Skip to content

Technology | People | Progress A Brighter Tomorrow

Get in Touch
Logic Networks transformation journey from strategy to lasting impact

Automation

Remove repetitive work.Achieve more.

Connect processes and give people time back with Power Platform, APIs and AI.

  • Fewer Manual Tasks
  • Fewer Errors
  • Connected Systems
  • More Time for People

Business process automation for the work nobody should do by hand

Every organisation has a task somebody does every week that a computer could do in a second.

Logic Networks builds those pieces of automation for organisations across London and the rest of the UK. Not large programmes, but the specific steps where information is copied from one place to another, a form is retyped into a system, or somebody chases an approval by email.

The work is deliberately small. One task at a time, measured before and after, using the tools you already pay for wherever possible.

We will also tell you when a task should not be automated. A step that changes every month, or one where judgement is the point, is better left to a person.

Automating a bad process makes the bad process faster.

Before we build anything, we ask why the step exists. A surprising number of them are there because of a system that was replaced years ago.

  • One task at a time, with the time saved measured
  • Built in tools you already own where possible
  • Documented, so it is not dependent on us
  • Monitored, because an automation that silently stops is worse than none

What is included

Six kinds of work, from a form to a full approval chain.

Most organisations start with the first or second. They are the ones where the saving is visible within a fortnight.


Moving information between systems

No more copying by hand.

Where somebody currently exports from one system and types it into another, we build the connection instead.

The important part is what happens when something does not match. A connection that silently drops a record is worse than a person doing it slowly, so every one we build reports its exceptions to somebody.

  • Integrations
  • Scheduled transfers
  • Field mapping
  • Error handling

Forms and the data behind them

Filled once, used everywhere.

Paper and email forms replaced with something that puts the information straight into the right place, with a record of who submitted it and when.

Holiday requests, expense claims, new starter details, equipment requests and incident reports are the common ones. Each removes a small amount of work from a lot of people.

  • Online forms
  • Approvals
  • Records
  • Notifications

Approvals and handovers

The work moves without being chased.

A request that goes to the right person, reminds them, escalates if nothing happens and records the decision.

Most delay in an organisation is not work, it is waiting. Making the waiting visible usually reduces it without anybody being asked to work faster.

  • Approval chains
  • Reminders
  • Escalation
  • Audit trail

Documents and correspondence

A first version produced in seconds.

Documents produced from information you already hold, rather than typed again into a template: contracts, letters, certificates and routine reports.

Where the wording varies with judgement, we generate the first version and leave the person to edit it. That is the split that works, and it saves most of the time anyway.

  • Templates
  • Document generation
  • Letters
  • Reports

Reading information out of documents

Invoices and forms, without the retyping.

Invoices, forms and scanned documents read automatically into structured fields, with anything uncertain sent to a person to check.

The rate is never perfect and does not need to be. If nine in ten are handled automatically and the tenth is queued for a person, the saving is already large.

  • Invoices
  • Forms
  • Scanned documents
  • Exception handling

Monitoring and keeping it working

Because an automation that stops is invisible.

Every automation watched, so a failure raises an alert rather than being discovered three weeks later by a customer.

We also document how each one works and hand that over, so you are not dependent on us to understand your own process.

  • Alerts
  • Logs
  • Reviews
  • Documentation

How it works

Three arrangements, and the first one is intentionally tiny.

Automation goes wrong when it is bought as a programme. One task, built and measured, tells you whether the next one is worth doing.

Tell us the task somebody in your organisation complains about every week and we will say whether it is a good first candidate.

  • One task

    A single step automated, with the time before and after measured. Small, quick, and the honest test of whether this is worth pursuing.

  • A process review

    We follow one process end to end, mark where the time goes and come back with what to automate, what to simplify and what to stop doing.

  • Ongoing work

    Automations built, monitored and maintained as part of managed IT, with changes made as the process changes.

Who it is for

This work pays for itself fastest in organisations where a small number of people do a large amount of repetitive administration, and where hiring another person is not the answer.

It also suits organisations facing a volume increase without a budget increase, which describes most of the public and charitable sector.

  • Businesses where order or invoice volume has outgrown the admin team
  • Charities and non profits with grant reporting and beneficiary records to maintain
  • Education organisations with forms, consents and returns to process
  • Organisations where two systems are joined by a person and a spreadsheet
  • Teams where an approval routinely waits a week on somebody’s desk
  • Anyone about to hire for a role that is mostly copying data

What we need from you

Three things, and the first one is half an hour that saves a month.

The person who does the task is the most important contributor to this work. Automating a process from a description written by somebody who does not do it is the usual way this fails.

Time with the person who does it

Half an hour watching them do the task. What they actually do is always different from the written procedure, and the difference is where the exceptions live.

Access to the systems involved

Including from the supplier, where a system belongs to somebody else. Some charge for connections and some do not allow them, and both change the plan.

An honest estimate of the time

How long the task takes and how often. Without a before figure there is no way to say afterwards whether it worked.

What we usually find when we follow a process

These come up most often, and two of them mean the answer is not automation at all.

  • A step that exists because of a system replaced years ago
  • Two people doing the same check without knowing about each other
  • A spreadsheet acting as the bridge between two systems
  • An approval that nobody has ever refused
  • Information retyped because a connection was never built
  • A report produced monthly that nobody reads
  • Exceptions handled differently by each person who touches them
  • The whole process resting on one person’s knowledge

Our steps for working together

Every automation we build follows the same six steps, whether it takes a day or a fortnight. The first two are the ones that decide whether the rest is worth doing.

  1. Watch the task being done Once, in full, by the person who does it. Including the parts they do not think are worth mentioning.
  2. Ask why each step exists Some steps disappear at this point, which is the cheapest improvement available and needs no software at all.
  3. Measure the before How long, how often, how many mistakes. Written down, because memory is generous about these numbers in both directions.
  4. Build it small The common path first, running alongside the manual process rather than replacing it.
  5. Handle the exceptions Anything the automation cannot manage goes to a named person, visibly. This is the part that makes it trustworthy.
  6. Measure the after, then hand it over With documentation, so your team can change it without us.

How we decide what to automate first

We look for tasks that are frequent, boring and stable. Frequent so the saving is real, boring so nobody objects to losing them, and stable so the automation does not need rebuilding every quarter. A task that is rare, or that changes constantly, stays with a person no matter how tedious it looks.

What we will not automate

  • A decision about a person Recruitment, eligibility, assessment and anything with a legal duty behind it stays with a human being.
  • A step where judgement is the work If the value is in somebody thinking, removing the thinking removes the value.
  • A process nobody understands We will map it first. Automating a mystery produces a faster mystery.
  • Something that changes every month The maintenance would cost more than the task.
  • A task that should simply stop Sometimes the honest recommendation is to delete the report rather than to automate it.

Working out what an automation is worth

The arithmetic is simple and it is worth doing before anybody builds anything, because it settles most arguments about whether a piece of work is justified.

Take how long the task takes, multiply by how often it happens in a year, and you have the hours. Multiply the hours by a sensible internal cost and you have the annual figure. Then compare it with the cost of building the automation and the cost of keeping it running.

A task taking twenty minutes, done twice a week, is about thirty hours a year. That is worth automating if the build is measured in days and not in weeks. The same task done twice a year is not worth touching, however irritating it is.

Two things are usually left out of that calculation and both matter. The first is the cost of the mistakes: a figure retyped wrongly can cost far more than the time it took to type. The second is the cost of the knowledge sitting in one person, because a manual process is a process that stops when that person is away.

We produce this calculation for each candidate before recommending anything, and we include the ones where the answer is no.

What we will not claim

Automation does not usually reduce headcount in a small organisation, and we would not sell it on that basis. What it does is remove the work people resent, reduce the errors that come from retyping, and let a team absorb more volume without a proportional increase in cost. We will give you the measured figure rather than a percentage from a brochure.

What good looks like

Nobody copies information between two systems by hand. Approvals move without being chased. Exceptions arrive at a named person with a note about why. Every automation has documentation and an alert if it stops. And you can say, for each one, how much time it saves, because it was measured before and after.

Related services

Automation is usually the second piece of work rather than the first, because it needs the systems and the information underneath it to be in order.

Where you already own a tool that can do this, we build in it rather than proposing another subscription.

Articles and news on automation

Notes on which tasks are worth automating and which are not, with the measurements rather than the promises.

Talk to us about automation

Tell us the task somebody in your organisation complains about every week. We will watch it being done, tell you whether it is worth automating and give you the measured saving rather than an estimate.

No obligation, and no sales call unless you ask for one.

Common questions

What should we automate first?

A task that is frequent, boring and stable. Frequent so the saving is real, boring so nobody minds losing it, and stable so the automation does not need rebuilding every quarter.

How do you work out whether it is worth it?

Time taken, multiplied by how often it happens in a year, gives the hours. A task taking twenty minutes twice a week is about thirty hours a year, which is worth automating if the build takes days rather than weeks. We produce that calculation before recommending anything, including when the answer is no.

Will this replace jobs?

In a small organisation, rarely, and we would not sell it on that basis. What it does is remove the work people resent, reduce the errors that come from retyping, and let a team absorb more volume without a proportional increase in cost.

What will you not automate?

A decision about a person, a step where judgement is the work, a process nobody understands yet, something that changes every month, and a task that should simply stop being done.

What happens if an automation breaks?

It raises an alert to a named person. An automation that stops silently is worse than none at all, which is why monitoring is part of every one we build.

Do we depend on you afterwards?

No. Each automation is documented and handed over, so your own team can change it. Where it is built in tools you already own, it stays yours entirely.