Seven real problems. Zero code.
Real cases, across different industries. In every one, the rules were written by the people who know the work: with fridAI you don't need to be a programmer.
With fridAI you do it without writing code.
Rules written by those who know
Supervisors, engineers, planners, HR: whoever knows the process configures it from the interface or in natural language.
Nothing to migrate
fridAI connects to the systems you already have. Data stays where it is.
Change it yourself
A parameter, a field, a new rule: no tickets, no quotes, no waiting for a release.
Security and logistics · mid-to-large company
The report that was written twice
A rules engine takes the service report from where it is created to the back-office system, with no retyping and with each operating unit's own rules.
- 0.5 FTE
- per day of transcription eliminated
- +5%
- customers served with the same structure
- Zero
- retyping: data is born where it happens
Across dozens of sites, service reports were filled in on paper and typed into the system by hand, the next day and by someone else. The system learned about events late, and in a rewritten version.
Rules become explicit
The engine applies them at the source
Straight into the system
- The same data is written twice, by different people.
- The rules exist but live in no system.
- The delay between event and record is considered normal.
Contract rules are defined by supervisors from the interface: not a single line of code.
Large-scale retail · accounts payable
Ten thousand invoices, ten minutes
fridAI reads unstructured supplier invoices, by template or in natural language, and loads them into the ERP after a validation stage.
- 10,000
- invoices processed in 10 minutes
- 1 person
- moved from transcription to higher-value work
- Peaks
- absorbed with no backlog
Every supplier sends documents in its own format. One full-time person opened invoices one by one and filled in the ERP: work done well, but costly and unable to absorb peaks.
Template reading
Natural-language reading
Validation before the ERP
- High volumes of documents in formats you don't control.
- People dedicated to data entry, seen as a fixed cost.
- Seasonal peaks the manual process can't handle.
Templates and validation rules are configured, not programmed.
Custom special machinery · engineering office of about 50 people
The know-how that was about to retire
Design rules formalised in an engine connected to CAD: assemblies, drawings and bills of materials are generated consistently.
- −90%
- hours per order
- −90%
- errors and rework
- 160h → 8h
- for a complex custom assembly
Senior engineers knew which solution to use before opening CAD; junior ones made mistakes because that knowledge was written nowhere. And the person who knew thirty years of rules was about to retire.
Formalising the rules
Connecting to CAD
Generation
- Recurring choices are made by one person's experience, not a written rule.
- The rules are reliable but live only in the head of whoever applies them.
- When that person is away, everything slows down.
Rules are written by the engineers, visually: no programming skills needed.
Specialty chemicals · patented formulations
A recipe archive frozen for years
Years of formulations written with different conventions, on a system that could not be updated, brought to a single structure, verified and connected to the ERP.
- 1 month
- archive migration and verification
- Zero code
- the company writes the rules
- Zero legacy
- old system switched off
The data was all there; consistency was missing: every reorganisation had left a different convention. Reading an old recipe required someone who was there when it was written, and updating the system cost more than living with it.
The platform ingests
The data is restructured
The ERP connects
- Historical data exists, in formats layered over time.
- Only those who were there can interpret it.
- Updating the system costs more than living with it.
Constraints, tolerances and raw-material equivalences are written from the interface, with no outside help.
Industrialised construction · components for building sites
Planning with the company's own rules
An engine that plans by the logic the company already applies, learns from historical data and replans when parameters change.
- 1 month
- from kick-off to plan in operation
- No code
- the company changes rules and parameters
- No migration
- no system to replace
There already was a planner, and it was a person: redoing plans all day at every new order and every site emergency. Standard software asks the company to adapt to its model, and the plan gets corrected by hand on day one.
Collect the real rules
The engine applies them to data
The plan is redone
- Someone recalculates plans all day long.
- Stock is kept to be ready for anything, without ever having calculated it.
- Planning and stock are managed by different functions.
Rules and parameters are changed by the planners, with no developers.
Consulting firm · marketing department
One tool for all of marketing
A CRM whose data model is configured by the people using it, with contact search built in, for a team that worked across five disconnected systems.
- One place
- CRM, contact search, activities, calendar
- Configurable
- users change fields and relations
- No tickets
- no developer for a new field
CRM, spreadsheet, lead generation, email marketing, calendar: five tools that did not talk to each other. Every step was an export, a re-import and a hunt for duplicates, because the standard data model did not describe the real work.
The client builds the data model
Built-in contact search
Activities and calendar together
- The same data lives in several places and nobody knows which copy is right.
- The standard model doesn't describe your work: a spreadsheet opens next to it.
- Someone spends hours realigning systems.
Users add fields and relations themselves: no tickets, no quotes.
Pay transparency · companies with 100+ employees
The pay gap, before it becomes an official document
A module that applies Italian Legislative Decree 96/2026 to personnel data: it flags categories above threshold, simulates corrections and produces the compliant report.
- 7 Jun 2027
- the first report
- 5%
- threshold for the joint assessment
- 100+
- employees: where the obligation starts
Almost no company knows its pay gap by category. The aggregate figure is not enough: an average can look fine while two categories are above threshold, and that is where the law looks.
Flag
Simulate
Report
- Headcount exceeds 100 employees.
- The gap is known only in aggregate, or not at all.
- There is no quick way to estimate the effect of a correction.
Categories and assumptions are set by HR, without going through IT.
This is not legal advice: the scope of the obligations must be checked with your employment adviser.
Confidentiality: clients are not named at their request. Every result states its source; figures "reported by the client" come from the client's own measurements. Results refer to the context described and are not a forecast for other contexts.
Is your case the next one?
Tell us about the process: we show you live how to configure it in fridAI.
Book a free demo
We show you fridAI live: connecting sources, asking in natural language, dashboards. Pick a slot on the calendar or write to us.