What Anthropic is.
Anthropic is an AI company. It makes Claude, a family of AI models, and published the Model Context Protocol (MCP), an open standard that lets an AI assistant connect to tools and data.
Claude is good at reading, drafting and organising. It is not a technician, and at ALCO it is never treated as one.
How it works for us.
- 01
Not a lot, and only small things
We do not use AI a lot in house. Where we do, it is minor: a first draft of a note, sorting an inbox, a routine check of a log. A person reviews every output before it goes anywhere.
- 02
A person is always accountable
Nothing Claude produces reaches a client, a server or a ticket without a human at ALCO reading it, approving it and signing their own name. Every use is logged.
- 03
Built into client tools on request
Our headless CMS exposes an MCP server, so a client's own Claude can propose website changes into a staged preview for the client to approve. We build similar features into client applications when asked.
Our role.
What ALCO is to Anthropic, and what it is not.
- 1
API customer and developer
ALCO pays for Claude through Anthropic's API and builds on it for clients. We claim no Anthropic partner tier or certification.
- 2
The humans in the loop
Our role is to decide where a model helps, keep it on a short lead, and take responsibility for the result.
- 3
Your data stays yours
Client information is never used to train a model, by us or by anyone we pass it to. That is written into how we work.
What it means for you.
People answer, people fix
When you call ALCO, a person answers and a person does the work. AI is not running your support.
AI features you control
If you want Claude in your own site or app, we build it so it proposes and you approve. It cannot publish on its own.
Honest about where it is used
Ask us where AI touched anything we did for you and we will show you the log.
Questions people ask.
Does AI handle my support tickets?
No. A person reads your request and a person does the work. At most, AI helps us sort or draft, and a person checks it before anything happens.
Is our data used to train AI?
No. Client information is never used to train a model, and we do not send it to any AI service without your written agreement.
Can you build Claude into our website or app?
Yes, when it is useful. Our headless CMS already lets your own Claude propose changes for you to approve. We build it so it proposes and you decide.
Why work with Anthropic at all?
Claude is careful, it follows instructions well, and MCP is an open standard that lets our clients bring the AI subscription they already have.
- The client area
Everything ALCO does for you, in one place.
Requests, your team, invoices, documents, shared credentials and sign-in security. One sign-in, any device. Take the tour.
Tour myALCO - New · Our work
See the sites we have built.
A newspaper rebuilt headless, a parts storefront, two Boise nonprofits and our own site. Every screenshot is the full page, scrolling live.
Browse our work
Clients we’ve worked with
We’ve worked with businesses, nonprofits and community organizations on managed hosting, website development and everyday IT support. These are some of the current and past clients we’ve supported, each with different systems, priorities and people. Explore their stories to see what working with ALCO looks like.




















