The problem Documents How it works Results Trust Scope a pilot
Home/GoCore/Asset Management
For / Private markets, credit and public funds

The fee terms are in the side letter, not in the system.

We read the documents your teams handle by hand — side letters, capital account statements, K-1s, credit agreements — and turn them into clean, checked fund data. Inside your own environment.

Runs in your environment Every value traced to its page People approve the important steps
Same dayPosition data, not a quarter late
1 answerTo “what are the live fee terms?”
Every fieldLinked to the clause it came from
0Documents leave your environment
the problem

Where the work slows down.

Four places where people still read and re-type documents by hand.

01

Re-keying is the operating model

Fund ops reads GP statements and notices, then types the numbers into the accounting system.

Cost grows with every fund
02

Nobody knows the live terms

One relationship is described by an LPA, several amendments and a dozen side letters.

Fee and expense leakage goes unnoticed
03

Positions land late

Private-market data arrives as a PDF, waits in a queue, and is typed in a quarter later.

Decisions made on old numbers
04

Every team built its own extractor

Reporting, risk and tax each wrote their own scripts over the same documents, and they disagree.

Arguments about which number is right
the documents

The documents we read.

Start with the one that costs you the most time today. That is where a pilot starts.

LPAs & side letters

Fee terms, offsets and MFN elections buried in signed PDFs.

Capital calls & distributions

Notices from dozens of GPs, each in its own shape.

Capital account statements

Quarterly statements re-typed into the book of record.

K-1s & tax packs

The annual fire drill for tax and fund-ops teams.

Credit agreements

Covenants, pricing grids and layered amendments.

Research & DDQs

Broker research, call transcripts and consultant questionnaires.

how it works

Five steps from document to data.

One pipeline, inside your environment. A document becomes usable data at step three, and trusted data at step four.

One file· five steps
As it arrivesReady to use
Step 01

Collect

from every source

Step 02

Read

scans, photos, PDFs

Step 03

Extract

the fields that matter

Step 04

Check

scored against your rules

Step 05

Deliver

into your systems

01 Collect Documents arrive from data rooms, GP portals and admin mailboxes Collected into one queue
02 Read Scans, faxes and PDFs are made readable Clear text, tables kept intact
03 Extract The system finds the fields that matter Parties, dates, amounts, clauses
04 Check Each value is checked and scored Low-confidence fields go to a person
05 Deliver Clean data is delivered where it is needed Book of record, search, reporting
Most tools stop at step two. Reading a document is not the same as trusting the data in it.
results

What you get.

One source of clean data that every team and system can use.

No more re-keying

Capital account and notice data lands on your schema without a keystroke.

Live terms in one place

Fee terms resolved across the LPA, amendments and every side letter, each linked to its clause.

Positions on time

Documents picked up where they land and booked the same day.

Numbers you can defend

Every reported value shows the page and clause behind it, ready for audit and filings.

How much runs without a person· by document typetypical targets
Capital calls & notices
94%6% reviewed
K-1s & tax packs
93%7% reviewed
Capital account statements
92%8% reviewed
Credit agreements
90%10% reviewed
LPAs & side letters
88%12% reviewed
The rest is not failure. It is the small share a reviewer checks, with the source page already open. Your actual numbers are measured on your own documents during a two-week discovery.
trust

Built by gothink.ai. Runs inside your environment.

This system reads side letters and LP documents and writes to your records. It belongs inside your walls, not on someone else’s servers.

Your data stays with you

Your cloud, your access controls, your keys. Our software comes in; your documents never go out.

  • No outside AI provider sees a page.
  • Nothing you process trains our models.
  • A named person approves every important action.
More than text recognition

Basic OCR gives you text. We give you checked data, traced to its source, that gets better over time.

TextOCR FieldsExtracted TracedTo the page CheckedYour rules LearnsOver time
See the full comparison →
Small steps, your decision

Each phase ends with a measurable result. You decide whether to continue at every gate.

2–3 weeks Discover
2 weeks Design
4–6 weeks Pilot
6–10 weeks Scale
ongoing Operate
How an engagement runs →
Next step / a 4–6 week pilot

Start with one type of document.

Pick the paperwork that costs you the most time. We prove the results on your own files, in weeks.

1

Choose the documents

One fund family’s side letters, or a quarter of capital account statements. Whichever hurts most.

2

Two-week discovery

We map the fields you need, agree how confident the system must be, and measure today’s manual effort.

3

See the result

A working pipeline on your documents, scored against answers your own experts approved.