Least necessary access
The system receives only the data and permissions required for the agreed task.
AI SOLUTIONS · AUTOMATION · DATA
Automating repetitive processes
Document processing, archive search and ERP or CRM integration. We first validate one process on your data, with traceable sources and human review.
What we build
AI for a defined task
01
Custom AI software
We design and build internal systems, assistants grounded in company data, and AI features for existing products.
02
Integrations and automation
We connect APIs, documents, and workflows. Exceptions and critical steps stay under explicit control, and so do permissions.
03
AI analysis and team training
We map opportunities, choose a sensible first step, and train your team safely using the work they actually do.
Who this is for
When this makes sense
- First step
- 2-6 weeks for audit, safe AI workflow, quality checks, and usage rules.
The team loses time processing similar inputs repeatedly.
You can define what a good answer or output looks like.
You need AI with rules, audit, and control, not an unbounded experiment.
Our AI projects · Owned digital product
MojaFirma
Our own product MojaFirma shows document intake, data processing and human review before approval. Screens use synthetic data. A separate pilot measures the value for your business.
- Problem addressed
- Scattered documents, manual data entry and missing context.
- Available evidence
- Our own product, in active development. OCR and source-linked answers are shown with synthetic data. No client savings are measured here.
Pricing
Cost of a small pilot
First step
AI audit
A good fit when: You are unsure which process to automate first.
We map your processes and data and tell you where AI pays off and where it does not. No commitment to continue.
- A map of your processes and data readiness
- Opportunities ranked by payoff and effort
Implementation
Starter
- A good fit when
- You want to test one repetitive process safely.
- Outcome
- One automation, one integration. The cheapest way to find out whether this works.
- Total package price
from 4 500 € + VAT
Timeline: 2 to 3 weeks
I'm interested in StarterWhat you get
- Initial analysis in 2 to 3 days
- Runs alongside the old process
More processes
- A good fit when
- You have several proven processes that connect.
- Outcome
- Three automations and two integrations for the admin work that comes up most often.
- Total package price
8 500 € + VAT
Timeline: 4 to 6 weeks
I'm interested in More processesEverything in Starter, plus
- Accounting and banking hooked up
- Monitoring and alerts
Connected workflows
- A good fit when
- You want to connect documents, payments, and approvals.
- Outcome
- A selected workflow covering documents, payments, approvals and reporting. Analysis confirms the scope and exceptions.
- Total package price
15 000 to 25 000 € + VAT
Timeline: 6 to 10 weeks
I'm interested in Connected workflowsEverything in More processes, plus
- Roles and permissions
- Three months of support
Optional maintenance
Maintenance
A good fit when: You need ongoing technical care after launch.
Security updates and small changes. Larger work uses the technical hourly rate.
- Hosting review
- 2 developer hours a month
BRIEF FOR PRICING
Want a more accurate estimate?
Email the file to rise@rise.sk or paste a share link into the form.
Our AI projects · Grant SaaS
GrantAI
GrantAI connects funding sources, the organisation profile and call documents. One product covers finding a call, verifying eligibility, preparing materials and running delivery after approval.
- Project type
- Owned SaaS product
- Our role
- Our own product. We designed it from scratch, including the data model, the interface, and the AI layer.
Before commissioning the work
Get clear before you commit.
Review the risks, cost and terms of working together.
Which process is worth automating?
Value, data readiness and the risk of a first pilot.
How should we govern agentic development?
Approvals, permission boundaries and evidence before merge.
Will this investment pay off?
Total cost and a realistic return.
What must remain under our control?
Data, access, integrations and changing supplier.
Your data. Your control.
Rules for data and decisions
The solution must fit the way your organisation handles data, permissions, and accountability.
A person at the decision
Critical steps remain under human control with a clearly responsible owner.
Traceable sources
Document-based answers show their evidence, and the system admits uncertainty when support is missing.
Decision brief · AI, automation, and data
A controlled pilot with quality measurement, human approval, and data-use rules.
Repeated input processing takes time, but quality and automation boundaries have not yet been measured.
First paid step
AI audit
We map your processes and data and tell you where AI pays off and where it does not. No commitment to continue.
3 to 4 daysfrom 900 € + VAT
Quick estimate
What does manual work cost?
Three values are enough for a first estimate. The result updates instantly.
Indicative annual cost
€40,250.00 / year
A cost model using 46 working weeks, not guaranteed savings.
What we need from the client
- • Representative inputs and examples of correct and incorrect output.
- • Data-access rules and the person who will approve the output.
Email the file to rise@rise.sk or paste a share link into the form.
How to prepare a project briefWhen it is not the right move
When AI is not the right answer
You expect fully autonomous decisions without human review.
The data is sensitive, but access rules cannot be defined.
Common questions
Before the first AI pilot
You receive a process map, ranked opportunities, and a recommended first step. Each one includes expected value, effort, and risks.
Not always. The approach depends on data sensitivity, required integrations, and your organisation's rules. We design the smallest necessary access scope and explicit permissions.
If the system provides an API or another secure way to exchange data, it can be connected. Before proposing a solution, we verify the technical options, data quality, and human approval points.
We compare extraction with verified data on a representative sample. We track corrections, missing fields and uncertain matches. Writing to another system requires the agreed checks.
We first assess formats, OCR, permissions and archive quality. Answers should cite a specific document. If evidence is missing or contradictory, the system should acknowledge uncertainty.