Custom D turns agricultural research into reliable answers with trusted AI on Amazon Bedrock

Agricultural research organizations, industry bodies, and advisory organizations manage valuable knowledge across reports, websites, and technical documentation, but users often struggle to access trusted information when they need it.

To address this challenge, Custom D built Caitlyn on Amazon Web Services (AWS), turning approved organizational knowledge into trusted AI experiences and workflows. Built on Amazon Bedrock, Caitlyn supports security, data residency, governance, and flexibility across foundation models.

Where source content is prepared, Custom D can deploy the platform in as little as two weeks, helping organizations make decades of research easier to access, verify, and apply.

Agricultural research organizations, industry bodies, and advisory organizations often manage large volumes of valuable knowledge. However, that information is frequently spread across reports, technical publications, websites, and research repositories, making it difficult for people to access relevant information and apply it effectively in real-world decisions.

Traditional search can help users locate documents, but it still requires them to review technical content, compare sources, and determine which information is most relevant to their situation. As organizations looked to improve access to knowledge using generative AI, they also needed confidence that responses would remain accurate, explainable, and grounded in trusted source material.

Custom D saw the same need emerging across multiple knowledge-rich organizations. "While the content and audiences differ, the goal is the same: helping people get more value from trusted knowledge," says Matthew Rhodes, head of marketing, Custom D.

In response, Custom D set out to build a platform that could make organizational knowledge easier to access, understand, and apply while maintaining the governance and transparency these organizations required. 

Custom D, an AWS AI Competency Partner, built Caitlyn on AWS for organizations that manage large volumes of research, technical documentation, and institutional knowledge. The platform can be embedded into products, websites, and internal systems, helping users access trusted answers and supporting intelligent workflows without compromising accuracy, governance, or control.

Caitlyn transforms approved content into a governed knowledge base that users can query in plain language. Responses are grounded in trusted source material and include citations, allowing users to understand where the information originated and review the underlying sources. Organizations can also use the same foundation to build role-specific assistants and AI-powered workflows for different audiences and use cases.

Amazon Bedrock is central to the platform's architecture. By building on Amazon Bedrock, Custom D can access multiple foundation models through a single service and avoid being tied to a single model provider. This gives the team the flexibility to select the most appropriate model for different tasks as customer requirements and AI capabilities continue to evolve.

Custom D uses Amazon Bedrock alongside Amazon OpenSearch Service to retrieve relevant knowledge, Amazon Simple Storage Service (Amazon S3) to store customer content, and Amazon QuickSight to provide insights into user engagement and knowledge gaps. The company is also exploring Amazon Bedrock AgentCore capabilities to support multi-step workflows and agentic business processes.

Beyond the technology, AWS has supported Custom D through technical guidance, go-to-market collaboration, and partner enablement. "AWS proactively reaches out with suggestions and new services we could try. It feels like they are part of our team too, and we are all working toward a common goal for our customers," says Josh Smith, growth and partnerships, Caitlyn AI. 

By building Caitlyn as a reusable platform on AWS, Custom D can rapidly deliver trusted AI capabilities across industries while allowing customers to configure experiences for their own users and workflows. Where source content is prepared, Custom D can deploy Caitlyn in as little as two weeks because the core AWS architecture is already in place, allowing the team to configure each deployment rather than rebuild the platform from the ground up.

"For many parts of how Caitlyn operates under the hood, we will swap out and use different models for different actions and tasks," says Smith.

Caitlyn is already being used by organizations to improve how specialist knowledge is accessed and applied. For the Foundation for Arable Research in New Zealand, it powers Ask FAR, an AI-powered research assistant that helps members find answers 90 percent faster by making decades of scientific research easier to access. Beef + Lamb New Zealand has applied the platform to Bella, its AI-powered advisory assistant, helping farmers make evidence-based decisions by providing access to nearly 2,800 trusted resources across 14 content formats and extending specialist knowledge to approximately 10,000 farmers across the country. In Australia, the platform supports emerging workflow and persona-based experiences for Forest Growers Research, helping the organization tailor guidance for different users and use cases.

Together, these deployments show how Custom D can adapt Caitlyn for different organizations while keeping the platform consistent at its core. “Caitlyn is designed for organizations that need AI to be accurate, secure, and grounded in their own knowledge. AWS helps us deliver that with the flexibility to scale,” says Rhodes.

Custom D is a New Zealand-based AWS AI Competency Partner and software development company that builds custom applications, workflow systems, and AI platforms for organizations with complex knowledge, data, and operational requirements.