AI Engineering
We develop production AI solutions that understand documents, use information from your systems, and reliably reduce recurring work.
Location, Hosting & Memberships
Made in Switzerland · Swiss, EU, or global hosting options · Member of Swico & swissICT
Whether you need to identify the right opportunity or already have a solution in mind, we take the work into a sound pilot project and onward into production.
Not sure what to build?
Already know what to build?
Client testimonial
“The collaboration with Tanner Lab and eventlokale.ch was outstanding. Their ability to visualize entropy and consistently reduce it is particularly impressive. They are highly skilled, clearly structured, focused, and at the same time dynamic and flexible. Problems are addressed directly and solved in a solution-oriented manner. It is a partnership that creates order and enables real progress. - Simply outstanding!”
We automate clearly scoped tasks and workflows where AI can reliably create value using your data and systems.
Classify incoming messages (support, sales, internal) and route them or draft a prepared response. Humans approve before sending.
Draft emails or chat replies that pull context from your CRM, ticketing, or internal knowledge. A human reviews and sends. Saves real time on repetitive correspondence.
Digitise documents reliably using modern Vision-Language Models. Far better than legacy OCR on real-world inputs: scans, photos, mixed layouts, handwriting.
Answer targeted, well-scoped questions by pulling from multiple systems (CRM, accounting, SaaS databases, internal docs). RAG done properly, with citations.
Agentic RAG combines retrieval, tool use, and multi-step reasoning. Useful when a system needs to search your knowledge, call APIs, compare sources, and return cited answers.
Multi-step business processes (read invoice, extract fields, match to PO, route for approval) with AI handling the fuzzy parts and humans approving the consequential steps.
A structured first step
In 2 to 3 weeks, you will know whether and how AI can improve one important process in your business, and what to do next.
Tanner Lab Process-First Framework
We start with the process in the selected business area: how work flows, where people get slowed down, and what outcome matters. Where the process is ready, we recommend a contained AI solution. Where it is not, we recommend redesign first, so automation strengthens the right process rather than scaling the wrong one.
Initial conversation. Define the business area, desired outcome, current concerns, and the people involved in the work.
Conversations with the people involved in the work. Understand how work actually happens, where time is lost, and where friction appears.
Problem identification. Describe bottlenecks, recurring work, and the improvement opportunities worth assessing.
Automation and Reengineering Analysis. Each relevant opportunity is considered through three questions:
Written report. Executive summary, current situation, identified problems, build-now or redesign-first assessment, recommendation, and next steps.
Presentation and discussion. Walk through findings, answer questions.
AI Potential Assessment
Report
1. Executive summary
3.2 Two clear buckets:
Ready for a contained tool, with start-here recommendation, roadmap, and rough ROI.
Not ready for automation yet. We name the issue and recommend the redesign path.
3.3 Recommended next steps
Describe your idea or recurring workload. We will respond with a pragmatic next step.
Tanner Lab delivered the RAG/LLM component: multimodal knowledge retrieval to match a photo to the right technical documentation. The overall PoC was developed in collaboration with Hatimeria.
Automation designed, implemented, and deployed to production, complemented by advisory work on digital product development and further processes.
AI is most useful when the underlying process is understood first. If a workflow is unclear, slow, or suboptimally designed, putting a model on top can make the problems harder to see. Our framework is built on Business Process Reengineering, Michael Hammer's HBR work from 1990, updated for the AI era. The discipline is simple: measure, analyse, redesign, then automate.
We're explicit with you about both possible outcomes. If your process needs rethinking before any AI lands on it, we'll say so, even when that means we don't get the immediate build engagement. When we do recommend building, you'll know we mean it.
Methodology: Tanner Lab Process-First Framework. Grounded in Business Process Reengineering (Hammer, 1990) and modern process-aware AI practice.
Describe your idea or recurring workload. We will respond with a pragmatic next step.