Who they are
Cass.ai is an early-stage startup building AI-powered mental health support: a conversational companion designed for empathetic, always-available help.
It had a product in market, a clear vision, and investors to convince — and no senior technical leadership to carry them.
The challenge
Cass.ai had a product and a clear vision: accessible, empathetic mental-health support through conversational AI. What it did not have was senior technical leadership — someone to own the architecture, build out the engineering team, direct the AI development, and hold the technology story with investors and the board.
The timing made that gap sharper. This was the earliest period of the LLM era — GPT-3 had only just been released — and in a product whose users may be in genuine distress, the cost of getting the AI wrong is not commercial. It is human.
Without a senior technical leader in place, the risk was not just slower development. It was building in the wrong direction entirely.
What we did
AHM Technology was engaged to provide fractional CTO leadership, with Neil Markey stepping in as CTO and taking direct ownership of Cass.ai's technology strategy and execution.
Technology Leadership and Organizational Growth
Stepping into the CTO role, AHM Technology provided strategic direction to Cass.ai's executive leadership and board — defining the technology roadmap and architectural vision for the platform, and hiring and onboarding engineering resources to build out the team.
Senior technology leadership without a full-time hire
AI Platform Architecture and Development
AHM Technology led the design and development of Cass.ai's own internal AI platform, architecting scalable systems to support real-time, empathetic user interactions. It worked — and it was deliberately built to evolve as the model landscape did, which mattered more than anyone could have known at the time.
An AI platform of the company's own — built to evolve
Machine Learning Model Development
Bringing machine-learning expertise into the team, custom models were built and trained specifically for mental-health conversational use cases — natural language understanding, sentiment analysis, conversational flow — on transformer models including Google's BERT, in the days before large language models were the obvious choice. The engineering team was mentored through the whole process, building durable AI expertise internally rather than creating dependency on external resources.
Durable AI expertise inside the company, not a dependency outside it
Adopting GPT-3 — Safely
As OpenAI released GPT-3, the team worked with OpenAI to adopt the LLM into the platform — genuinely early adoption, at a time when hallucination was real and frequent. In a mental-health product that risk is intolerable: a fabricated answer to someone who is suffering is not a defect, it is a harm.
So safety controls were engineered on top of the model — detecting hallucination and dealing with it before anything reached the user. The platform got the power of the frontier model, with the guardrails the product's users deserved.
Frontier AI in production, with the safety controls a mental-health product demands
Product Innovation and Mobile Experience
AHM Technology oversaw the development of the full chatbot experience, leading product and design teams to deliver seamless, empathetic user interactions across the platform.
A chatbot experience users could open up to
Strategic Advisory and Investor Engagement
The fractional CTO acted as a trusted advisor to the executive team and board, representing the technology vision in investor discussions and providing guidance on AI adoption, product strategy, and go-to-market approach.
A technology story investors could back
GPT-3 was adopted almost the moment it released — wrapped in safety controls built to catch hallucination before it ever reached someone who was suffering.