Connected to your systems

Hours a week,
by hand?
That’s a job
for StoriBot.

StoriBot drafts the files and reporting pages your team rebuilds, every number traced to source. It completes the tasks they carry between systems, the way a person would. Anything costing three hours a week qualifies. Start with one.

Get started

Free to start. Best if you already know the job you’d hand over first.

Book a meeting

We’ll help you see where AI best fits into your workflow.

Drafted from Kalgarra · Operations
kalgarra.stori.bot/q1-fy26
Live
Kalgarra Resources
Q1 FY26 quarterly operations review
ASX: KGR  ·  28 OCT 2026
Gold produced
27,412oz
Up 8.4% on prior quarter
Production report · monthly · Jul–Sep 26
AISC
A$2,180/oz
Inside guidance of A$2,100–2,300
Cost model v14 · tab “AISC” · cell H22
Cash & bullion
A$164.2m
At 30 September 2026
Cash reconciliation · GL 1010 + 1040 · closed 12 Oct
Quarterly production (oz)
Q2 25
Q3 25
Q4 25
Q1 26
Production report · monthly · 8 quarters

Higher grades through the Mt Ivor pit drove the production lift, with mill throughput of 1.42Mt for the quarter.

Plant daily log · throughput.mtd · Sep 26

Guidance for FY26 is unchanged at 105,000–115,000oz.

Board paper · FY26 guidance · approved 14 Jul 26
  • Stripe
  • Xero
  • MYOB
  • Business Central
  • Shopify
  • HubSpot
  • PostHog
  • Amplitude
  • Jira
  • Airtable
  • Granola
  • OneDrive
  • SharePoint
  • Dropbox
  • Zernio

Read by default, and every write is one you granted. Uploaded files and grounded web research are sources too. See what connects

Your business runs on a dozen systems.The part that makes sense of them lives in someone’s head.

So the same person rebuilds the same slides every month, someone else reads one system and updates another by hand, and anyone with a question has to go and ask them. A Stori is where that part of the business lives. Connected to your systems, in your brand, open to your team.

What a Stori does

Two jobs, off the same connections

Connect it once, to the systems you choose. Both jobs run off those connections, so the number in the file, on the page and in the system it was written to is the same number.

Job 01 · Files and pages

Drafts what you'd have written

The board pack, the monthly report, the client update: whatever your team rebuilds on a cycle. Drafted from your own systems, in your template and your language, with every number traced to source. You add what you think.

  • fileA real PowerPoint, Word document or spreadsheet you can open and edit
  • pageOne live version everyone works from, not a screenshot in a thread
Job 02 · Tasks between systems

Finishes what one system starts

The work someone does with two tabs open: read it here, make sense of it, enter it there. StoriBot reads both ends the way a person would, so there's no field mapping to build first. It writes only where you've granted it, and if your system isn't connected yet, we build the connector.

  • liveAn import file dropped in the SharePoint or Dropbox folder your other system watches
  • in buildHubSpot updated from the Granola notes: what was said, decided and promised
  • on requestA connector built for the system only your business runs

Behind both jobs

UpdatesIn build

It comes to you

You don't go and check a dashboard. When something moves, or on the day it's due, the update lands — in your inbox, in Slack, or as a file.

  • month-end closeThe pack drafts itself and goes to review
  • grade drops 5%Flagged, with the drivers behind it
The agent

Ask it, argue with it

It knows this part of the business — the systems, the history, how you talk about it. Use it to check something, or to think out loud before you commit to a position.

  • ask“Why did unit costs move in September?”
  • ask“What's the case against expanding the mill?”
Real work

Sent, and still live.

Not mockups. Open any of these and you’re reading the finished artifact — built from live data, with every figure traced back to the source it came from.

See more of what we make
How it works

Four steps, then it runs

01

Point it at something

Upload last month's files to get started — see a Stori with real output on your own numbers. When it makes sense, connect your systems to keep the numbers current without anyone rebuilding them.

02

Set it up, and wire it in

Set it up yourself, or we'll do it with you, then connect the systems it works from. It reads by default, and every write is one you granted — one switch turns them all off. We'll build any connector you're missing, and any business file works as a source.

  • Stripe
  • Xero
  • MYOB
  • Business Central
  • Shopify
  • HubSpot
  • PostHog
  • Amplitude
  • Jira
  • Airtable
  • Granola
  • OneDrive
  • SharePoint
  • Dropbox
  • Zernio
See what connects
03

Review before anything leaves

Every draft goes through a review gate — numbers checked against source, content checked against your rules. Nothing goes out on its own unless you decide it can.

04

Then it just runs

Files and pages draft on demand, tasks run between your systems, updates push, and your team can ask it anything. Setup is versioned, so you can see what changed between runs and why.

Security and data

Built to get through your risk committee

Every permission is granted one operation at a time, and withdrawn the same way. Here’s what finance, IT and legal will want to see.

Residency

Stored and processed in Sydney

Hosted in AWS ap-southeast-2. Storage, processing and the audit trail stay in region; inference is the one call that leaves.

Retention

You choose what the model keeps

By default, inference is restricted to models under zero-retention terms, with no fallback to one that collects. You can turn that off per instance.

Certification

ISO 27001:2022 in progress

Controls implemented and monitored continuously. Internal audit report available under NDA.

Open our Trust Centre
Lineage

Every value traces to source

System, table, cell and timestamp for each figure — on the output itself, not buried in a log.

Access

You choose who sees what

Public, access-code or authenticated — set per share link, not per Stori. Send the same material twice, with different pages open to each audience.

Control

It reads your systems. Every write is one you granted.

Scoped credentials per source, read by default. Nothing in a connected system changes without a grant you made — and one switch turns every write off.

See full trust & controls
FAQ

The questions we get asked first.

What's a Stori?
One part of your business — the board pack, operations, the investor story — wired to the systems that part runs on. It drafts from those numbers, in your brand and your template. You take it the last stretch, the part worth your time, then send it to the people you name and no further. Most teams start with one and add another once it’s running.
Do you connect to our systems, or do we upload?
Both, and most people start by uploading — last month’s pack and the files behind it is enough to see what comes out. Connections come after that, once it’s worth wiring in. They’re scoped per source, read by default, and every write is one you granted. If we don’t connect to yours yet, we’ll build it.
Can it update our systems, or only read them?
It can update them. Give it the task someone does with two tabs open, and it reads one system, works out what belongs in the other, and writes it there. Every write is one you granted, one operation at a time, and one switch turns them all off. Where a system imports from a folder it watches, a file written there does the job with no connector at all.
We already use Zapier or Power Automate. What's different?
Those move a field when a trigger fires, so someone builds the map between the two systems first, and keeps it working. StoriBot reads both ends the way a person would, which is what the work between systems usually needs: a meeting note isn’t a field, and no mapping turns one into a CRM update. Keep them for what they’re good at. This is for the tasks that still need someone to make sense of what they’re reading.
Can the agent be wrong?
It can be wrong about a judgement. It isn’t the thing calculating your numbers — those come from your data through code, and the model narrates them. Every figure carries the system, table, cell and timestamp it came from, so you can check it in a glance rather than taking it on faith. And nothing leaves without going through the review gate.
Can you match our brand?
Yes. We tune to your house voice and brand — fonts, colours, tone, your existing template — so what comes out looks like you, not like a template. If we’re engaged to help you, voice and brand settle over the first 1–3 weeks.
Can we keep editing in Word or PowerPoint?
Yes. Ask for the Word, PowerPoint or Excel version and you get a real file — track changes, comments, the tooling your reviewers already use. Bring the marked-up version back and it runs through the review gate again, then the page updates to match. You don’t have to come back; the document is yours either way.
Can I create inline HTML for my email tool?
Yes. We output an inline-HTML version that drops into the common tools — Mailchimp, HubSpot, Campaign Monitor, Brevo — or your own client, with merge tags for your platform if you use them.
What about material non-public information?
Hosted in AWS Sydney, with inference restricted to zero-retention models by default. Access is per user and set per share link, every access event is logged, and nothing goes out until someone approves it. Every number is auditable end-to-end.
How long to onboard?
Self-serve, there isn’t one — sign up, upload last month’s files, and you’re working on your own numbers. If we’re doing it with you, 24 hours to first draft from when we have the inputs.

Pick the part of the business that eats the most time

Tell us what your team rebuilds every month, or carries between two systems by hand. We’ll build it on your systems and show you what actually comes out — not a slide about what it could do.

Book a meetingSign up