TenBeltzA feature to define
You have an idea on the roadmap and need to clarify data, architecture, costs and quality criteria before building.
AI for software companies
Bring AI into your product with an engineering team alongside you.
We help SaaS companies and consultancies define, build and integrate AI systems. Architecture, evaluation and delivery coordinated with your team and your product’s constraints.
Your starting point
Technical support when you need it
A project that needs to move forward
You need implementation and integration capacity, with clear responsibilities and a handover your team can maintain.
A foundation for your consultancy
You want to prepare proposals, start projects and train your team with a reusable technical foundation.
Where we can help
Systems connected to your product
Voice agents and tools
Conversations that check availability and carry out actions through integrations with your software.
Document processing
Structured extraction, classification and validation, evaluated against representative documents.
Search and generation with sources
Assistants and reports that retrieve documents, respect permissions and provide references for review.
Local AI and conversation analysis
Model and infrastructure choices guided by privacy, workload and operational needs.
Documented experience
Engineering explained through projects
From phone agents integrated into a SaaS product to conversation analysis and private AI. Each case describes the contribution and working relationship.
Qamarero · Client project
Phone bookings integrated into the product
Voice agents connected to the restaurant’s CRM, tables and opening hours.
Read the project: Qamarero
Irontec · Technical collaboration
Konect: turning calls into useful information
Transcription, LLM analysis and quality metrics, with a local AI adaptation for RETA.
Read the project: Irontec
Biiak · Co-founded product
A child protection platform with private AI
Case files, documents and reports in a platform with local AI, access controls and traceability.
Read the project: BiiakAI in your product
Explore a few concrete challenges.
AI evaluation for SaaS products
See how a system performs on real cases before choosing a model, expanding access or shipping a change.
Explore the approachVoice agents integrated into SaaS products
A useful voice agent needs to read the product’s live state and complete a task, not just hold a conversation.
Explore the approachPrivate AI for software products
When data, latency or infrastructure control matter, the architecture must decide what runs locally and how it will be operated.
Explore the approachA shared way of working
Clear decisions at every step.
- 01
Understand
We review a real example, the people involved and the systems you already use.
- 02
Define
We agree on scope, data access, expected results and the points that need human review.
- 03
Build and evaluate
We test a first version against representative examples, measuring quality and operating cost.
- 04
Integrate and hand over
We connect the system, document how it works and agree on support and monitoring.
Aritz leads consulting and technical definition. The team develops and integrates the system under his technical direction. Meet the team ↗
Ways to start
Five ways to work with TenBeltz
Start with a technical decision, add development support or prepare your team. Each engagement has its own scope.
AI Gap Analysis · Assessment
A context meeting, technical report and roadmap presentation to decide how to approach your AI feature.
See the serviceAI Project Foundations
Reusable technology choices, planning guides and technical pre-sales criteria for your consultancy’s AI projects.
See the serviceAgent MVP · First version
A working system with reference data, quality checks and a guide to what remains before production.
See the serviceProduction Delivery · Integration
Implementation, integration, operational controls and handover documentation for your team.
See the serviceApplied AI Training
Practical sessions on local models, RAG, agents and evaluation, adapted to your team’s skills and work.
See the serviceBefore we talk
Working alongside your team
Can you collaborate with our development team?
Yes. We agree responsibilities, interfaces, review and documentation from the start. The work fits your product and your team’s capacity.
Can we handle the final delivery ourselves?
Yes. Agent MVP provides a first version with tests and a guide to what remains, so your team can continue. Production Delivery also covers integration through production.
Do you work with software consultancies?
Yes. We can support technical pre-sales, reusable foundations, development and training for your customers’ projects.
How do you assess quality?
We define reference cases and acceptance criteria for the task. We assess responses, errors, latency and consumption where relevant, and agree operational controls.
Who leads the work?
Aritz leads the initial meetings, consulting and technical definition. The team develops and integrates under his technical direction, with responsibilities agreed for each project.
Next step
What does your team need to move forward with AI?
Tell us about your product, the AI system and its current stage. We will review the fit and outline a concrete next step with you.
Prefer a conversation?
Book a 15-minute phone callWe will discuss your goal, available tools and data, and whether we can help.
Want to assess your AI project first?
Technical AI diagnosis · 2 minAssess evaluation, observability and control. Get your priorities in a downloadable report.
Technical AI diagnosis