100's of workflows

Build a Tax and Financial Assistant with AI

Integrations
HTTP Request
Edit Fields (Set)
Loop Over Items (Split in Batches)
Switch
Compression
Wait
Sticky Note
Execute Workflow Trigger
Manual Trigger
Filter
AI Agent
OpenAI Chat Model
Window Buffer Memory (easiest)
Recursive Character Text Splitter
Call rantir Workflow Tool
Extract from File
Split Out
Default Data Loader
Embeddings Mistral Cloud
Chat Trigger
Qdrant Vector Store

This rantir workflow builds another example of creating a knowledgebase assistant, but demonstrates how a more deliberate and targeted approach to ingesting the data can produce much better results for your chatbot.

In this example, a government tax code policy document is used. While we could split the document into chunks by content length, we often lose the context of chapters and sections, which may be required by the user.

Our approach is to first split the document into chapters and sections before importing it into our vector store. Additionally, using metadata correctly is key to allow filtering and scoped queries.

Example

Human: "Tell me about what the tax code says about cargo for international commerce?"

AI: "Section 11.25 of the Texas Property Tax Code pertains to 'MARINE CARGO CONTAINERS USED EXCLUSIVELY IN INTERNATIONAL COMMERCE.' In this section, a person who is a citizen of a foreign country or an entity..."

How it works

     
  • The tax code policy document is downloaded as a zip file from the government website, and its pages are extracted as separate chapters.
  •  
  • Each chapter is then parsed and split into its sections using data manipulation expressions.
  •  
  • Each section is inserted into our Qdrant vector store, tagged with its source, chapter, and section numbers as metadata.
  •  
  • When our AI Agent needs to retrieve data from our vector store, we use a custom workflow tool to perform the query to Qdrant.
  •  
  • Because we're relying on Qdrant's advanced filtering capabilities, we perform the search using the Qdrant API rather than the Qdrant node.
  •  
  • When the AI Agent needs to pull full wording or extracts, we can use Qdrant's scroll API and metadata filtering to do so. This makes Qdrant behave like a key-value store for our document.

Requirements

     
  • A Qdrant instance is required for the vector store and specifically for its filtering functionality.
  •  
  • Mistral.ai account for Embeddings and AI models.

Customising this workflow

Depending on your use case, consider returning actual PDF pages (or links) to the user for extra confirmation and to build trust.

Not using Mistral? You can replace it, but make sure to match the distance and dimension size of the Qdrant collection to your chosen embedding model.

Other Workflows like this one

Your connected stack awaits to automate AI workflows with 24-7 uptime performance and engagement

Voice Activated Multi-Agent Demo for Vagent.io using Notion and Google Calendar
AI Agent
OpenAI Chat Model
Window Buffer Memory (easiest)
Call rantir Workflow Tool
HTTP Request Tool
Automated Agentic News Monitoring with perplexity.ai
Gmail
AI Agent
Basic LLM Chain
OpenAI Chat Model
Markdown
Building a WhatsApp Chatbot
HTTP Request
Switch
Sticky Note
WhatsApp Business Cloud
Manual Trigger
AI: Ask questions about any data source (using the Rantir workflow retriever)
Question and Answer Chain
OpenAI Chat Model
Workflow Retriever
Edit Fields (Set)
ETL pipeline for text processing
Postgres
Slack
MongoDB
X (Formerly Twitter)
Google Cloud Natural Language
Building Chatbot for Movie Recommendations
GitHub
HTTP Request
Merge
Edit Fields (Set)
Sticky Note

Compare features across plans

Computir Cloud Suite All Access

$99/m

Per team/per month, with 10 GB of data and storage
Everything in Free, and:
Icon
Host up to around 4-5 Applications
Icon
Advanced user roles
Icon
Unlimited AI applications & workflows
Icon
Custom onboarding & Customer management
Icon
Advanced integrations
Icon
International capabilities
Unlimited Team Plan & Custom Integration

$299/m

Per $1K Tokens or 1 TB added, custom integration (per month)
Everything in Professional, and:
Icon
Host up to around 20+ Applications
Icon
Tailored implementation services
Icon
Advanced ERP integration capabilities
Icon
Extra bandwidth and open-source AI models
Icon
Fine-tuning & data logic
Icon
SOX or integration customization
Icon
Dedicated premium support
Cloud Suite

$99/mo

Team Plan

$299

Computir Cloud

AI Application & Automation platform suite
Get access to generate dashboards, websites or content
Chat to Explore Data
Icon

Custom Develop  integrations

Chat to Transform Data
Icon
Direct or Enterprise application connections
Webflow, Wix or Wordpress
+ Acumatica, Microsoft, Netsuite & Sage
+ Oracle & Workday
Rules to automate AI
Basic
Advanced
Advanced

Custom Integrations

Build & Share Live Reports
Icon
Generated
Human-Led
Train Classification Models
Icon
Human-Led
Train Time Series Forecasts
Icon

"I highly recommend Computir, they are a great dev team with quick turn around on all projects and requests. We recently worked with them on updating our website and any changes, updates or modifications I needed were always taken care of quickly!"

Paige J, VP of Marketing, Heavy AI