Skip to content
Koderead

Solution

Business Intelligence

Ask Your Own Company a Question

Most mid-sized companies are data-rich and answer-poor. The information is spread across a CRM, an ERP, a warehouse, a dozen spreadsheets and the head of the one person who has been there longest. The work is making it answerable.

“The data exists somewhere. Getting an answer takes three days and a favour.”

How you know

You probably have this problem if…

The change

What the workflow looks like before and after.

Amber marks a step the system performs. Solid ink marks one a person still performs. The count above each column is how many steps still need someone's attention.

Today

6 human steps

  1. Manager needs a number

    Performed by A person

  2. Emails the analyst

    Performed by A person

  3. Analyst joins three exports

    By hand, in a spreadsheet

    Performed by A person

  4. Number returned

    Three days later

    Performed by A person

  5. Definition questioned

    Work partially repeated

    Performed by A person

  6. Decision already made without it

    Performed by A person

After

3 human steps

  1. Manager asks in plain language

    Performed by A person

  2. Query built against the metric layer

    Performed by Automated

  3. Governed data returned

    Permission-aware

    Performed by System

  4. Answer with source and timestamp

    Performed by Automated

  5. Checks the query if it matters

    Performed by A person

  6. Decides, same meeting

    Performed by A person
  • Automated
  • A person
  • System
  • Customer

The system

What this looks like running in a business.

The same components we build with.

Governed question — management reportingAnswer with source

Asked in plain language

“Which accounts are overdue by more than 30 days, and what changed since last month?”

Query it ran

select account, days_overdue, balance
from finance.receivables
where days_overdue > 30
  and period = current_period

Source: finance warehouse · refreshed 06:15 today · permission-aware

  • Harlow Group · 47 days£62,400
  • Calder Holdings · 38 days£28,100
  • Northgate Partners · 34 days£12,900

Up £19,300 on last month. The movement is Harlow Group, which crossed 30 days on the 11th.

Metric layer — agreed once, in writing

Overdue counts from invoice due date, not issue date, and excludes accounts in a formal payment plan. Two departments defined this differently before the engagement — which is why their numbers disagreed.

It does not generate numbers. It generates a query, runs it against governed data, and shows both so the answer can be checked.

The work does not disappear. It moves — and the steps left are the ones that needed a person.

Koderead · scope of authority

What we build

Four things, in this order.

01

A defined metric layer

Before any AI touches it, we agree what 'active customer' and 'gross margin' actually mean, once, in writing. Disagreement here is the real reason your numbers conflict.

02

Natural-language querying over governed data

Managers ask in plain language and get an answer with the query, the source and the timestamp shown — so it can be checked rather than trusted blindly.

03

Scheduled narrative reporting

The weekly pack assembles itself, including the commentary on what moved and why, ready for a human to review and sign.

04

Permission-aware answers

The system respects the access rules you already have. A regional manager's question returns that region's data.

Boundaries

What stays with a person.

Agreed in writing before the build, not discovered afterwards.

  • Every answer shows its query and source — we do not ship an oracle you cannot audit
  • Statutory and board reporting is always reviewed and signed by a person
  • The system declines rather than estimates when the data does not support an answer
  • Access permissions are inherited from your existing identity system, never re-implemented

Integrations

We connect to what you already run.

Replacing a working system mid-transformation adds risk without adding value. If the existing tool genuinely cannot support the workflow, we say so at assessment rather than three months into a build.

More specific

Narrower services under this capability.

If you arrived looking for one of these by name, each page answers it directly — including where the honest answer is that you need something else.

AI Copilots for Staff

How do we give staff an assistant that knows our business?

Read the answer

Questions we get

Asked and answered.

We do not have a data warehouse. Is this out of reach?

No, but be realistic about sequence. For most mid-sized companies the first phase is consolidating a few sources into one governed place. That is unglamorous and it is the part that determines whether anything later works.

Can it just read our spreadsheets?

It can, and for a narrow question that is sometimes the right answer. It is not a foundation. Spreadsheets have no access control, no change history and no agreed definitions — three things you need the moment the output influences a decision.

What stops it from confidently inventing a number?

It does not generate numbers. It generates a query, runs it against your governed data, and shows both. If the metric layer cannot express the question, it says so instead of approximating.

Where this lands hardest

Industries feeling this most.

Next step

Find out whether this is actually your highest-value fix.

Eleven questions, scored in your browser, with the arithmetic shown. It may well tell you that a different area matters more.

Start the assessment