When Justin Kim launched his startup four years ago, he wasn’t thinking about banking regulations or sales quotas. His company began as Ami, a mental wellness platform focused on helping people manage pressure and build better habits. But today, that same company – now called Hupo – is helping major financial institutions like Prudential, AXA, and HSBC train their sales teams using AI-powered coaching. What happened in between reveals a crucial lesson about how AI is actually being adopted in enterprise settings: it’s not about flashy technology, but about fitting into existing workflows and solving specific industry problems.
The Pivot That Wasn’t Really a Pivot
“The core problem in both cases is performance at scale,” Kim told TechCrunch. “In banking and insurance, results vary, not because of motivation, but because training, feedback, and confidence differ. Traditional coaching can’t reach everyone, and managers can’t sit in on every conversation.” This insight – that the fundamental challenge was human performance rather than mental wellness specifically – led Hupo to pivot from general wellness to specialized sales coaching for the banking, financial services, and insurance (BFSI) sector.
The company recently raised a $10 million Series A led by DST Global Partners, bringing total funding to $15 million since its 2022 founding. More importantly, Hupo has demonstrated real traction: customers typically expand contracts 3�8x within the first six months, and the startup now serves dozens of customers across APAC and Europe. Kim’s background selling enterprise software at Bloomberg and working on product development at South Korean fintech Viva Republica gave him unique insight into both the buyer and end-user perspectives in financial services.
The Enterprise AI Reality Check
Hupo’s approach stands in stark contrast to many AI startups that begin with technology and then search for problems to solve. According to a Financial Times analysis, AI startups typically use standard AI models like ChatGPT or Claude and build specialized software on top, promising efficiency gains but struggling with high costs and integration challenges. These startups spend double what traditional SaaS companies spend on compute and infrastructure, yet often lack sustainable business models and face difficulties integrating with broader business processes.
“One of the biggest lessons I’ve learned is that, especially with large enterprises, you have to understand their business and industry in detail,” Kim emphasized. Hupo built its platform around how banks and insurers actually operate, training its models from the start on real financial products, common objections, client types, and regulatory requirements. This industry-specific approach helps explain why established platforms like Salesforce, SAP, and Microsoft have advantages in the AI race – they already have decades of industry knowledge, existing software integrations, and regulatory compliance infrastructure.
The Regulatory Storm Clouds Gathering
Hupo’s expansion into the U.S. market comes at a particularly turbulent time for the financial sector. Just this week, President Donald Trump called for capping credit card interest rates at 10% for one year, sending shares of major banks and credit card companies tumbling. The proposal, which Trump says would take effect January 20, 2026, has created uncertainty across the financial industry, with banking associations warning it would “reduce credit availability and be devastating for millions of American families and small businesses.”
This regulatory uncertainty creates both challenges and opportunities for AI companies like Hupo. On one hand, financial institutions facing margin pressure may be more motivated to adopt efficiency tools. On the other, they may become more risk-averse about new technology investments. “Distribution-heavy financial models create a strong need for scalable coaching,” Kim noted about the U.S. market, suggesting that the very pressures creating industry turmoil could drive demand for his product.
The Dark Side of AI Innovation
While Hupo focuses on legitimate business applications, other AI developments highlight the technology’s potential for harm. The UK-based Internet Watch Foundation recently found criminal sexual imagery of girls aged 11�13 on a dark web forum that appears to have been generated using xAI’s Grok model. “We are extremely concerned about the ease and speed with which people can apparently generate photo-realistic child sexual abuse material,” said Ngaire Alexander of the IWF.
Separately, researchers discovered that ChatGPT remains vulnerable to data exfiltration attacks through a new vulnerability called ZombieAgent. “Guardrails should not be considered fundamental solutions for the prompt injection problems,” warned Pascal Geenens, VP of threat intelligence at Radware. “As long as there is no fundamental solution, prompt injection will remain an active threat and a real risk for organizations deploying AI assistants and agents.”
The Acquisition Trend and What It Means
Hupo’s growth comes amid a wave of AI acquisitions, with OpenAI recently acquiring the team behind executive coaching AI tool Convogo in an all-stock deal. This marks OpenAI’s ninth acquisition in a year, signaling how major AI players are using M&A as a talent and capability accelerator. The Convogo founders explained their move by saying, “We’re convinced now more than ever that the key to bridging that gap lies in thoughtful, purpose-built experiences.”
This trend suggests that while specialized AI startups like Hupo can find success in specific verticals, they may eventually face competition from or acquisition by larger players with deeper resources. The question becomes: can focused startups maintain their industry-specific advantages as general AI platforms expand into their domains?
The Road Ahead for Enterprise AI
Looking forward, Kim envisions Hupo expanding beyond sales coaching to help large teams perform at scale across entire organizations. “In five years, I want Hupo to go beyond sales coaching and help large teams perform at scale, giving managers and employees clearer insights and practical guidance, even across tens of thousands of people,” he said.
But the broader landscape suggests that enterprise AI success requires more than just good technology. It demands deep industry knowledge, regulatory awareness, and the ability to integrate with existing systems and workflows. As financial institutions navigate regulatory changes, economic pressures, and technological transformation, tools like Hupo’s that address specific pain points with industry-specific solutions may have the best chance of surviving – and thriving – in the complex world of enterprise AI.