By Volha Hurskaya
ELVTR, a live online learning platform, has announced the launch of Wingman, an AI-powered teaching assistant designed to augment human instructors in live cohort-based courses. Helmed by CEO and co-founder Roman Peskin – an AI expert who previously co-founded DealAngel (a TechStars-backed startup acquired by OneTwoTrip) – Wingman represents a significant innovation in the edtech space. Rather than replacing educators, this AI “co-pilot” works alongside them, handling routine tasks and providing real-time support to both teachers and students. The result is the ability to teach larger cohorts without sacrificing quality, marking a new chapter in human-centric online education.
Founders Blend Tech and Education Expertise
Wingman’s development draws on the combined expertise of ELVTR’s founders. Roman Peskin led the product development and AI architecture of Wingman, leveraging his background as a “product guy” with a track record in tech innovation. His co-founder Yevhen Feldman – also founder of Laba Group, a multinational online education company – ensured the solution addresses real market needs in live online learning. Feldman is recognized as one of the world’s notable authorities on cohort-based education; under his leadership Laba Group has expanded across Europe, operating in Poland, Hungary, the Czech Republic, Romania, Ukraine, and Turkey. This blend of tech and educational expertise led Wingman’s design.
“We built Wingman to amplify our instructors, not automate them out,” Peskin explains. “The goal is to use AI where it adds value – grading, feedback, analytics – so our experts can focus on what only humans can do best: inspire and mentor students.”
This philosophy stands in stark contrast to trends in some corners of edtech. Many companies have explored fully automated tutors or MOOC-style platforms that minimize instructor involvement, often with mixed results. Research indicates that cohort-based courses, which emphasize live instructors and peer interaction, yield better outcomes than self-paced massive open online courses (MOOCs). ELVTR has built its model around such live, expert-led courses and boasts high completion rates (over 70% compared to under 5% for MOOC) thanks to this methodology. Wingman is designed to further enhance this human-centric approach by offloading menial tasks to AI while preserving instructors’ central role in teaching.
Augmenting Instructors – Not Replacing Them
A key principle behind Wingman is that the AI serves as a “teaching assistant” in the most supportive sense. It frees educators from administrative burdens – such as grading and answering repetitive questions – so they can devote more attention to mentorship and high-value interactions. In practice, this means instructors spend less time on paperwork and more time engaging students in meaningful discussion and feedback.
Crucially, Wingman is positioned as the opposite of a “cheap AI substitute teacher.” Where some edtech solutions attempt to cut costs by removing humans from the equation, ELVTR’s approach doubles down on the irreplaceable value of human instructors. The AI is there to assist – acting as a 24/7 support system and information resource – so that one expert teacher can effectively reach many more students without diminishing the learning experience. This human+AI partnership is a unique combination in the sector, aiming to make education more scalable and more personalized at the same time.
Inside Wingman: Three Pillars of Innovation
To achieve these goals, Wingman is built around three core modules, each addressing a critical aspect of online cohort-based course delivery. Together, they form an integrated AI toolset that boosts both efficiency and educational quality:
- Auto-Grading & Feedback: Wingman’s Auto-Grading module goes beyond standard AI assessment tools by evaluating open-ended responses, including essays, projects, and portfolios. Unlike conventional AI auto-graders that rely on generic publicly available knowledge, Wingman bases its judgments on an internal rubric derived exclusively from the instructor's live lectures and unique course materials. This approach ensures the system's functionality is closely aligned with ELVTR’s mission of providing education grounded in the instructor’s proprietary real-life expertise rather than public domain knowledge.
- Cohort Health Analytics & Instructor Insights: This integrated module continuously monitors cohort engagement and performance, synthesizing signals from live discussions and lectures, student's questions, and homework to proactively identify confusion or disengagement—even before grades begin to slip. Beyond merely flagging these issues, Wingman offers real-time suggestions for adapting teaching strategies and generates customized learning paths for student clusters struggling with specific topics. Instructors retain full control over these adaptive insights, deciding how best to integrate them into their teaching flow to maintain personalized education at scale.
- Student Knowledge Base & Q&A: Wingman transforms the traditional notion of course Q&A tools by embedding a course-specific knowledge base directly into each syllabus. Unlike typical education-focused AI chatbots, this system remains tightly aligned with the content of instructor’s live lectures, coursework, and distinct instructional style. Responses are re-enforced to reflect the instructor’s expertise and voice, reinforcing course content and maintaining consistent instructional quality, thus avoiding pitfalls common in more general-purpose chatbots.
Scaling Up: Early Results and Impact on ELVTR’s Model
Though Wingman was only in beta launch during 2024, ELVTR reports striking preliminary results. Traditionally, ELVTR would cap live cohort classes at around 40 students or else bring in human TAs to assist the instructor. With Wingman in the virtual classroom, a single instructor can confidently handle 120+ students – a threefold increase in capacity – because the AI is handling the repetitive queries, grading, and monitoring that extra TAs used to manage. This has immediate financial and operational implications. Hiring more staff to handle growing class sizes isn’t always feasible… AI can manage support at a fraction of the cost of additional personnel. By replacing the need for human teaching assistants in large classes, Wingman dramatically improves unit economics. The cost per student drops, and instructors can reach more learners at once, which in turn boosts revenue potential per course offering.
Just as importantly, these efficiencies come without eroding educational quality – in fact, the quality has improved. Students in Wingman-assisted courses receive more feedback and support than in a conventional setup; no raised hand goes unanswered, even in a crowd of one hundred, because Wingman is always there to help. And with instructors freed to concentrate on content delivery and mentorship, their teaching effectiveness goes up. ELVTR’s early pilot experience aligns with this: internal metrics from the Wingman pilot indicate higher assignment completion rates and course satisfaction scores compared to previous cohorts. With more students engaged and fewer falling through the cracks, the overall cohort performance rises.
From a business perspective, Wingman’s impact on the bottom line is twofold. First, by enabling larger class sizes and better student retention, it grows top-line revenue and ensures more students successfully complete (and potentially return for additional courses). Second, by reducing personnel costs (eliminating a need for human TAs) and automating time-consuming processes, it trims the expense side. This combination of higher scalability and improved margins strengthens ELVTR’s position in the competitive edtech market. It’s a timely boost as the company eyes expansion and potential strategic partnerships. Industry analysts note that edtech startups with proprietary AI innovations have become attractive acquisition targets, as larger education companies seek to modernize their offerings – Wingman clearly puts ELVTR in that innovative league.
Looking Ahead: Expansion to New Markets
As of this article's publication in October 2024, Wingman has long moved beyond its alpha release, which launched in Q4 2023. Currently in a stable beta phase, it is already delivering promising results in select U.S.-based courses, from data science to video gaming. Now, Wingman is preparing for a full rollout in its General Availability release in the near future. The feedback from these trials is being used to further refine the AI. ELVTR has confirmed plans to roll out Wingman to its broader course catalog in 2025, including an expansion to serve the UK and European student base. (It’s fitting, given Laba Group’s European roots, that UK cohorts will soon benefit from the technology.)
What will remain unchanged is the core philosophy: human-led education with AI in a supporting role. Wingman’s early success suggests that the right balance of AI and human expertise can overcome one of online education’s toughest challenges – scaling up instruction without losing the personal, high-touch experience that defines quality learning. By augmenting teachers rather than sidelining them, ELVTR is making a case that technology in education is most powerful when it amplifies the human element. It’s a vision of the future where an “AI wingman” at every instructor’s side could make elite, interactive education accessible to far more people, without driving up costs. In the words of Roman Peskin: “We’re fierce proponents of human-led live education. Wingman just gives each of our instructors an extra pair of hands – or rather, an extra brain, and a very smart one – to ensure every student gets the attention they deserve.”
