Intelligence Corner

AI analyst for operations businesses

An analyst that works on its own

It watches the numbers you name and tells you when one moves. It reads your data and can't change it.

Intelligence Corner is an AI analyst for mid-sized operations businesses in the US and Canada that run on a SQL database and have no data team. Once you set up a job in plain words, it watches the numbers you name, tells you when one moves, and runs your approved reports and forecasts on schedule. It can read your data but cannot change it.

Book a demo

We show it working on made-up demo data, then look at whether it fits your setup.

Illustrative mockup of the Watches screen A watch set in plain words, "Alert me when stock drops below 500", with a month of made-up stock levels. The line falls below the 500 line on day 3, where the alert is marked. Day 30, month-end, is marked further along. Watches Illustrative mockup, made-up data You typed "Alert me when stock drops below 500" Checked on a schedule. Runs as fixed code. 500 Day 3: you get the alert Day 30: month-end Day 1 Day 30 Stock on hand, made-up demo business
Illustrative drawing of the Watches screen on made-up data, not a screenshot.

Why problems surface at month-end

The numbers are already in your system: stock on hand, late orders, margin by job. Someone has to pull them and look. In a business without a data team, that someone is you or your controller, when there's time. So a number that slipped on the 3rd gets noticed on the 30th.

A dashboard shows the number to whoever opens it. It doesn't tell you the day the number moves.

What it does without being asked

Intelligence Corner runs the checks you would run yourself if you had the time. You tell it in plain words what to watch and what to send. A watch starts as soon as you type it. Reports, forecasts and KPIs start once you approve them. From then on it works on a schedule and tells you when something needs a look.

  • "Alert me when stock drops below 500"

    Type that and it sets a watch. It checks the condition on a schedule and messages you when it changes. Once the watch is set up, the checks run as fixed code and use no AI calls.

  • "Send me the dispatch summary every day at 9 PM"

    After you approve it, the report goes out every day at that time, built by fixed code.

  • "Forecast weekly orders every night"

    After you approve it, the forecast refreshes nightly. You can pause it, resume it or run it now.

  • "Track this daily"

    It proposes a KPI. Once you approve it, the history builds up and shows as a sparkline, so you can see the trend at a glance.

It also notices new rows in tables you choose, such as new orders or new returns, and can open a case from them. Any alert or finding can become a tracked case with "open a case for this", and it stays open until someone closes it. If one of its own background checks stops, or your database can't be reached, it tells you once.

Book a demo
Illustrative mockup of the KPIs screen Three approved KPIs, late orders per day, stock on hand and margin by job, each with a sparkline of its daily history on made-up data. KPIs Illustrative mockup, made-up data Late orders per day Daily history Stock on hand Daily history Margin by job Daily history
Illustrative drawing of KPIs with daily history sparklines, made-up data.

Ask a question, get the number from your own data

Ask in plain words, such as "sales by product last month", and Intelligence Corner writes a read-only query against your database and answers in the web app. You can keep several chats going, and a status line shows what it is doing.

Ask for a report and it sends back a PDF, Word or Excel file with charts. Ask for a dashboard and it drafts one; you preview it, publish it and share it by a private link that refreshes itself.

Measures you approve, such as how revenue is calculated, are used the same way in every answer, dashboard and report.

Illustrative mockup of the Ask panel The question "sales by product last month" in a chat, a status line, and an answer shown as a bar chart of four made-up products. Ask Illustrative mockup, made-up data sales by product last month Status: running a read-only query Product A Product B Product C Product D
Illustrative drawing of an answer in the Ask panel, made-up data.

It reads your data, and you decide what it may do

It cannot write to your database

Intelligence Corner connects with a read-only database login and a read-only session. It runs only single SELECT queries, each with a time limit and a row cap. Those four layers mean it can read your data but cannot update or delete anything.

You approve what it builds

New KPIs, measures, dashboards, data sources and reports wait in one approvals queue. You approve, edit or reject each one, or approve several at once.

You set how much each routine may do alone

Each scheduled routine has a level you choose: notify you only, send messages on its own, or draft them for your approval. The software's tools enforce that level.

It runs on its own server, and every step is logged

Each business gets its own copy of Intelligence Corner on a dedicated server. Every query, AI call, approval and action is logged, and you can open a step-by-step trace of any run. You can set a daily AI budget; when it is reached, background work pauses. See where your data goes, in the FAQ.

Illustrative mockup of the Approvals queue Three proposals waiting for the owner, a new KPI, a scheduled report and a new dashboard, each with approve, edit and reject buttons. Approvals Illustrative mockup, made-up data New KPI: late orders per day Scheduled report: dispatch summary, 9 PM New dashboard: stock by site Approve Edit Reject Approve Edit Reject Approve Edit Reject
Illustrative drawing of the approvals queue, made-up data.
Book a demo

How setup works

We set it up for you. Your part is answering questions about the business and approving what it learns.

  1. We check the fit. On the demo call we ask which database your system runs on and where your team talks day to day, and tell you straight what works today.
  2. We install it on its own server. One server per business, installed and maintained by us. We create a read-only login on your database for it to use.
  3. It learns your business. It interviews you about how the business runs, then proposes descriptions of your tables, business rules and KPI definitions. You approve each one. If you already have a document describing your database, it can import it as proposals.
  4. You approve the first routines. Pick the numbers to watch and the reports you want. Once you approve them, it starts working on its own.

Will it work with our systems and our team?

It fits a business whose system of record runs on a Postgres, MySQL or SQL Server database. Postgres can be read live; the others are copied in on a schedule, and you get an alert if a copy goes stale or fails.

Your team works with it on WhatsApp, and you get a web app for questions and approvals. Tell us where your team talks and we'll tell you straight what's ready today.

It is not a fit yet if your numbers live only in QuickBooks Online, Excel or SaaS apps, if you need it to write back into your ERP, or if every staff member needs their own login to the web app.

Tell us about your setup when you book a demo.

Questions owners ask

Will it change anything in our system?

No. It connects with a read-only login and runs only single SELECT queries with a time limit and a row cap, so it can read your data but can't update or delete it.

Will it do things on its own without asking me?

Only what you allow. New reports, KPIs, dashboards and data sources wait for your approval, and for each routine you choose whether it only notifies you, messages on its own, or drafts for your approval.

Where does our data go?

It runs on a dedicated server for your business. Data leaves that server only to reach your database, WhatsApp, the AI model providers it uses and, when you ask it to research something, the web page or search it reads.

What systems does it work with?

It reads Postgres, MySQL and SQL Server databases, and it can also bring in CSV files. It doesn't connect to SaaS apps through their APIs yet, so a QuickBooks Online-only setup isn't supported today.

Which tools does my team use it in?

Today the team side runs on WhatsApp, and the owner also has a web app. Tell us which tools your team uses and we'll tell you straight what works today and what doesn't.

Can I trust the numbers it gives me?

Measures you approve, such as how revenue is calculated, are used the same way everywhere it answers. Every run is logged with a step-by-step trace you can open to see how it got there.

How much does it cost?

We go through pricing on the demo call and put it in writing in a proposal.

See it work on demo data

In the demo we type an alert in plain words, trigger it on made-up business data and show you the alert, a scheduled report and the approvals queue. Then we ask about your database and where your team talks, and tell you straight whether it fits today.

It helps to know which system you run and what database it sits on. If you don't know, bring the name of the system.

Book a demo

Opens a booking page for a 30-minute call. Pricing is covered on the call.