AI Solutions & Automation
We build AI systems that take over repetitive knowledge work: reading documents, classifying, drafting, checking. Each one is scoped to a clear outcome and built so your own team can understand and maintain it.
AI & Technology Consulting
Elpis Technology builds practical AI systems for enterprises. We help teams take AI from idea to daily use, on whatever systems they already run.
What we do
We build AI systems that take over repetitive knowledge work: reading documents, classifying, drafting, checking. Each one is scoped to a clear outcome and built so your own team can understand and maintain it.
New tools only help if they fit the systems you already have. We connect automation to your existing landscape, whether that is SAP, another ERP or a set of internal tools. Years inside large enterprise systems taught us to handle the messy parts.
Scattered spreadsheets and half-connected tools make every decision harder than it should be. We build pipelines and workflows that keep data clean, checked and easy to reach.
How we work
A short sprint to understand your process, your data and your constraints, and to find out where AI actually pays off.
A concrete plan: what gets built, what it should achieve and how we will measure it.
Short iterations you can see. Working software early, on your real documents and data.
A clean handover to your team, or we keep running it for you. Documented and monitored either way.
Selected work
For clients in a dozen industries. Names withheld, references available when we talk.
Invoice OCR pipeline for an enterprise finance team
Over 70% less manual effort in pilots
Document automation inside SAP for a manufacturing group
Processing time from days to minutes
AI-assisted master data quality checks
Thousands of records validated per run
Smart triage for a shared-services finance inbox
Routed, prioritized and pre-drafted automatically
Elpis, the one that stayed.
About
Elpis Technology is a one-person consultancy, and that is a choice. The person you talk to is the person who builds. Before Elpis I spent more than eight years inside large enterprise landscapes, part of that around SAP, working with everyone from food producers and furniture makers to mining groups, restaurant chains and multi-sector holdings. What that background really gives you is judgment: knowing what survives contact with real operations, and what quietly dies a few weeks after the demo. The practice is based in Istanbul, Turkey, and runs engagements remotely for clients worldwide.
Umut Şanlı
Founder & Principal Consultant
FAQ
Elpis Technology is a boutique consultancy that builds AI systems for enterprises. It designs, develops and operates automation on top of the systems clients already run, such as SAP and other ERPs, covering document processing, prediction, verification and compliance. Every system is built so the client's own team can understand and maintain it.
Elpis Technology has delivered projects in food production, furniture manufacturing, chemicals and industrial oils, food retail and restaurant chains, mining, energy and waste to energy, automotive, textiles, paper and packaging, agriculture, cosmetics and personal care, metals and steel, diversified holding groups, software and SaaS, and public transportation. Client names are shared as references in direct conversations.
Elpis Technology works across SAP ABAP development, S/4HANA migrations and the FI, MM, SD and TRM modules. Typical work includes custom development, integration between SAP and external systems, and AI-assisted automation of SAP-centered processes. SAP experience is a core strength, but none of the work requires an SAP landscape.
Elpis Technology builds agentic AI systems using LangChain, LangGraph and retrieval augmented generation (RAG). Language models are evaluated continuously and used only where they clearly beat simpler methods, and model choices are documented and swappable. Infrastructure runs on Google Cloud Platform with Kubernetes, Terraform and Docker.
Elpis Technology fits enterprises and mid-sized companies that want AI working inside real operations rather than in demos, especially teams running SAP or another ERP with repetitive, document-heavy processes. It is a one-person practice by design, so clients work directly with the consultant who builds the system.
An engagement starts with a discovery sprint of two to three weeks that maps the process, examines the data and produces a concrete design with measurable targets. The build then runs in short iterations, with working software on the client's own documents within the first weeks. It ends with either a documented handover to the client's team or ongoing operation at a fixed monthly rhythm.
Write a short note about your process. You will get an honest answer, even if the answer is that you do not need us.