---
title: "AI field reports from the Lab"
description: "16 anonymized field reports by SIMO GmbH on AI in business architecture: starting question, result, feasibility, and lessons learned, by topic."
canonical: "https://simosphereai.com/en/field-reports"
lang: en
---

# AI field reports from the Lab

Every report starts with a question from practice. It shows what was technically feasible, what it means for business architecture, and which lessons we take away. Client names, people, and cost figures have been removed.

## What the status means

- Assessment status: In testing · Built · In use in our own operations · Client case examined.
- Assessment result: Viable · Conditionally viable · Not viable · Pending. “Viable” applies only to the conditions we tested. We draw no conclusions from it for other companies or infrastructures.

## AI governance

Who may know what, who approves, and how to prove it.

- [Who may know what: need-to-know in code, not by discipline](https://simosphereai.com/en/field-reports/need-to-know-in-code.md): How an AI function sees only the data that role and tenant allow, even when a developer forgets a check. (Assessment status: In testing · Assessment result: Pending)
- [AI clearance: content an AI may use only with permission](https://simosphereai.com/en/field-reports/ai-clearance-for-content.md): How to control which AI system may use which piece of company information, separately from who may read it. (Assessment status: Built · Assessment result: Viable)
- [Better too careful: classifying data before every AI call](https://simosphereai.com/en/field-reports/classifying-data-before-ai-calls.md): How we decide, before calling a language model, whether an input is confidential and may leave the organization. (Assessment status: In testing · Assessment result: Pending)
- [Provable, not claimed: a tamper-evident log for AI agents](https://simosphereai.com/en/field-reports/tamper-evident-agent-log.md): How to prove after the fact what an autonomous AI agent did, without anyone being able to change the log unnoticed. (Assessment status: In testing · Assessment result: Pending)
- [Microsoft 365 and AI: the value lies in control, not access](https://simosphereai.com/en/field-reports/microsoft-365-and-ai-control.md): How AI assistants get access to mail, calendar, and documents without an all-or-nothing permission. (Assessment status: Built · Assessment result: Viable)
- [Four eyes on every post: a shared governance foundation](https://simosphereai.com/en/field-reports/four-eyes-on-every-post.md): How AI-assisted social media posts never go out without human approval, across several channels. (Assessment status: Built · Assessment result: Viable)

## Data sources and registers

Connecting external and public data so that provenance and gaps stay visible.

- [Annual accounts as data: making registers machine-readable](https://simosphereai.com/en/field-reports/annual-accounts-as-data.md): How published annual accounts of German companies become comparable figures that can be analyzed reliably. (Assessment status: Built · Assessment result: Viable)
- [Finding public tenders: when portals draw the line](https://simosphereai.com/en/field-reports/finding-public-tenders.md): How public tenders from many portals add up to one picture that matches your own profile. (Assessment status: Built · Assessment result: Conditionally viable)
- [Trademarks in an agent workflow: clean register access](https://simosphereai.com/en/field-reports/trademark-registers-in-agent-workflows.md): How trademark and design data from a European register flows straight into a workflow with AI agents. (Assessment status: Built · Assessment result: Viable)

## Identity and security

Sign-in, keys, and shared capabilities for people and machines.

- [Signing in without passwords, for machines too](https://simosphereai.com/en/field-reports/passwordless-sign-in.md): How static passwords disappear from the mail infrastructure, including for automations such as the ERP system. (Assessment status: In testing · Assessment result: Pending)
- [One mail route: reuse beats a new integration](https://simosphereai.com/en/field-reports/one-mail-route.md): Why new applications send their notifications through one shared, hardened mail route instead of their own providers. (Assessment status: Built · Assessment result: Viable)
- [Secrets do not belong in the process list](https://simosphereai.com/en/field-reports/secrets-out-of-the-process-list.md): How to automate the handling of keys and credentials without exposing them along the way. (Assessment status: In use in our own operations · Assessment result: Viable)

## Operations and observability

Measure, don’t estimate: monitoring, cost, and quality in daily operation.

- [When dashboards go quiet: observability needs cross-checks](https://simosphereai.com/en/field-reports/observability-needs-cross-checks.md): How to tell that monitoring actually monitors, and why “green” is not proof. (Assessment status: In use in our own operations · Assessment result: Conditionally viable)
- [Fewer tokens without losing the line that matters](https://simosphereai.com/en/field-reports/fewer-tokens-without-loss.md): How to cut the cost of tool output in AI agents without handing customer data to external services. (Assessment status: In testing · Assessment result: Pending)
- [Measure accessibility, don’t estimate it: a 32-page audit](https://simosphereai.com/en/field-reports/measuring-accessibility.md): How usable a website really is, for every user, in both color modes, and on every screen size. (Assessment status: In use in our own operations · Assessment result: Viable)

## AI on your own infrastructure

What works locally, what it takes, and when a hybrid architecture holds up.

- [Local AI on existing hardware: an honest feasibility check](https://simosphereai.com/en/field-reports/local-ai-feasibility-check.md): Can an AI assistant run entirely on your own server? A feasibility check for a firm with around 20 workstations. (Assessment status: Client case examined · Assessment result: Conditionally viable)

## From the Lab to consulting.

SIMO does not judge options on vendor claims alone. Assumptions that can be tested technically are examined in the SIMOSphere AI Lab. SIMO GmbH works with your executive team to assess whether an approach that proves viable in the Lab makes sense for your company, and supports the implementation: independent of vendors, with a fixed scope and a decision after every phase.

- [Start the check](https://simo-online.com/en/self-check)
- [Book an initial call](https://request.simo-online.com/en?intent=erstgespraech)
- [All services on simo-online.com](https://simo-online.com/en/services)

Decision Readiness Check: 12 questions, about 3 minutes, instant result. Initial call: 45 minutes, free of charge, by video. Engagements begin with the BISA baseline assessment or the BEIA enterprise assessment. All on simo-online.com.

Rendered version: https://simosphereai.com/en/field-reports
