Injecting human intelligence into AI workflows: How SanctifAI uses n8n to embed human intelligence into AI processes

n8n_image
Version of n8n used
Self-hosted
First workflow in
Two hours
Time saved
3x faster than Python or LangChain

Considering n8n Enterprise for your team?

Background

SanctifAI, a leader in Human-AI collaboration for over a decade, integrates Human Intelligence with AI agent workflows through its Human as a Tool (HaaT) and Human as a Model (HaaM) products. Its industry-leading Work Distribution Platform scales task distribution to both internal and external teams. SanctifAI's offerings also include Data Services and Managed Agent Services, providing companies with the resources to build efficient, AI-driven workflows.

Challenge

SanctifAI's Work Distribution Platform is made up of two primary components: a Customer Portal, where customers create unique workflows and assign workforces to specific tasks; and a Worker Portal, which vendors use to access and manage assigned tasks for human workers.

With a network of more than 400 workforce providers, SanctifAI needed to find a way of easily enabling human workers to complete tasks as part of AI workflows – whether obtaining sign-off on financial documentation, or providing a second opinion on medical diagnoses. The solution needed to be horizontally scalable, distributed, and highly composable. The team didn’t want to build from scratch, or to have to maintain and update an in-house codebase.

Solution

SanctifAI evaluated a number of LangChain tools that could cover their needs. However, they realized they needed greater flexibility to incorporate logic that was not novel to the LangChain framework.

Instead, they turned to n8n, which they used to configure HaaM and HaaT in the Customer Portal by selecting the appropriate credentials and workflowID within the node configuration. Tasks for human workers are then published to the Worker Portal and assigned to the appropriate workforce. Once completed, the response is streamed back to n8n.

“At first we asked, ‘are we going to rebuild Zapier?’,” Tyler Thompson, Head of Product at SanctifAI said. “That’s what led us to n8n. This is really one of the first low code tools that we’ve ever used and we’ve not encountered any problem that we haven’t been able to solve with n8n. Plus we’re having so much fun with it. It’s like playing a game, but we’re actually building powerful tools for our customers.”

SanctifAI spun up its first n8n workflow in little over two hours, leveraging n8n’s out-of-the-box capabilities including the unique visual interface, queuing systems, routing systems, and schedulers – three times faster than writing Python controls for LangChain.

“n8n provided an immediate method to build and test agent and agent workflows for ourselves and our customers,” Nathaniel Gates, Founder and CEO of SanctifAI said.

“We embraced this 'low code' method of programming and saw dramatic efficiencies in both prototyping and production workloads. The business logic is set out in a framework that we can interact with easily. We can literally change nodes without needing to change any values.”

Impact

Today, n8n is at the core of SanctifAI’s products, enabling access to 400 diverse and highly specialized workforces through the SanctifAI platform – matching customer needs to specific human expertise for simple compliance tasks through to highly specialized skills. n8n’s visual interface has also fundamentally changed the way that the SanctifAI team is able to build solutions, by bringing stakeholders into the product development process.

“SanctifAI product development is no longer constrained by the scarcity of engineering talent or budgets, instead we now train our product managers to build and test directly. This has opened up a whole different paradigm – those with the vision for the product can be involved with the development. And when we demonstrate SanctifAI, we do it on n8n because no matter your technical background, everybody can understand the visual workflows,” Nathaniel said.

Next steps

SanctifAI’s front-end is currently running on a Render server, but the team is ready to stand up its workflows in AWS via Docker Containers. Meanwhile, the SanctifAI team are now looking at ways to contribute back to n8n, publishing a SanctifAI node that the community can leverage.

"n8n provided an immediate method to build and test agent and agent workflows for ourselves and our customers"

Nathaniel Gates
Nathaniel Gates
CEO