Module · When the customer is a machine

The AI Ecosystem Readiness Check

Your next integration request will not come from a person filling in a form. It will come from a partner's AI agent, at machine speed, expecting to authenticate itself, read a contract, and transact without a human in the loop. Companies built for human customers are not built for this. This module checks whether you are ready to be part of an agent ecosystem: an API a machine can use, authentication for non-human clients, permission and rate models that assume automation, data contracts partners can rely on, and eyes on the agent traffic itself.

Question 1 of 5 · An API a machine can use

Can a partner's system connect to yours through an API without a human involved?

If integrating with you means emails, spreadsheets, or a portal a person has to click through, an agent cannot use you and will route around you. A documented, stable API is the price of admission to an ecosystem where the other side is increasingly software, not staff.

Question 2 of 5 · Machines can log in

Can a non-human client authenticate to your systems safely?

Human logins assume a person, a password, a second factor on a phone. A partner agent has none of those. Machine-to-machine authentication, keys, tokens, scoped service credentials, is a different discipline, and bolting agents onto human login flows is how credentials leak and access sprawls.

Question 3 of 5 · Limits assume automation

Do your rate limits and permissions assume a machine on the other end?

A human clicks a few times a minute; an agent can call you thousands of times before you finish reading this. Rate limits, quotas, and fine-grained permissions sized for people fall over or over-expose the moment the caller is software that never tires and does exactly what it is allowed to.

Question 4 of 5 · Partners can rely on it

Do your partners have a clear contract for the data they exchange with you?

An agent acts on the structure and meaning of your data exactly as documented, with no human to notice when a field changes shape. A data contract, the format, the fields, the guarantees, and how changes are announced, is what lets a partner's automation depend on you without breaking silently.

Question 5 of 5 · You watch the agents

Can you see and understand the automated traffic hitting your systems?

Agent traffic behaves nothing like human traffic: it is bursty, relentless, and a malfunctioning or hostile agent looks a lot like a healthy one until it does not. Without monitoring built for machine clients, you cannot tell a productive partner from a runaway loop or an abuser until the damage is visible.

For the statistics · one click each

Three questions for the public picture

These do not affect your score. They feed the anonymised, aggregated statistics; groups under 8 respondents are never shown.

How can partners integrate with your systems today?

Manually only

Private API on request

Documented public API

A full integration platform

We do not know

Are partner AI agents already calling your systems?

We are not sure

Just starting

Yes, regularly

No external integrations

How do non-human clients authenticate to you?

They cannot

Shared keys

Per-partner credentials

Managed machine identity

We do not know

Your context

Used to calibrate the report. Company size and sector remain in the anonymized dataset; your email does not.

Company size Sector (optional)

Your corporate email required

We do not accept Gmail, Outlook, GMX, etc. Use your company domain.

Anonymised answers feed aggregated public statistics (groups under 8 respondents are never shown); your email is deleted after confirmation. Details: Datenschutz.

I agree that my anonymized answers (without my email or company name) may be used in Thorsten's published analysis. Tell me when new assessments or question modules go live. One email per launch, opt out any time with one click.