IMPROBABLY.ART / BUREAU OF OPERATIONAL EFFICIENCY

Administrative work should not require quite this many humans.


IMPROBABLY.ART is a specialist practice that removes repetitive administrative work. We locate the tasks your organisation performs again and again, establish whether they need to exist, and take out those that do not.

SCOPE / recurring administrative work
METHOD / remove, simplify, automate
OUTCOME / fewer tasks, not faster ones

01 / The problem

THE ADMINISTRATIVE PROBLEM

Repetitive administration accumulates without anyone approving it. A spreadsheet is copied weekly. A report is re-keyed from one system into another. An approval is chased by email because nothing else exists.

Each instance is small. Together they form a standing cost in time, attention and error, paid every week and rarely measured.


TIME / hours spent on work that produces no decision
ATTENTION / interruptions that fragment skilled work
ERROR / transcription faults introduced by hand
DEPENDENCY / procedures held in one person’s memory


02 / Services

WHAT WE REMOVE


R1

Weekly spreadsheet copying

Data exported, pasted, corrected by hand and circulated, week after week. The task is traced, then removed or reduced to the exceptions that need a person.


R2

Re-keying between systems

Figures typed from one system into another, each time introducing the possibility of error. Entered once, passed on automatically, checked only where it matters.


R3

Approvals chased by email

Approvals that approve nothing are dropped. Those that remain are routed once, to the person accountable.


R4

Procedures held in one head

Each retained procedure is written down in plain language, so it no longer depends on one person’s memory or one person’s holiday.


R5

Reports nobody reads

Some work should simply stop. We establish who consumes the output, confirm that no one does, and retire the task.


R6

Confirmations sent by hand

Where a task is stable, rule-based and still necessary, it is handed to software. Staff handle only the replies that need a person.

03 / BEFORE / AFTER

BEFORE / AFTER

BEFORE

EMAIL → EMPLOYEE → PDF → EMPLOYEE → EXCEL → CRM → REPORT


AFTER

EMAIL → SYSTEM → CRM + RECORD + REPORT


Four administrative touchpoints disappeared. The employees did not.

04 / SELECTED INTERVENTIONS

SELECTED INTERVENTIONS


F1

DOCUMENT PROCESSING

Incoming documents arrived by email as PDF attachments and were processed by hand: opened, read, re-typed into a spreadsheet, and forwarded into the CRM. The extraction step was removed and replaced with a rule-based parser that reads the document directly into the record. Staff now review only the exceptions the parser cannot resolve.


F2

CRM ADMINISTRATION

Customer records were updated from several sources, each requiring a separate manual entry into the CRM. Duplicate entry points were consolidated into a single intake, and routine field updates were automated against defined rules. Administration staff retained responsibility for exceptions and for records flagged as incomplete.


F3

RECURRING REPORTING

A weekly report was assembled by hand from several systems, formatted, and circulated by email regardless of whether it was read. The compilation step was automated and the report is now generated directly from source data on a fixed schedule. Distribution was reduced to the recipients who act on it.

05 / THE COST OF ADMINISTRATION

THE COST OF DOING IT MANUALLY

INPUTS


EMPLOYEES


MINUTES / TASK


TASKS / WEEK


HOURLY COST


% AUTOMATABLE


OUTPUTS


ANNUAL HOURS


ANNUAL COST


RECOVERABLE HOURS


SAVINGS


PAYBACK



Figures are estimates based on the inputs provided. They are not a guarantee of outcome.

If the economics do not work, neither should the automation.

06 / THE METHOD

THE METHOD


M1 OBSERVE — watch the task and record frequency, duration and people.
M2 REMOVE — stop work with no purpose.
M3 STANDARDISE — define one procedure.
M4 AUTOMATE — give stable, rule-based work to software.
M5 ESCALATE — route exceptions to a person with context.


Good administrative automation should eventually become invisible.

07 / WE ARE DELIBERATELY NARROW

WE ARE DELIBERATELY NARROW


We do not automate judgement. Diplomacy, leadership, creative direction, sensitive negotiation and genuinely exceptional cases are left where they belong, with people who can exercise discretion.

The work we take on is deliberately limited to repetitive, stable, rule-based administrative tasks. Anything that requires a judgement call stays a judgement call.


P1

TECHNOLOGY IS THE SECOND QUESTION.

The first question is whether the work should exist at all. If it should, it is simplified and standardised before any tool is chosen. Technology is selected last, and only where the task is already stable enough to deserve it.


Good administrative automation should eventually become invisible.

08 / PRINCIPLE

The best administrative task is the one that no longer exists.

Remove before simplifying. Simplify before automating. Automate only what remains, and write it down.

IMPROBABLY.ART / Bureau of Operational Efficiency

Remove before simplifying.

09 / GET IN TOUCH

WHAT ADMINISTRATIVE TASK DOES EVERYONE HATE?

We are not asking for a polished brief. Send the evidence as it stands — the spreadsheet, the chased email, the form nobody updates. We will do the diagnosis.


SEND THE EVIDENCE


← Back

EVIDENCE RECEIVED.

Your submission has entered review. We will examine the process and determine whether it appears suitable for administrative automation.