Support intake
Trigger:
An email describes a problem in the sender's own words.
Agent:
The agent classifies it, creates the ticket in Zammad with the right category and priority, attaches known solution articles and acknowledges receipt.
AI agents for business: assistants that do not just answer, but act
Requests arrive unstructured, someone sorts them and carries out the same step in another system. An AI agent takes over that step: it creates the ticket, starts the workflow and writes the result back - with clear limits, logging and approvals where it matters.
- For teams with recurring workflows
- Tool catalogue instead of full access
- Connect systems, with approvals and logging
The situation
The task is always the same. The intake never is.
An assistant that finds knowledge is half the job. The work happens where a request has to become an action in a system.
Requests arrive unstructured
An email, a PDF, a call asking for a callback: every request is phrased differently, and before anything happens a person has to read and classify it.
The actual task is a routine step
Open a ticket, book an appointment, complete a record, trigger a workflow. Always the same steps, just in a different system with different values each time.
Errors surface late
Whatever was copied wrongly or forgotten only shows up in the follow-on process. By then it can hardly be traced.
Use cases
Where AI agents take work off your desk today
Quote preparation
Trigger:
An enquiry names requirements across several services.
Agent:
The agent pulls the matching building blocks from your catalogue, drafts a quote and submits it for approval - nothing is ever sent without a human.
Onboarding and provisioning
Trigger:
A new employee or customer needs access.
Agent:
The agent triggers provisioning through existing automation (n8n, Ansible, API), checks the status and reports deviations instead of glossing over them.
Data upkeep between systems
Trigger:
Information sits in one system and is missing in another.
Agent:
The agent reconciles, fills missing fields by defined rules and flags conflicts for clarification instead of overwriting blindly.
Scheduling and callbacks
Trigger:
Someone wants a conversation but only gives rough time windows.
Agent:
The agent checks calendars, proposes slots and books after confirmation - including the context of what it is about.
Recurring reports
Trigger:
The same question against the same data every week.
Agent:
The agent pulls the numbers via API, sanity-checks them and files the report - flagging values that fall outside the expected range.
How it works
How an AI agent operates
Not a model roaming freely through your systems, but a controlled chain with defined tools on your existing interfaces.
- 01
Understand the request
The agent reads the request and maps it to a known task. If nothing fits, it asks or hands over to a human - it does not improvise.
- 02
Gather context
Via RAG or direct queries it pulls what it needs to decide: customer data, history, rules. Permissions apply as with the internal assistant.
- 03
Choose a tool
From a fixed list of allowed actions it picks the right one and fills the parameters. What is not on the list does not exist for the agent.
- 04
Request approval
For critical steps - money, external effect, deletions - the agent stops and puts the proposal to a human. You define the threshold.
- 05
Execute and log
The action runs through your API. Request, context, chosen tool, parameters and result land in the log with a trace ID.
- 06
Report back
The agent reports the result - including failures. Silent errors are the most expensive failure mode, so they surface.
Anything with an API. These systems we integrate in projects or operate ourselves; where an interface already exists we use it rather than building a parallel world.
Basics: AI agents: How agents work technically - without the sales lens. →
- Ticketing & serviceZammad, Chatwoot, mail inboxes
- Automationn8n, Node-RED, webhooks, custom scripts
- Business systemsERP and CRM via REST, ERPNext, Odoo
- DataPostgreSQL, APIs, file storage, Nextcloud
- Calendar & communicationCalDAV, Cal.eu, Mattermost, SMTP
- InfrastructureProxmox, Ansible, monitoring APIs
Guardrails
Three decisions before the first line of code
Whose account does the agent act under?
The rule is the requester's passed-through identity: someone who may not open a ticket in a queue cannot do it through the agent either. One service account with every permission creates a union of rights and is only defensible if the agent does nothing beyond what everyone may trigger anyway. Both are logged: who triggered the action and under which identity it ran.
Permissions, identity and approvals →What the EU AI Act requires
Since 2 August 2026 the transparency duties apply: an action triggered by the agent must also remain recognisable as such. After the Digital Omnibus, the high-risk duties under Annex III apply from 2 December 2027; most business agents fall under transparency and documentation. Approvals and logging are therefore part of the architecture. The classification of your case is for your legal advisers.
The EU AI Act for businesses →How the agent reaches your systems
Through your existing APIs or the open Model Context Protocol (MCP). An MCP server built once for your ticket system keeps working when you swap the model. The agent gets a tool catalogue with defined parameters instead of full access; every run is recorded as a trace in Langfuse.
MCP in the enterprise →
Risks
What can go wrong - and what we build against it
The concern about agents is legitimate. It can be answered with architecture, not with optimism.
| Risk | Countermeasure |
|---|---|
| The agent performs an action nobody wanted | It only knows explicitly defined tools. What is not implemented cannot be called - there is no generic system access. |
| Wrong parameters, right action | Parameters are validated against a schema before the API sees them. Invalid values trigger a question, not an attempt. |
| Excessive permissions | When a user triggers the agent, it acts with that user's passed-through identity (on-behalf-of) and inherits their rights. For runs without a triggering user, such as nightly jobs, it gets its own technical account with minimal rights per task. It never gets admin rights. |
| Nobody notices what happened | Every action is logged with trace ID, timestamp, trigger, executing identity and result. The log lives in your system and is part of the handover. |
| An error repeats endlessly | Retries are capped, loops break, and the case lands with a human instead of in an infinite loop. |
| Prompt injection via incoming content | Content from emails or documents is treated as data, not as instructions. Tool selection and approval logic never depend on the sender's text. |
Scope and investment
A defined engagement with verifiable outcomes
Delivery timeframe: 6–10 weeks
First Production AI Agent
Introduce one workflow with defined tools, verified permissions and controlled actions into production.
from €39,900 excl. VAT
Included
- One workflow with up to three confirmed tools/APIs
- Role and identity model, parameter validation and human approvals
- Error paths, retry protection, tracing and agreed end-to-end tests
- Staging and production, technical documentation and training
Scope details
Outcome and verification
- Acceptance of allowed actions, permissions, approvals and error paths against the agreed test set.
Quoted separately
- New AI platform and arbitrary new APIs · Additional workflows, unrestricted autonomy and ongoing operations
Delivery timeframe: 4–6 weeks
Agent / Workflow Expansion
Deliver another use case on an existing suitable agent stack.
from €19,900 excl. VAT
Included
- One additional use case with up to two new tools
- Reuse of existing IAM and tracing
- Workflow and regression tests including approvals and error paths
- Updated documentation and handover
Scope details
Outcome and verification
- Acceptance of the additional workflow and assessment of its impact on the existing application.
Quoted separately
- New platform and fundamental IAM migration · Additional workflows and undocumented interfaces
Terms and notes
Delivery starts once agreed data, access and stakeholders are available. The binding scope and price are defined in your proposal. Hardware, hosting, model consumption, licences and ongoing operations are quoted separately. Follow-on projects are costed against reusable results. We credit the pilot price when you commission the production rollout; there is no requirement to purchase every stage. All prices are net and exclude statutory VAT. The offers are addressed to businesses.
Limits
When you do not need an agent
We would rather build a smaller system that holds than a large one nobody switches on.
The process is always identical
- Then a classic workflow (n8n, script) is cheaper, faster and easier to debug. An agent pays off once the intake is unstructured.
There is no API
- Where only a UI exists, any automation stays fragile. The order then is: interface first, then agent.
The task is rare
- Ten cases a year do not justify automation with maintenance overhead - the human wins there.
The data situation is unclear
- If nobody can say which system holds the truth, an agent only automates the confusion. Clean up first, then automate.
The failure case is not bounded
- As long as nobody can say what an agent is allowed to have broken when it goes wrong, it is too early. First the tolerable damage is defined, then we build.
| Chatbot | Workflow | Agent | |
|---|---|---|---|
| What it does | Answers from existing content. Changes nothing. | Runs a chain of steps: open a ticket, send mail, write a record. | Decides for itself, within set boundaries, which tools to use and in which order. |
| Who determines the path | Nobody - there is no path, only an answer. | You, in advance. The sequence is hard-wired; at most the model fills in fields. | The model, at runtime. You set the boundaries, not the route. |
| What a mistake costs | A wrong answer. Annoying, but reversible. | The sequence breaks or runs through with wrong values. Reproducible and therefore fixable. | A wrong action in the target system - and it was not predicted. That is why approval belongs here. |
- The difference is not the technology but who determines the path - and what happens when something goes wrong.
- Most projects that come in asking for an agent are workflows. A workflow is cheaper, finished sooner and easier to audit; we say so in the first conversation.
Might fit better
Common process patterns have a clear starting point
Not every workflow needs an open-ended agent project. These offers define pilot, price and optional ongoing operations for recurring intake patterns.
AI Opportunity Check
Assess one process, its data, systems, risks and the right automation path before implementation.
from €3,900 excl. VATAI Document Processing
Classify documents, extract and validate content, and hand results to downstream systems.
Automated Order Entry
Validate orders from email, PDF or spreadsheets against master data and prepare them for ERP or CRM.
Ticket and Inbox Automation
Classify, prioritise and route requests and prepare response drafts.
Internal assistant
When knowledge should be found rather than actions executed.
ViewOffer finder
When website visitors should be routed to the right offer.
ViewFrequently asked questions about AI agents
Answers to the most important questions
- A chatbot or assistant answers questions. An agent executes actions: it creates the ticket, books the appointment, queries the database or triggers the n8n workflow - and reports the result back. Defined tools draw the line, not the model.
- from €39,900 excl. VAT for the first production agent; from €19,900 excl. VAT for an additional workflow on a suitable existing platform. Scope, interfaces, permissions and acceptance criteria are confirmed before commissioning.
- We typically plan six to ten weeks for the first production agent once the agreed access, data and contacts are available.
- Through architecture instead of hope: the agent can only execute explicitly approved actions with validated parameters, critical steps require human confirmation, and every action is logged with a trace ID. What is not defined cannot be done.
- Anything with an API: ticket systems like Zammad, workflows via n8n, ERP and CRM systems, databases, email and calendars. Integration runs through defined tools with clear permissions per agent, via your existing APIs or MCP.
- Incoming content is treated as data, not as instructions. The agent selects tools based on the defined task, not on text inside a document - and critical actions require approval anyway. On top of that, such an attempt is visible in the log.
- The stack can run in German data centres or on your hardware with European or local models. Data flows, action logs, retention periods and access rights are documented. Whether the specific process complies with the GDPR also depends on purpose, data categories and your organisational implementation.
- Yes, that is the recommended path. Internal assistant and agent share the same stack (LiteLLM, Langfuse, knowledge base) - tools and approvals are added as a layer, without a rebuild.
Reviews & projects
What clients say about working with us
From a knowledge base with an AI assistant and local AI integration to data sovereignty and ongoing operations.
AI Platform
We were looking for a secure yet easy-to-use alternative to the large American providers and came across WZ-IT. WZ-IT advised us on choosing the right models and implemented the integration for us. Beforehand, we were able to try everything out calmly on a test instance. Today WZ-IT runs Open WebUI for us on our own domain, with a convenient Microsoft login and open-source models hosted by a German provider. Our team can now work productively with AI without resorting to external tools or shadow AI. Working together is straightforward, and questions are answered quickly and clearly. With WZ-IT we also have a partner at our side for further AI projects. (translated)
Carl-Frederic KornHead of E-Commerce & Online Marketing, Krüger Dirndl GmbH
Knowledge Management & AI
The knowledge from many years in the solar industry lived in the head of our former managing director, and it still lives in the heads of our employees today. We wanted to secure it, not as a folder full of notes, but in a way that lets us ask questions in daily work and get reliable answers. WZ-IT advised us from the start and weighed different approaches and knowledge management systems against each other with us before we decided. What was built is a custom knowledge base with an AI assistant that brings together conversations, notes, documentation and manuals: from the platform through onboarding our team to support in operation. Conversations are transcribed and facts are extracted; we review and correct everything ourselves before it is stored, because we do not want to trust an AI blindly. The system runs in our own environment, and the data stays with us. What surprised us was how efficient it was: WZ-IT used existing open-source software and extended it specifically for our requirements, instead of selling an expensive custom build from scratch, as is often usual with IT service providers. Timo and Robin understood our actual problem, preserving knowledge during the generational handover, and implemented it step by step: from the proof of concept through setup and validation with our own data to productive operation, explaining everything clearly along the way. Next, we want to extend this to the knowledge of our employees, so that this does not depend on individual people either.
Joshua von AmelnManaging Director, solarwerkstatt berlin GmbH
Read client reviews
International
Selected projects
- LiveRemote access platform for distributed water treatment plantsABCO Water Systems · Australia · Secure Access & OT
- 81%AWS migration with substantially lower operating costsEVA Real Estate · UAE · Cloud exit & infrastructure
- 98%Provision new sites in minutes instead of hoursnextGYM GmbH · Germany · Automation & IoT
WZ-IT designs, develops and operates infrastructure and software for clients in Germany and internationally. We deliver projects remotely and continue supporting them in ongoing operations after go-live.
Enquiry
Plan a clearly scoped AI agent
Describe the process the agent should support or execute.
- Straight with Timo and Robin - no sales team, no pitch
- A professional assessment of your situation
- Concrete next steps for infrastructure, software or AI
The conversation is free and non-binding.
WZ-IT's advice on our Azure migration was technically sound and completely non-binding right from the intro call - we took away a great deal.
Jakob ÖschlbergerInno7 GmbH
WZ-IT's advice on our Azure migration was technically sound and completely non-binding right from the intro call - we took away a great deal.