Monitoring filings in real time — now onboarding design partners for 2026 · request access ›
Agentic AI for earnings

Earnings out. Model updated in minutes.

Qubera reads company press releases the moment they're published and updates the correct input drivers across your Excel workbooks — automatically, within minutes, with you in full control.

No workbook changes. Complete audit control. Built for how analysts already work.

Press Release · 07:00:04 ET · detecting…
Acme Corp Reports Third Quarter 2026 Results
Q3 revenue $4.82B
Diluted EPS $2.14
Qubera
Qubera
watching the wire…
Acme_Model_Q3.xlsx · Drivers
Driver
Value
14
Revenue ($B)
4.61
15
Gross margin
61.4%
16
Diluted EPS
1.98
awaiting update
You approve every change
Built by operators
20+ years on institutional trading floors  ·  AI & compliance infrastructure at scale  ·  engineering from the ground up
The earnings-season fire drill

The data is structured. The workflow isn't.

Every quarter, analysts race to pull the same numbers out of press releases and type them into models by hand — under deadline, across dozens of names, with their P&L on the line.

Manual

Copy-paste from PDFs and press wires into the right cells — line by line, name by name, every quarter.

~45 min per model, by hand

Slow

Numbers move markets in seconds. By the time the model is updated, the call has often already happened.

Minutes matter at the open

Error-prone

One mistyped driver cascades through every formula. Fat-finger risk is highest exactly when it's busiest.

A single cell breaks the chain
How it works

From the wire to your workbook, automatically

Qubera runs the manual loop for you — and shows its work at every step so you stay fully in control.

01

Detect

Qubera watches the wires and filings for the names you cover and catches the release the second it drops.

02

Extract

It reads the unstructured text and pulls the reported metrics — revenue, margins, EPS, segment data — with their context.

03

Map

Each metric is matched to the exact input driver in your model. Your formulas, structure and logic stay untouched.

04

Update

You review a clean diff of proposed changes, each linked to its source, and approve. The model recalculates.

Your model, your way

We automate the typing — not how you work.

Qubera never rebuilds, reformats, or "improves" your workbook. It writes values into the input cells you already use, exactly where you'd type them — so your model stays yours.

Input drivers only

Values land in your existing input cells. Formulas, links and formatting are never altered.

Source-linked every time

Click any updated value to jump to the exact sentence in the release it came from.

proposed-changes · Acme_Model_Q3.xlsx
Sheet: Drivers
+B14  Revenue ($B)4.614.82
+B16  Diluted EPS1.982.14
+B22  Cloud rev ($B)1.201.37
B15  Gross marginunchanged
Approve all · or accept line by line
Why analysts use Qubera

The speed of automation, the control of doing it yourself

Updated in minutes

Models reflect reported numbers within minutes of the release — before the call, not after.

Human in the loop

Nothing changes without your approval. Review a clean diff and accept all or line by line.

Full audit trail

Every value is timestamped and linked to its source sentence. Defensible by construction.

Works with your models

No migration, no template, no rebuild. Point Qubera at the workbook you already use.

Coverage that scales

One analyst, the whole coverage list. Qubera handles the long tail of names in parallel.

Nothing slips

Catch every filer the moment it reports — even the names you'd otherwise update late.

See it run

Two minutes from press release to updated model

Watch Qubera read a real release and propose driver updates — and see exactly where the human stays in control.

Full demo

Press release → Excel

End to end: detection, extraction, mapping, and the recalculated model.

Walkthrough

The human in the loop

How analysts review the diff, check sources, and approve changes.

Security & trust

Built for the standards of a trading floor

Designed by people who shipped compliance and infrastructure inside global banks — security isn't an afterthought, it's the starting point.

Your data stays yours

Models and credentials are isolated per client. We never train on your workbooks.

Encrypted end to end

Data is encrypted in transit and at rest, with access scoped to the people you authorize.

Complete audit log

Every detection, proposal and approval is recorded — exportable and review-ready.

No black box

Every number is explainable and traceable back to the source it was read from.

The team

Operators who lived the problem

Two decades on trading floors and in AI infrastructure — building Qubera to fix the workflow they ran by hand for years.

Ronti Pal
Co-founder · CEO

Ronti Pal

  • Built & ran billion-dollar trading businesses at Citi & Barclays over 20 years.
  • Started the algorithmic trading business at Barclays.
  • CFO / COO at PayActiv and Molekule.
  • BS Computer Science, Caltech.
MarketsTradingOperations
Xavier Mandeng
Co-founder · CTO

Xavier Mandeng

  • Developed and implemented AI-powered technology solutions at Quantile (acquired by London Stock Exchange) and Credit Suisse.
  • Conceptualized and built THEAMA — an automated NLP-driven compliance system that eliminated millions in regulatory fines for Credit Suisse.
  • AI-native, with hands-on experience developing Deep Learning systems (CNNs).
  • Extensive experience with infosec and technical workflows at financial firms.
NLPDeep LearningInfoSec
Request access

Get Qubera on your coverage before next earnings season.

We're onboarding a small group of design partners. Tell us a little about your models and we'll be in touch within one business day.

  • Works with your existing Excel models
  • Hands-on onboarding with the founders
  • No commitment to start
We'll reply within one business day.

Thanks — request received.

We'll be in touch within one business day.