AI analytics and forecasting
Every business already holds the data needed to answer what will sell next season, which customer is about to stop paying, and which lead is worth calling first. We build the models that answer those questions from your own history, and put the answers where the decision actually gets made.
Is this you?
- Stock is planned on last year’s figures and a feeling about the season.
- You find out a customer has stopped paying when the receivable is already ninety days old.
- Sales call leads in the order they arrived, not in the order worth calling.
- Every report is somebody’s personal Excel file that nobody else can reproduce.
- A simple question about your own numbers takes two days and comes back disputed.
What we build
- Demand and stock forecasting, by item, branch and season
- Receivables risk — which invoices are likely to go bad, early enough to act
- Customer churn and repeat-purchase prediction
- Lead scoring, so sales calls the ten most likely first
- Anomaly detection in transactions, stock movement and expenses
- Plain-English querying — ask a question of your own data and get the number
- Dashboards that show the decision, not forty charts nobody opens
- Honest accuracy reporting, measured against what actually happened
What this is not
This is not a dashboard product. A chart that tells you what happened last month is reporting; this is about what is likely to happen next, and what to do about it.
Businesses with at least a year or two of reasonably clean transaction history. If your data is scattered or unreliable, the honest first step is fixing that, and we will say so.
- A written scope you approve before any code is written
- A price agreed against that scope, not discovered later
- Source code, database, documentation and IP yours on final payment
- Training for your team, and close support through the first month
AI software development
AI built into the system you already run, doing work your team currently does by hand.
ERP software
One system for production, purchase, stock, sales and accounts, instead of five.
Business process automation
The approvals, reminders and hand-offs your team currently does by hand.
Questions people ask about this
Do we need clean data before any of this works?
For document reading and for agents, no — those work on documents and conversations as they arrive.
For forecasting and scoring, yes, and this is where we would rather lose the project than take it. Prediction needs one to two years of reasonably consistent transaction history. If your data is scattered across systems, full of duplicates, or was entered inconsistently, the honest first step is fixing that. A forecast built on bad history is worse than no forecast, because people act on it.
How accurate is it, and what happens when it gets something wrong?
Nobody honest will quote you a single accuracy figure before seeing your actual documents and data. What we do instead is measure it on your real material during a pilot, and tell you the number we got.
The design matters more than the number. Every system we build sets a confidence threshold: anything below it goes to a review queue for a person rather than being posted silently. Anything touching money gets checked. You will always be able to see what the system decided, what it was unsure about, and what a person changed.
Where does our data go when you process it with AI?
AI processing is performed through OpenAI’s API. Your data is sent to that API for the specific task, and the result comes back into your system.
Before any project involving your data begins, we confirm the current provider data-handling terms in writing — retention, whether anything may be used for model training, and where processing takes place — so that it forms part of your agreement rather than an assurance in a sales call. If those terms do not work for your business, the honest answer is that this is not the right project for you, and we will say so at that stage rather than later.
What is the sensible way to start with AI?
One narrow process, measured. Pick the task that is highest volume and most repetitive — usually document entry or first-line customer questions — and run it on your real material alongside the existing manual process for a few weeks. You get an accuracy figure from your own data rather than a claim from a vendor, and the cost of finding out is small.
What we would advise against is an open-ended “AI transformation” programme. Those are expensive, slow to show anything, and usually end with a pilot nobody deployed.
Tell us what isn’t working
A first conversation costs nothing and commits you to nothing. Bring the problem — the mess of spreadsheets, the report nobody can produce, the process that breaks every month-end. We will tell you honestly whether custom software is the answer, and roughly what it would take. Sometimes the answer is that you do not need us. We will say so.