AI agents & automation | Connected to your own systems | Newblend
Newblend Book an introduction
AI 02

AI agents & automation

An agent receives an instruction, consults your systems, makes a decision within your rules and carries out an action. Not a chatbot that invents answers, but software that takes over work.

In short
  • Five steps per agent: trigger, decision, system integration, action, feedback.
  • You decide which actions the agent may take on its own and which pass by a human.
  • Every step is logged, so you can demonstrate at an audit what happened.
  • A first agent for one clearly bounded process is usually live within a few weeks.
  • The agent runs on your own systems and data; company data is not used to train external models.
Last updated: 5 August 2026
Examples

What we build agents for

Processes with a lot of repetitive manual work, fixed rules and a clear owner are the best fit.

Incoming post and email

Classify documents, extract the data and file it in the right system. When in doubt, to a human.

Quotes and requests

Turn a request into a draft quote based on your price list and earlier orders. The salesperson checks and sends.

Service desk and tickets

Categorise reports, merge duplicate tickets and handle standard questions with an answer from your own documentation.

On- and offboarding

Arrange accounts, licences, permissions and devices when people join or leave, with an audit trail.

Reporting

Compile recurring reports from several sources, with a note on what has changed.

Make knowledge findable

Search your own contracts, procedures and case files, with source references so the answer can be verified.

Process

From analysis to production

  1. STEP 01
    Process analysis

    We record the steps, the exceptions and the decision rules as they really are, not as the procedure describes them.

  2. STEP 02
    Business case

    Expected time saved, integrations needed and costs. Plus the outcome of the GDPR and AI Act assessment.

  3. STEP 03
    Build and pilot

    The agent first runs alongside the existing work. We adjust decision rules based on real cases.

  4. STEP 04
    Production and management

    Live, with an owner inside your organisation, logging and management under the same SLA as the rest of your IT.

Preconditions

What we arrange up front

  • Permissions: an agent sees what the user may see, so the permission model is put in order first.
  • Boundaries: which actions may the agent take on its own, and when does it escalate to a human.
  • Logging: every step recorded, with retention periods that match your policy.
  • Ownership: someone inside your organisation maintains the decision rules and reviews exceptions.
  • Fallback: if the agent fails, the work has to be able to continue manually.

Frequently asked questions

What does having an agent built cost?

The investment depends on the number of system integrations and the complexity of the decision rules. We start with a fixed price for the analysis, so the business case is clear before anything is built.

What if the agent makes a mistake?

It works within fixed rules and logs every step. When in doubt it escalates instead of deciding. During the pilot it runs alongside the existing work, so mistakes have no consequences.

Can you integrate with our ERP or CRM?

Usually yes, provided the system offers an interface. We check that first during the analysis, because without an integration it becomes manual work with extra steps.

Does this work with our own documents?

Yes. An agent can search your contracts, procedures and case files and points to the source, so the answer can be verified.

Read on

Want to talk it over?

Book a half-hour introduction. We are happy to think along, even if you are not sure yet what you are looking for.