AI’s USB-C Moment
The Boring Breakthrough That Could Make AI Useful in Your Finance Department
One week it’s a dazzling new generativeAI model demo; the next, a headline about failed enterprise rollouts. The noise is loud, the stakes feel high, and for finance leaders it can feel like more distraction than direction.
If generative AI is indeed drifting toward the hype cycle’s trough of disillusionment, engineers are already laying the groundwork for what follows.
In a somewhat boring look back, history shows it’s not the splashy breakthroughs that matter most, but the quiet protocols that unlock scale.
In the early days of the internet, that was HTTP. Nobody outside of engineering circles paid attention, but HTTP made the web interoperable. It gave us a way to connect information across machines, browsers, and networks. Without it, there’s no Amazon, no Google, no digital economy.
A similar protocol moment is happening right now in AI. It’s called the Model Context Protocol (MCP), and though it sounds esoteric, finance leaders should pay attention. It might be the difference between AI as a parlor trick and AI as a dependable member of your finance team.
Why a Protocol Matters More Than Another Model
Pundits have been calling MCP the USB-C of AI, which makes sense for the most part. It’s meant to be the one format to rule them all.
Instead of every software vendor building bespoke integrations with every AI model, MCP creates a standard connector. Once in place, any AI assistant can plug into your systems (ERP, CRM, warehouse, you name it), and do useful work under the right permissions.
That means fewer custom projects, faster deployments, and most importantly, a common governance boundary. Every AI action from pulling a P&L to drafting a journal entry could flow through the MCP interface, which could provide consistency, auditability, and security by design.
Sounds like just what finance leaders need, right?
The significance is less about technical plumbing and more about leverage. Protocols stabilize the foundation so innovation can compound. They give you bargaining power against vendor lock-in. They cut your integration costs. They make AI projects easier to justify in a budget meeting.
Special Bonus for Deep Finance Dispatch Subscribers:
In last week’s issue, we discussed Rosie, the Excel add-in that brings ChatGPT capabilities directly to your spreadsheets. The team at Ask Rosie has extended a free month to our subscribers who use the promo code: DeepFinance.
Thanks to Dennis and team for the offer!
From Toys to Tools
Right now, most finance teams dabbling in AI are living in copy-paste purgatory. Export a CSV, paste into a chatbot, get a narrative back. It’s a proof of concept, not a system you’d run a close on.
MCP changes that dynamic. By giving AI agents structured, governed access to your live systems, they can:
Fetch data directly: No more spreadsheets shuttled between apps.
Retain context: Remember prior steps in an analysis, reducing rework.
Invoke tools: Trigger allowed workflows like generating draft flux commentary or creating a case for a flagged invoice.
Stay within guardrails: Respect role-based permissions, with logs auditors can actually review.
It’s not magic. It’s just the kind of boring standard that transforms AI from a curiosity into a workhorse.
The NetSuite Example: A Protocol in the Wild
If you want proof this is real, look at Oracle NetSuite. The ERP giant recently rolled out an MCP server through its AI Connector Service. In plain English, that means you can bring your own AI assistant, whether it’s OpenAI, Anthropic, or another, and let it query and act on your NetSuite data via MCP.
For a controller, that might mean the AI can pull unreconciled accounts, propose journal groupings, and draft variance explanations, all tied back to NetSuite records. For FP&A, it might mean conversational queries into actuals vs. budget with drill-through to transactions. And for a CFO, it’s leverage: you can experiment with different AI models without re-architecting integrations each time.
It’s the first ERP-grade signal that this protocol has moved from theory to practice; and it sets a precedent other SaaS vendors will be under pressure to follow.
Why Finance Leaders Should Care
The last thing most finance chiefs want is another acronym. But MCP isn’t a vendor pitch; it’s an industry standard in the making. And standards shape markets.
Operational impact: Shorter closes, faster variance analysis, less time spent on manual reconciliations.
Cost control: One integration instead of many. Fewer IT tickets. Lower consulting spend.
Governance: Audit trails, permissioning, and evidence that satisfies SOX and board audit committees.
Strategic flexibility: The ability to swap models or vendors without tearing up the plumbing.
Protocols don’t trend on social media, but they change the rules of the game. Finance leaders who recognize this early will be the ones who get real productivity gains from AI while their peers are still chasing the next shiny demo.
The Bottom Line
Generative AI will keep grabbing headlines with bigger models and smarter demos. For finance, the real story is less about raw intelligence and more about the infrastructure that makes it usable.
HTTP quietly built the web. MCP may quietly build the AI economy. For finance leaders, that’s the part worth paying attention to.
🔒 Pro Version Extras
Subscribers get access to two downloadable resources designed to help finance teams put MCP into practice:
MCP Finance Server List – A list of 10 Finance SaaS vendors that have released official MCP servers, with links and finance-specific use cases.
MCP Explainer – A visual walkthrough of how MCP servers work, including the interaction diagram and data funnel, with plain-language notes for finance teams.
Unlock these downloads and all our Pro content: