Book a call

Delivered

Case study · Claims software for P&C insurance

From a simple chatbot to a deep agent for Adjusted

Adjusted is a claims management platform for property and casualty insurance professionals. Central Dev rebuilt its chatbot into a deep agent, integrated with the backend, that takes actions on the user's behalf, and delivered the CRM, web application and mobile application.

adjusted.com (formerly deltaclaims.ai)

Key facts

Client
Adjusted
Industry
Claims software for P&C insurance
Scope of work
AI agent · CRM · Web app · Mobile app
Agent actions
On the user's behalf
Status
Delivered
Timeline
4 months

Challenge

Adjusted's platform included a chatbot that answered basic questions about the product. It could explain the platform, but it could not act inside it on the user's behalf.

What the rebuild had to achieve

Turn a question-answering chatbot into an agent that understands Adjusted's system, works through its backend and takes actions on the user's behalf.

Approach

We rebuilt the chatbot as a deep agent instead of extending the existing one. A deep agent is an AI agent that understands the system it works in, operates through its backend and takes actions for the user. The rebuild rested on three requirements.

  1. 01

    Understand the system

    The agent works with an understanding of how the Adjusted platform operates.

  2. 02

    Integrate with the backend

    Full integration with Adjusted's backend, so the agent works with the same system as its users.

  3. 03

    Act on the user's behalf

    The agent completes actions inside the platform for the user, beyond answering questions.

Solution

A deep agent inside the Adjusted platform that understands the system, is fully integrated with the backend and takes actions on the user's behalf. Alongside the agent, we delivered the CRM, the web application and the mobile application.

Before and after: chatbot to deep agentBefore: a user asks a chatbot and receives a basic answer. After: the user works with a deep agent that overlaps with Adjusted's backend, understands the system and takes actions on the user's behalf.BEFOREAFTERUserChatbotAnswers basicquestionsUserDEEP AGENTADJUSTED BACKENDUnderstandsthe systemActionActs forthe user
FIG. 1  A conceptual view of the change: the agent overlaps with the backend, so it understands the system and can act on the user's behalf.

What we delivered for Adjusted

  • AI agent

    The deep agent described above

  • CRM

  • Web application

  • Mobile application

Outcome

Adjusted's users moved from a chatbot that answered basic questions to an agent that understands the system and acts on their behalf.

Before

A chatbot that answered basic questions about the product.

After

A deep agent, fully integrated with the backend, that takes actions on the user's behalf.

Need your assistant to act, not only answer?

Confidentiality

We sign an NDA and talk through how your data is handled in the first conversation.

Book a call
Book a call

Start a project

Let's discuss your project

Book a 30-minute call or write to us. Scope, estimate and timeline are agreed with you after the first conversation.

Book a call Email us

Introductory call: we discuss your goals, constraints and existing systems, and the engagement model that fits.

09:00–19:00 Moldova time