When Your LCA Tool Won't Answer: Why Pause Beats Silent Auto-Fill

I sat in the audit room, EPD printout in hand, when the question came: “Where did you source the GWP value for virgin polyester?”

I scrolled back through my project. The software showed a number—47.3 kg CO₂-eq per kg material—but the path that got me there was opaque. The tool had “retrieved” it, I explained. From where? Best guess: some weighted average across three databases. Which three? The software didn’t say. Why those databases over others? Silence. An auditor’s pencil hovered over a checkbox marked “defensible” and didn’t move.

That moment taught me that not all LCA tools are built the same way. Some are optimized for speed. Some for simplicity. But when an auditor or a regulatory body asks you to reproduce your work, speed and simplicity become liabilities. What you actually need is transparency—the kind that keeps decision authority where it belongs: with you.

This is the difference between auto-fill and pause-and-ask.

Auto-Fill: Fast, Fragile

Silent auto-fill sounds helpful. You tell the tool you need a GWP value for a material. It disappears into its databases, runs some logic you can’t quite see, and surfaces a number. Done. Move forward. Your project finishes faster.

But here’s what happens to the audit trail: the tool selects a database. It takes an average. It outputs a number. The chain of reasoning that produced that result stays locked inside the software. When an auditor asks where the value came from, you have no answer that isn’t a guess. When a client asks why you chose this material data over that one, you have no documented reasoning. When you try to submit the result in an EPD or ISO 14040 report that requires you to justify your choices, the form asks: “Why did you select this source?” Your answer—“the tool picked it”—is not defensible.

Auto-fill breaks the audit trail inside the tool. The practitioner—you—loses decision authority without losing accountability.

Pause-and-Ask: Slower, Defensible

Cortex works differently. Instead of hiding the decision, it surfaces it.

Let’s say you’re looking for GWP data for the same virgin polyester. Cortex retrieves from 14 databases. Across those databases, the values might cluster tightly or they might scatter. Maybe database A shows 45 kg CO₂-eq/kg, and database B shows 98 kg CO₂-eq/kg—a spread of more than two times. That spread signals a real disagreement in the scientific record, not a rounding error. The tool could average them and move on. Instead, Cortex pauses.

At that pause, you see options. Not a single number: a ranked set of candidates, each individually scored across the five DQI dimensions—Temporal, Geographic, Technology, Completeness, and Reliability. These dimensions, derived from the Pedigree-Matrix methodology, let you compare data quality head-to-head. You might see:

• A recent, geographically matched dataset with spotty completeness
• An older, globally averaged dataset with complete coverage
• A database entry for the specific technology you’re using, but from a different region
• A proxy material that covers the gap

Each has trade-offs. The tool doesn’t hide them. You don’t have to guess. You compare them, decide which trade-off makes sense for your EPD, and record your choice. That decision—your decision—becomes part of the reasoning chain that you submit with your result.

Cortex pauses whenever it matters: when coverage falls below 80% and a proxy is needed; when geographic or technology representativeness is uncertain; when DQI dimension scores conflict; when cross-database spread signals genuine scientific disagreement. In each case, you get 2 to 4 concrete examples to weigh. The tool doesn’t average them. You decide.

Why This Matters for Audit Defense

An auditor’s job is to ask: Did you make a defensible choice? A defensible choice is not automatically the most conservative one or the most recent one. It’s one you can justify—and in writing, with reasons—given the data you had available.

When you use auto-fill, you can’t. The tool made the choice in the dark. You inherited the result.

When you use pause-and-ask, you can. Every decision is documented. Every choice points to the data you compared, the dimensions you weighed, the constraints you were working under. If the auditor questions the choice, you’re not saying “the tool picked it.” You’re saying: “I selected this database because it matched our geography and production method, even though another source was more recent, because recency was less critical than representativeness for this material, and here’s my reasoning.” That’s defensible.

The reasoning chain isn’t a compliance checkbox. It’s the record of your judgment—the evidence that you thought carefully about your data sources instead of hoping an algorithm would do your thinking for you.

The Practitioner Decides; The Tool Records

Cortex produces the reasoning chain. You—the practitioner—submit it. Cortex does not auto-file. You decide what goes into your EPD or your ISO report. You decide which data quality trade-offs are acceptable. You decide how to align your choices with the standards that govern your sector. The tool’s job is to surface the options and record your decision. Your job is to decide.

That distinction is everything. It moves you from being a passenger riding along with a black box to being the driver, with the route map visible on your dashboard.

I don’t know what happened in that audit room if the question had been asked of an auto-fill tool. Maybe the auditor would have signed off on the magic number. Maybe not. What I know is that I don’t want to find out. The tools that are easiest to use—the ones that hide the work and just give you an answer—are the ones that make you most vulnerable when someone asks to see the work.

Pause and ask. Make the tool show its options. Make yourself decide. Record what you chose and why. That’s not slower. That’s just honest.

And honesty is what audits are for.