solution that supported Bayer’s acquisition of Monsanto in June 2018:

Bayer inherited Haystack, an internal metadata store, that had:

a glossary of nearly 500 terms
Nearly 1 million fields of metadata

Using the application, Bayer better understood frameworks shared between Monsanto departments.

But it raised an interesting question: What is the most efficient way to distribute this need-to-know information when everybody is proficient in a different platform?

We should be able to leverage the latest technology to consolidate and distribute information. Working with Daugherty, Bayer did exactly that.

Responsive image

Pairing with Bayer’s Data Platform Engineer lead, Daugherty helped provide cutting-edge technology for an innovative solution that exceeded Bayer’s expectations — an application named Alfred.

Alfred is built on Amazon Web Services (AWS) infrastructure to leverage Natural Language Processing (NLP) and Machine Learning requirements, and it grants access to the Bayer workforce to the metadata within.

Responsive image

What began as a simple Proof of Concept chatbot with minimal functionality has been enhanced by our team to respond to employee questions on multiple platforms. The chatbot sifts through low-level data and provides answers based on relevance. While most of the data is internal, Alfred can also retrieve external data like Wikipedia pages.

Alfred is now integrated with Slack, Amazon Alexa and Amazon Sumerian as an access point. Future work will see Alfred integrated with Yammer, Skype and Microsoft Teams. All of these will use NLP from AWS to ensure customers are getting the most relevant responses to their questions.


The best part? Because the Bayer workforce has widely adopted Slack, Alfred is the fastest way to access this information.

The second best part? Alfred makes documentation fun. Since its introduction, Bayer has seen a 400 percent increase of documentation stored within Haystack, and over 45,000 visits from 28 countries. Alfred has helped unify Bayer and foster collaboration among teams.

Throughout the process, Daugherty was able to strengthen our relationship with Bayer. We had previously partnered with Monsanto through many of the technological advancements in the past 30 years – from BlackBerries to IoT devices to cloud dashboards. Plus, we had a wealth of expertise around transforming organizations that had gone through mergers and acquisitions – leading them through pre- and post-merger decisions. Not only could we bring familiarity and expertise, we could tailor proven tools and techniques like Business Solutions and Advisory frameworks.

Consider this: If Bayer’s 12,000 employees ask a mere two questions a week, and each question takes 20 minutes to answer, this solution would save Bayer $37 million annually.

Now, Bayer’s glossary of terms and metadata can be accessed via a chatbot designed to fit whatever platform its teams are using. Not only will this save them millions, it will transform their teams to keep efficient processes in place, while enhancing their productivity.