Regulatory intelligence tools mostly deliver a feed. Agency announcements, guidance updates, warning letters, inspection outcomes — filtered by product name or therapeutic area and delivered to a mailbox.
Feeds are easy to build and easy to demonstrate. They also, in practice, serve about one of the six roles they are bought for.
What a news item does not tell you
An item says what happened. It does not say which of yours it bears on, and it does not say what that now obliges.
Every reader is working both out in their head, from a product tag and a paragraph of agency prose. That reconstruction is the work — and it requires knowing your own portfolio well enough to connect a warning letter about a facility to the fact that the facility makes an intermediate for one of your products.
Which is why the CMC lead gets nothing out of a page organised around news. No amount of better filtering fixes that, because the information they need was never in the feed.
Three things, not one
DnXT’s regulatory intelligence makes the unit of the page an impact rather than an item. An impact carries three things:
- The finding — what the source said.
- The asset — the specific thing of yours it concerns. Not just the product: the site, or the substance, or the supplier itself.
- The obligation — what this now requires of you, if anything.
Around those sit the decision, the functions involved, an owner, a deadline and the change control reference — an assessment record attached to something specific enough to actually assess.
The asset is free at the moment you find it, and unrecoverable later
This is the observation the whole design turns on, and it is not obvious.
A watch-list item was found by asking a source about one substance or one site. The search that found it was “tell me about inspections at this facility”. So the thing the item concerns is the term that found it. It is known, exactly, at the moment it arrives. Nothing has to be worked out.
Afterwards, all that survives on the item is a product name and some prose. “Which of our sites does this concern?” becomes a guess — a plausible guess, from a person or from software, that can be wrong.
That asymmetry is why DnXT records the asset at the moment of the search rather than trying to work it out later on the page. Free now. Unrecoverable ten minutes from now. A surprising amount of good design in this domain is simply noticing which facts are cheap today and expensive tomorrow.
Raised for every hit, not only the important-looking ones
The tempting shortcut is to raise impacts only for serious-looking findings, so the list stays short.
DnXT raises one for every watch hit. By definition a watch hit is about your organisation — it was found by asking about something you own. Whether it matters is what the assessment decides, and filtering it out beforehand is software making a judgement that belongs to a regulatory professional.
The same logic governs routing. A hit on a site or a substance goes to CMC, because the asset says so — that is a fact, not a guess. Everything else is left unassigned. Guessing a function puts the item on the wrong desk, and an item on the wrong desk is worse than one on no desk, because the wrong owner closes it.
The history nobody budgets for
Items collected before impacts existed would never become impacts. Duplicate detection means an item is never collected twice, so there is no later moment at which an old one gets picked up.
Left alone, that gives you a capability that works perfectly from its go-live date and is empty behind it — which, for a team whose live supplier concerns started six months ago, means it is empty for everything they care about.
So each watch-list run picks up the historical items first, matching each back to the term that found it through the evidence already attached to it. It fixes itself rather than depending on when a customer happened to switch it on. A small point, and the difference between something that demonstrates well and something that is useful in week one.
The view groups by asset, because that is the question
A list puts things in order. The question a regulatory lead is actually asking is where pressure concentrates.
Not “eleven findings this month” but “three of them are at one supplier”. That is a different shape of information and it needs a different display: cards grouped by asset, most findings first, with colour reserved for the ones nobody has decided on yet.
It sits above the working list rather than replacing it. The list is where a decision gets made item by item. The cards are where you decide which items matter this week. Both are needed, and combining them produces a page that does neither well.
What to ask when you evaluate one of these
Three questions separate a portfolio-aware system from a filtered feed:
- Does a result name the asset, or only the product? If a supplier finding cannot tell you which of your sites it concerns, somebody is doing that work by hand every single time.
- What happens to items that predate the setup? If the answer is “nothing”, the tool is empty for exactly the history you need.
- Does it guess an owner? A confident wrong routing costs more than no routing, because it looks handled.
DnXT builds eCTD publishing, submission planning, document management and dossier review software for regulatory operations teams. Book a demo to see portfolio-aware regulatory intelligence against your own submissions.
