)
With up to 80 per cent of software features introduced to the market rarely or never touched by the end user, Melbourne startup Cor has just raised $2 million in pre-seed funding to further its AI coach 'Obi' that can guide new customers at any time of day or night.
Founded by ex-Googler Mantas Aleksiejevas and former Uptick CTO Luke Hodkinson, who both bootstrapped the company until this formal financing round, Cor uses generative video AI to create interactive agents that onboard users to new software in real time while watching the user's screen.
The round was led by Rampersand with participation from Archangel, Skalata and Black Sheep Capital, with the funds going towards scaling operations as the startup hires new staff, focusing on sales and growth investments.
"From Salesforce to early-stage startups, only enterprise customers and top-tier accounts get white-glove onboarding, the kind that helps users become truly productive, let alone power users," says Rampersand partner Andrew Poesaste.
"The reality is that 85 per cent of the long tail gets left behind, forced to figure out complex products on their own.
"Cor's AI agent Obi changes this equation entirely. It delivers personalised, high-quality onboarding at scale, meeting users where they are and giving them personalised support to master the product."
Poesaste says Cor's product has been developed at a time when AI is allowing teams to ship product changes at "maniacal pace", daily rather than quarterly.
"Obi closes this widening gap between what software can do and what users actually adopt, driving faster activation, stronger retention, and clearer ROI )return on investment) for software teams," he adds.
"The early results speak for themselves, and we believe Cor is building the category-defining platform for how companies onboard, enable, and grow their users throughout the entire customer lifecycle."
Another investor, Archangel co-founder and venture partner Quentin Wallace, says it is impressive how Obi learns from every user session, understands where people struggle and feeds that insight back to the business.
"That kind of feedback loop has been missing in software for far too long. Cor is by far the most capable product I’ve seen in this space," he says.
"I saw one of the earliest versions of Obi and the experience instantly felt magical. Now, whenever I use software, I find myself wishing every product had a resource like this."
During his time at Google Startups, Cor co-founder Mantas Aleksiejevas saw how clients were struggling to onboard customers to their software products, representing a major bottleneck for international growth.
"We built Obi because we saw companies struggling with a fundamental problem: existing AI tools can't maintain context long enough to guide someone through complex product workflows," he says.
"Think of Obi as a product expert available to join your session whenever you need them. Obi meets users where they are, at any time of day and jumps in exactly when it's needed."
Early adopters of Cor's offering include Gold Coast-based AI-powered virtual receptionist startup Sophiie AI and Sydney-headquartered site assessment prop-tech Canibuild.
Obi has been utilised by Sophiie AI to onboard non-technical users, including tradespeople such as plumbers and electricians, as well as SME business owners.
"Obi has been essential to our hypergrowth over the last 12 months. It manages hundreds of onboarding sessions every month, helping users train and integrate our AI agent Sophiie into their workflows," says Sophiie AI's co-founder Jacob Banks, who alongside fellow co-founder Luke Kelleher was a finalist in the 2025 Gold Coast Young Entrepreneur Awards.
"Obi helped us move our onboarding from completely manual to AI-led," adds Canibuild chief commercial officer Mark Deacon. "Customers now spend an average of 35 minutes learning our software with Obi at their convenience, time savings that free our team to focus on strategy and relationships."
The 80 per cent rare or non-usage claim at the start of this article comes from a 2019 white paper from US-based software experience management company Pendo, which aggregated anonymised product usage data to find 56 per cent of features were used rarely and 24 per cent were never used.
)
