VectorShift Platform Overview
An overview of the VectorShift Platform
Last updated
An overview of the VectorShift Platform
Last updated
VectorShift's platform allows you to design, prototype, build, deploy, and manage generative AI workflows and automation across two interfaces: no-code and code SDK. Example use cases:
Chatbots: Build a fully functional AI chatbot and embed it into your website. AI chatbots can respond to user queries and answer questions based on a knowledge base (e.g., product documentation, support articles, live-synced Google Drive).
Search: Centralize and live-sync your data across your applications (e.g., Google Drive, Slack, One Drive, etc.) in one knowledge base and perform search queries over all your data.
Automations: Automate workflows end-to-end, schedule workflows to run at certain time intervals, or allow certain triggers (e.g., email, Slack message) to trigger the running of a workflow; have the VectorShift platform automatically perform actions (e.g., send an email, create a notion document, add data to an Airtable database).
Content Creation: Create marketing copy, personalized outbound emails, call summaries, proposals, and graphics at scale and in a predetermined format and style.
Workflow Automation: Replicate the logical thinking of a process by breaking down a task into sub-tasks and having different LLMs perform them.
The VectorShift platform's core is a pipeline, which is an AI workflow. On the "Pipelines page," click "New" at the top left to create a new pipeline.
This will bring up a modal that either allows you to "Create Pipeline from Scratch" or to start from a template. You can filter the different template categories with the buttons on the left.
For example, you can choose the "Customer Support Chatbot template" to build a workflow that automates customer support requests.
The no-code builder is comprised of three components:
Nodes: Nodes are found on the top bar and contain the tabs named "General", "LLMs", "Knowledge Base", etc. Nodes are components of AI workflows that perform various functions, such as fetching data, performing actions, and leveraging LLMs.
Canvas: The canvas is where nodes can be connected to form a workflow.
Control bar: The control bar is on the top right. These allow you to "Deploy" a workflow to production, "Deploy as" your pipeline as a chatbot, automation, or search, "Run" the pipeline, or Name the workflow / share it with teammates (under "Details").
You can add relevant data to the "Knowledge Base Reader" node to operationalize this customer support chatbot pipeline. This node allows you to centralize your data wherever it sits (e.g., files, Google Drive, Notion, etc.) to perform a semantic search (e.g., to find the most relevant pieces based on a query). You can do this by selecting an existing knowledge base through the drop-down menu or creating a new one by clicking "Create New Knowledge Base."
You can create a "Knowledge" that centralizes and live-syncs to your data (e.g., Google Drive, OneDrive, Notion, Airtable). You can then use knowledge bases within pipelines to perform semantic searches and return relevant information that LLMs can use.
Create a Knowledge Base by clicking " New " in the "Storage" tab.
You can then "Add Documents" to the knowledge base, including files, data from "Integrations" (e.g., live sync to data within Google Drive, Airtable, OneDrive), URLs, and more. Using the "Knowledge Base Reader" node within the no-code builder will return the most relevant pieces across the knowledge base.
Pipelines can be deployed as Chatbots by clicking on "Deploy as" within the No-code Builder.
Once created, you can customize the functionality (e.g., suggested initial prompts, Welcome Message) and UI (e.g., color).
You can export the chatbot via URL (with a ChatGPT-inspired interface) and embed it directly into your website via an iFrame, as a WhatsApp / SMS bot, as an API endpoint, or as a Slack App bot.
Finally, you can also Manage the chatbot:
View recorded conversations,
See Analytics (number of messages, AI cost, user feedback, etc.), and
Download the conversation history.
Pipelines can be deployed as Automations by clicking on "Deploy" within the No-code Builder.
Automations allow you to run pipelines automatically. Here, you can schedule workflows to run at certain time intervals (Cron) or allow defined triggers (e.g., email, Slack message, Typeform message) to trigger the running of a workflow.
