Business visibility
What is happening, where we deviate and why.
- Connecting data from different systems
- Profitability by product or customer
- Write-off analysis
- Detecting deviations
- Automatic alerts
Business processes, data and software solutions
We start from how your business actually works and build solutions that improve everyday work and decision-making.
We help companies improve how they plan, decide and carry out their work. We turn business problems into clear rules, data models and working solutions, and we verify their effect.
Example decision specification
Confirming an incoming order
- Metric
- Time from receipt to confirmation, share of corrected orders
- Decided by
- Sales representative
- Rules
- Customer and item codes exist in the ERP. Quantity is in stock or in the plan.
- Exceptions
- New customer. Delivery date less than 48 hours away.
- Calculated
- Available stock, earliest delivery datesystem
- Estimated
- Likelihood of a supplier delaymodel
- Judgement
- Exceptions are approved by the representativehuman
- Value
- Less manual checking, fewer wrong confirmations
OI-SPEC-01Rev. A
Connecting AI subscriptions
Your AI assistant, connected to your business
No additional paid application. Employees get the new tools inside the assistant the company already pays for.
Your employees already use ChatGPT, Claude or Gemini. Through MCP servers and plugins, we connect them to your ERP, documents and internal rules. The assistant then works with the company's actual data and performs only permitted actions.
This is one of several options. Where a purpose-built application or automation without a language model fits better, we choose that.
Your existing subscription
- ChatGPT
- Claude
- Gemini
OI connection
- MCP server
- Access rights
- Business rules
- Action log
Your systems
- ERP
- Documents
- Internal rules
Reads
Stock levels, open orders, price lists and contracts.
Acts
Prepares a quote, enters an order or opens a task, within the rights it has been given.
Logs
Who asked for what and what the assistant did.
Access rights, traceability and data processing are set up in line with the GDPR and the EU AI Act.
Talk to us about your caseAreas
What we improve
We organise our work around business needs and combine the underlying technologies to fit the problem.
Planning and decision-making
What the company should do given its goals and constraints.
- Demand forecasting
- Inventory optimisation
- Work scheduling
- Production plans
- Scenario simulations
Getting work done
How to reduce manual work, errors and delays.
- Processing orders and documents
- Connections between systems
- Approval workflows
- Purpose-built applications
- AI assistants connected to processes
Tools
Analytics · Statistical models · Mathematical optimisation · Automation · Purpose-built software · Language models
We choose the technology to fit the problem. Some projects don't need a language model at all; in others it is a useful part of the whole.
Examples
How it works in practice
Sales assistant
A sales representative asks their AI assistant (Claude, ChatGPT or Gemini) whether 400 units can be delivered to a customer by Friday. Through the connection, the assistant checks stock and open orders in the ERP, applies the delivery rules and drafts a quote for approval.
- Claude / ChatGPT / Gemini
- OI connection
- ERP
- Approval
Profitability and write-offs
Data from the ERP and the point-of-sale system is combined into a single overview every day. The system raises an alert when write-offs or margin deviate from the usual.
- ERP + POS
- Analytics
- Alert
Voice order entry
A dispatcher speaks the order. A language model understands what was said, program logic checks codes, quantities and conditions, and the connection to the business system prepares the order.
- Speech
- Language model
- Validation rules
- ERP
Automated equipment check
Before a shift, a worker inspects a machine on their phone using a checklist and attaches photos. The system checks that the inspection is complete, notifies maintenance if there is a defect and stores a record for internal and external audits.
- Mobile check
- Rules
- Maintenance alert
- Record
Method
Formalise first
Before we choose a technology, we work through five questions with you.
Formalisation has value in itself. Sometimes it shows that the company needs a better process, a single source of data or clearer responsibility. Other times it makes a concrete tool possible.
- MetricWhat do we want to improve, and how will we measure it?
- DecisionHow is the decision made today, and who makes it?
- RulesWhich rules, constraints and exceptions must the solution respect?
- JudgementWhat can we calculate, what can we estimate, and where is human judgement needed?
- ValueWhat does the solution bring: money saved, easier work, less responsibility to carry or more convenience?
Process
From the problem to a verified effect
We work in three consecutive services. Each one has a clear result.
Business review and solution plan
Review of a specific process, formalisation of the problem, feasibility assessment and choice of intervention. The result is a written plan you can also use on its own.
from €490
- OP 010
Understand the problem
We look at the process as it actually runs, and at the people who make decisions in it.
- OP 020
Define metrics and values
We set the baseline, the success metric and the values before we start building.
Development and rollout
A bounded project with clear acceptance criteria, use in the real process and handover to the people responsible.
from €690
- OP 030
Build the solution
We build the solution and connect it to your existing systems.
- OP 040
Put it to work
The solution runs in the real process and the people responsible take it over.
Monitoring and improvement
Maintenance, quality monitoring, rule adjustments and a regular before-and-after comparison.
from €90/month
- OP 050
Monitor and adjust
We monitor the quality of the solution and adjust the rules when the business changes.
- OP 060
Compare before and after
We regularly show how much time, money and how many errors there were before, and how many there are now.
Hours of manual work per week
Before10After7−30%
Prices exclude VAT.
Book an introductory call Free of charge, 45 minutes, on Google Meet.
Compliance and security
Solutions that respect the rules of your country and industry
We treat regulation as part of the specification, just like business rules. In the first step we determine which rules apply to the solution and document them: risk classification, impact assessment, security policies and access traceability.
OI-REG-01
Regulations and standards we work with
- European Union
- GDPREU AI ActMDRNIS2Data ActDORACyber Resilience ActEuropean Accessibility Act
- National laws
- ZVOP-2 SIDSG ATBDSG DEZZPL RSNational rules HR BA MK ME
- Outside the EU
- revDSG CHUK GDPR UKHIPAA USCCPA US-CA
- Standards
- ISO/IEC 27001ISO/IEC 42001ISO 13485IEC 62304
- Industry rules
- Labour law in work schedulingOccupational health and safetyRetention of business records
The list is not exhaustive. For each project we check the rules of the country and industry where the solution will operate.
Frequently asked questions
- Can you connect our ChatGPT, Claude or Gemini to our ERP and documents?
Yes. Through an MCP server or plugin, the assistant your company already uses gets access to selected data and permitted actions. You define the permissions, and every action is logged.
- Is our business data used to train public AI models?
No. We only use business plans and APIs for which the model providers (OpenAI, Anthropic, Google) do not use data to train their models. Data retention is limited according to the provider's terms. Where a provider offers zero data retention, we use it.
- What is MCP (Model Context Protocol) and why do you use it?
MCP is an open standard through which an assistant securely accesses your systems, such as an ERP or documents. The assistant only gets the data it needs for an answer, without copying it into a separate database. Data goes to the model for processing, not for training, and you decide on access.
- When does our investment in AI automation pay off?
We plan the rollout so that the initial investment pays off in less than a year. After the rollout, monthly costs with measurable results are lower than before, or the agreed non-financial goals are reached, for example faster response or fewer errors. We prepare the calculation during the assessment.
- How do we introduce AI in the company without breaching the GDPR?
By not sending personal data to the model provider unless necessary, and by having a legal basis, a data processing agreement and EU hosting in place for every processing activity. We connect the assistant to your systems through an MCP server that determines which data it can access and logs every action.
- Where should we start with AI in our company?
With one process that is repeated often, takes a lot of time and has clear rules. We first describe it as a decision (criterion, rules, exceptions) and only then choose the tool.
- Can we use ChatGPT with confidential company data?
Yes, but not with a consumer account. Use a business plan or the API with a data processing agreement and retention settings, and set up access to company data through a connection that defines permissions and logs actions.
- Can we start with a small trial?
Yes. A trial on one process with your data costs from €690 to €890 and shows whether the rollout pays off before you decide on more.
Prices exclude VAT.
Contact
Book an introductory call
45 minutes, no obligation. Tell us which process or decision you want to improve.
Book a time Free of charge, 45 minutes, on Google Meet.
- Phone
- +386 31 651 802
- Based in
- Na Pušo 7, 2000 Maribor, Slovenia
Write to us
Describe the process or decision you would like to improve. We will reply and suggest a time for an introductory call.
Send an emailThe button opens your email program. If it does not open, write to info@operative-intelligence.com. We use the details in your message only to reply.