AI
The Context Moat: Why 95% of AI Projects Fail and What the Winning 5% Know
As AI becomes a commodity, the firms creating lasting value are building something far harder to replicate: a context moat.
"The world doesn't need another AI assistant."
A conversation with Dasseti's Head of Product on MCP, AI and the future of investment technology
Dasseti
Aug 11, 2026
Artificial intelligence is rapidly reshaping the way investment firms work.
From Microsoft Copilot to Claude and ChatGPT, AI assistants are becoming part of everyday workflows. But as adoption accelerates, a new challenge is emerging: how do you connect those assistants to the trusted investment intelligence that firms rely on every day?
We sat down with Arjun Patel, Head of Product at Dasseti, to discuss why Model Context Protocol (MCP) is generating so much attention, why Dasseti has invested in it, and what it means for the future of due diligence and investor relations.
"The conversation has changed."
According to Arjun, the biggest shift isn't the rapid pace of AI innovation, it's the questions customers are asking.
"A year ago the question was, 'Should we be using AI at all?' That question's basically gone now. Firms have answered it. They've moved on to the harder question: how do we point AI at our own data, inside our own processes, without losing control of it?"
Investment firms are looking for AI that understands their approved investor responses, due diligence findings, manager assessments and internal knowledge, while being trusted to work with that information securely.
AI isn't the problem. Context is.
Today's AI assistants are incredibly capable, but they all share the same limitation. They don't know your organisation.
"A general assistant has read the whole internet. It hasn't read your firm."
Without access to trusted internal information, AI can only provide generic responses, or worse, generate answers that sound convincing but are inaccurate.
"The gap isn't how smart the model is. It's that it hasn't got trusted context sitting behind it."
That's the challenge MCP is designed to solve.
So what is MCP?
If you've been following developments in enterprise AI, you've probably heard the term Model Context Protocol (MCP).
Arjun explains it simply:
"Think of MCP as a universal adapter between AI assistants and the systems that actually hold your data."
Instead of building separate integrations for every AI platform, MCP provides a standard way for assistants like Microsoft Copilot, Claude and ChatGPT to securely connect to enterprise applications.
"MCP isn't another chatbot. It's the pipe. It's what lets the assistant you already use reach the data it needs, and only the data you've allowed it to see."
Why Dasseti isn't building another chatbot
This is where Dasseti's strategy differs. Rather than creating another AI interface, the focus is on making customers' existing AI tools significantly more useful.
"We made a deliberate call not to build another AI assistant."
Customers have already invested in platforms like Microsoft Copilot, Claude and ChatGPT, the opportunity is to make those tools smarter for their users.
"We're building the intelligence layer behind the assistants our customers already use."
Why Knowledge Base matters
One question Arjun is asked regularly is how Dasseti’s Knowledge Base fits into the wider AI strategy. His answer is simple.
"MCP is the connection, but a connection's only worth anything if there's something solid on the other end of it."
Knowledge Base creates that foundation by organising approved content, manager assessments and institutional knowledge into a structured, searchable layer that AI can reliably access.
Without that structure, AI is left trying to interpret raw documents. With it, organisations can provide accurate, governed answers built from trusted information.
What this means for investment teams
The technology itself is important, but the real value comes from the workflows it enables.
“Imagine an investor relations professional asking Copilot for the latest approved response to an investor question and receiving an answer drawn directly from Dasseti.
Or a due diligence analyst generating a meeting brief, reviewing recent manager changes and updating records, all without leaving their AI assistant.”
Because MCP connects multiple enterprise systems, those workflows don't stop at Dasseti. AI can bring together information from CRM platforms, document repositories and other business systems into a single conversation.
Governance first
As AI capabilities grow, so do questions around governance. For Dasseti, Arjun says, governance isn't an afterthought.
"The assistant only ever reaches what that specific person is already allowed to see."
Permissions, auditability and human oversight remain central to every AI-powered workflow. The goal is to give investment professionals better information, faster.
Looking ahead
Arjun believes the biggest shift over the next few years will be the way people interact with enterprise software altogether.
"People stop living inside individual applications."
Instead, AI assistants become the interface, drawing intelligence from trusted business systems behind the scenes.
That future is exactly what Dasseti is building towards.
As Arjun puts it:
"We're not asking you to change how you work, or which AI you trust. We're taking the tools you already use and making them genuinely useful across your investment processes by connecting them to intelligence you can trust."