Over a hundred colleges and high schools are turning to a new AI tool called Nectir, allowing teachers to create a personalized learning partner that’s trained on their syllabi, textbooks, and assignments to help students with anything from questions related to their coursework to essay writing assistance and even future career guidance.
The company announced Thursday its $4 million seed funding round, bringing the total amount raised to $6.3 million. The new capital will be used to develop new features and expand the team.
With the ongoing teacher shortage, it’s becoming harder for educators to provide individualized, one-on-one feedback to every student. Kavitta Ghai, the mind behind Nectir, witnessed this issue while studying at the University of California, Santa Barbara (UCSB).
“Everyone that I was talking to could not see the value of a college education anymore, and we’re seeing that across the nation right now,” Ghai told TechCrunch. “The education that we’re providing is not valuable. It’s not modern. It’s not setting them up for career success.”
Additionally, Ghai has her own learning challenges. “I am autistic, and I have ADHD, so no classroom that I’ve ever been in before has felt comfortable or safe for me or my brain to be in,” she told us.
Ghai and co-founder Jordan Long (CTO) decided to create Nectir—a 24/7 chatbot that provides constant support and assistance to students at any time.
With Nectir, teachers can create an AI assistant tailored to their specific needs, whether for a single class, a department, or the entire campus. There are various personalization options available, enabling teachers to establish clear boundaries for the AI’s interactions, such as programming the assistant to assist only with certain subjects or responding in a way that aligns with their teaching style.
In a demo with TechCrunch, Ghai used an example of a Calculus AB course assistant that helps students with study tips, assignments, and course-related questions. The AI was specifically trained to interact with students by providing support rather than completing assignments for them. It was also instructed to use the Socratic method, which involves asking a series of questions to guide students toward finding the final answer independently.
What we found most noteworthy about Nectir is the ability to limit where the AI collects information from. Users can define the level of access the assistant has to the language model (LLM), which in this case is OpenAI’s GPT-4o.
Nectir’s “General Knowledge” option gives the assistant complete access to ChatGPT, enabling it to retrieve broad information. Meanwhile, the “Topic Knowledge” option restricts the AI to pulling only the most relevant information based on the specific topics that users have uploaded.
A “Document Only” option limits the AI to extracting information solely from the data provided by users. It will not access any information from ChatGPT. This is particularly beneficial for educators creating something like a financial aid assistant as it should improve the accuracy and relevance of the information since ChatGPT has been known to hallucinate answers.
“Our proprietary system that we’ve built on the back end is not just using [retrieval augmented generation]. We’re having it pull from the documents that the teacher has uploaded first, and then if the AI needs to, it can pull from the LLM. But that’s how we’ve improved the hallucination to not allow it to access the LLM first. It’s going to go to that much smaller pool of data from the teacher,” Ghai said. She claims that Nectir has over a 95% accuracy rate.
Nectir has a disclosure at the bottom of the chat that warns the AI may make mistakes. Users can also rate the responses at the bottom to inform the platform whether the questions were answered correctly.
While implementing generative AI technology in educational settings has raised concerns, especially given the widespread use of tools like ChatGPT for cheating, educators are increasingly recognizing the need to understand how the tech is evolving.
Notably, Nectir allows teachers to monitor interactions to see exactly how their students are using the tool and make sure the AI is providing support without simply giving away answers. An analytics dashboard also lets teachers see insights like which students are using it most, what hours they’re using it, and so on.
In addition to concerns about cheating, many AI tools currently available aren’t suitable for use in schools as they don’t comply with privacy laws, such as the Family Educational Rights and Privacy Act (FERPA), which protects personal information found in education records. Nectir asserts that its AI protects student privacy, claiming compliance with FERPA standards.
“There was nothing out there that allowed them to utilize any LLM in a FERPA-compliant way…[With Nectir,] none of this [student] information is sent back to the LLM. We use our own private endpoint using Azure, and so we’ve built this specifically for educational institutions to be able to use,” Ghai said.
The company’s recent round was led by Long Journey Ventures with participation from Behind Genius Ventures, Entrada Ventures, and Precursor Ventures. The company previously raised a pre-seed $2.3 million.
Nectir is using the funding to build the second version of its product, which will include a new “sentiment analysis” feature. The tool will analyze the conversations that students have and provide insights into their strengths, weaknesses, and areas for improvement. Additionally, it will suggest classes for students to consider for the following semester and outline potential career paths. This feature is currently being tested in a few schools and will be widely rolled out in the coming months.
“It’ll really be that customized learning partner. Every single conversation that a student has with any of their assistants will then be fed into that student profile for them to be able to see based on what the AI thinks, what should I be doing next, not only in my educational journey, but in my career journey,” Ghai said.
Launched in May, Nectir is already active with over 80,000 students at more than 100 schools, including Boston University, Los Angeles Pacific University, Questrom School of Business, Stanford Graduate School of Business, and UCSB.