Skip to content
Agentic AI Enablement/5 Min Read/

UNLEASH THE PLATFORM.IGNITE EVERY USER.

Rody Arantes

Rody Arantes

Director of Digital Technology

Head of Platform Engineering & IT

ShareLinkedInSubstack
A wide cinematic editorial sci-fi illustration. A luminous translucent data vault at the center of the frame unleashes bright amber and starlight energy beams outward in graceful arcs, reaching a constellation of workstations arranged around it. Four people work at their stations, each illuminated by the energy lines connecting them to their tool of choice: a glowing analytics dashboard, a luminous code editor, an AI chat window, a holographic document. Every user is visibly supercharged by the platform reaching them. Warm amber and starlight tones against a cosmic deep navy background.

A data platform is one of the most valuable assets a company ever builds, and a petabyte-scale one is a genuine feat of engineering: information gathered from everywhere, cleaned, secured, and kept current, standing ready to answer almost any question the business can throw at it. The exciting question has always been what happens when you put that kind of power in front of everyone at once. Until recently, you mostly could not. The people who could have been using the data spent their day somewhere else entirely, in a spreadsheet, in a slide deck, in a code editor, while the platform stood ready with everything they needed.

In a startup, resources are finite, and building tailored interfaces for every group was never on the table. So the platform shipped the one interface it could afford to make great.

Here is what changes everything. Agentic-powered platform development now lets a team stand up enterprise-grade middleware, the production-quality software that carries the data straight into any tool the teams already use, in hours to a couple of days at most, instead of the quarter to half a year it once demanded. The platform can finally reach out and meet people exactly where they are, and the strategy that was always too expensive to attempt becomes the obvious one to implement. The data goes to the scientist inside their spreadsheet, to the engineer inside their IDE, into the AI chat window the analyst or scientist already has open, and all at once the whole company is standing on the same ground truth data for decision making.

And this breaks the bargain every engineering leader has lived under for years. Production-quality engineering was slow to build, and slow building was the price of quality. Agentic-powered platform development collapses that choice. Enterprise-grade and startup-speed now arrive together, and every user lights up the moment the platform reaches their tools.

The data was always the asset. Agentic-powered platform development is what finally unleashes it, everywhere work happens.

WHAT IT LOOKS LIKE

This is what agentic-powered platform development now puts within reach. Picture the same platform and the same petabytes, reaching every person through the door that fits the way they already work. An add-on lights up the spreadsheet or the document they are already building. A plug-in puts the platform inside the code editor, where an engineer's assistant reaches for it without breaking stride. A single connector built on an open standard lets every AI assistant in the company draw on the platform directly, and a ready-made template hands someone a running start with the data already wired in. Multiple doors into the platform, each of them putting the full force of the data behind a person who used to be making decisions on quickly outdated data.

WHAT IT GIVES PEOPLE

None of this is really about the tools. It is about what they unleash in the people who use them. A scientist who can put a question to the platform from inside their own spreadsheet, and get back an answer drawn from the entire data asset in seconds, moves faster and decides with real confidence, with nobody to wait on. Now multiply that across every group in the company. The asset that cost so much to build can pay back in time savings, increased decision quality, and reliability, because the value was never in storing the petabytes. It was always in what people do with them the moment the data is within reach.

THE DASHBOARD STILL MATTERS

And none of it means the dashboard goes away. It is still where a leader turns to take in the whole company at a glance before a decision that needs the whole picture, and designing that view with the right altitude and clarity is some of the most valuable work a platform team does. The new surfaces simply join it, reaching all the people a single dashboard was never going to serve. Nothing is replaced. The reach explodes outward.

ONE FOUNDATION

What makes this safe to do at full scale is that every door opens onto the same secure vault. All of these surfaces read from a single foundation, sharing the same data, the same permissions, the same security, and the same record of who did what, so a question answered inside a spreadsheet returns exactly what the same question returns through an assistant or on the dashboard. What this really buys is speed without compromise: ship a new surface on Monday, live by Friday, because the foundation has already done the security work every door will inherit. The surfaces multiply while the foundation holds steady, and that is what lets you hand the whole company the keys without ever losing control of the vault.

WHERE TO INVEST

For a leader deciding where the next dollar should go, here is the good news: the hard part is already built and paid for. The data is there. The return now comes from the last mile, the middleware that carries the data into the tools people already use and the foundation that keeps it safe in transit, and that is a far smaller investment than the platform itself. Fund it, and your most valuable asset goes to work across the entire company at once. The teams that move on this now will set the pace for everyone else: every user lit up, every decision on ground truth, and the platform they spent so much to build finally running at full power.

You have already built the hard part. The only question left is how fast you put it in everyone's hands.

Notes & References

  • Building the middleware got dramatically faster. The largest measured gains from coding assistants land on new, routine code; McKinsey measured new code written in roughly half the time. A 2025 METR study marks the honest boundary: experienced engineers working deep in large, familiar codebases saw no speedup, which is the opposite of the focused connectors described here.
  • One connector, every assistant. The Model Context Protocol, introduced as an open standard in November 2024 (announcement) and since adopted across the major AI assistants, standardizes the connection itself, so one server is reachable by every compatible assistant (spec).
  • AI is moving into the tools people use. Gartner projects that 40 percent of enterprise applications will feature task-specific AI agents by 2026, up from less than 5 percent in 2025.

Tags

  • agentic AI enablement
  • meet users where they work
  • AI middleware
  • browser extension AI
  • Excel add-on AI
  • Model Context Protocol
  • MCP server
  • embedded AI
  • AI in the flow of work
  • platform engineering
  • internal developer platform
  • AI platform leadership
  • Cambridge biotech AI
  • Rody Arantes
ShareLinkedInSubstack