Your AI system, managed by our team
EdgeKeeper. Your AI and business tools, working together.
By Automated Edge
EdgeKeeper is our software on a dedicated server with no public internet address and restricted outbound access. It gives you private AI connected to your business systems. We connect the systems you choose under agreed permissions. Your team works in Switchboard, while we manage the server and support what we build.
Built forBusiness owners who want useful automation and one team to look after it, without having to run an AI system themselves.
01 / The outcome
What you get.
- Connect the business tools you choose so information does not need to be copied by hand
- Give people and AI access to the information they need, under the rules you agree
- Automate the work identified in your assessment, starting with the priorities in your proposal
- Use outside services under agreed permissions when the work calls for it
- Have one team to contact for help with the system and automations we build
The scope
What's included.
- Our proprietary EdgeKeeper software, engineered and developed by Automated Edge
- A dedicated Microsoft Azure server that we set up and manage
- Connections to the tools agreed in your proposal
- Agreed access rules, read-only questions in temporary sandboxes, and records of access for review
- The automations in your proposal, built to an agreed scope and delivery date
- Switchboard for answers, reports, dashboards and AI agents across the systems we connect for you
- As many read-only AI agents as your team wants, for any job, just by asking
- Ongoing monitoring, maintenance, and support under your service agreement
- A named lead who knows your business and reviews how the system is working
- Ownership of your business data and knowledge, with full export rights in the contract
02 / Delivery
How we do it.
- 01
Start with the proposal
Your Edgerton assessment identifies the work worth changing. We agree the build, connections, price, and delivery date. You see the estimated savings and payback before deciding to go ahead.
- 02
Set up, connect, and build
We set up the server with no public internet address and restricted outbound access, install EdgeKeeper, and connect the tools you approve. Each question runs read-only in a temporary sandbox. We agree what it can access, and a person separately approves a specific change before it runs. Then we build the automations in your proposal.
- 03
Keep it working, together
We manage the server and connections, fix problems with what we build, and help your team use it. Contact us whenever you need help with the system. As your needs change, we can agree new tools and automations to add.
Explore the evidence
See how the findings become a proposal.
Explore a sample assessment and proposal to see how the recommended work is explained. The businesses and estimates are illustrative. Your build starts with the findings from your own business.
Fictional companies and illustrative data. Explore without signing in.
Open the live sample reportShadow AI
Give people something better than a public chatbot.
When the approved process is slow, staff reach for personal chatbot accounts. Client documents can end up in tools the business never reviewed. Banning them alone leaves the work unchanged.
Switchboard gives your team an approved workspace connected to your own systems and data. It can answer about your actual business, clients and work, with read-only AI agents your team creates for the jobs it needs.
EdgeKeeper runs on a dedicated Microsoft Azure server we manage, with no public internet address and restricted outbound access. Each question runs read-only in a temporary sandbox. A person approves specific changes, and access is recorded for review. Your data remains yours.
How to detect shadow AI in your business ↗03 / In practice
What this can look like.
Illustrative workflows. Your assessment determines the scope.
Prepare a quote
Bring together the right customer details, past work, and prices to draft a quote. Your team reviews it before it goes out.
Connect an enquiry to the next step
Prepare customer details and a proposed next task from an approved form or email source. Your team approves changes to the customer system.
Assemble a recurring report
Gather information from the tools you use so someone can review the result instead of spending hours putting it together.
Work with outside services
An AI assistant can use an approved outside service as part of a task. We agree what information it can share and the approval required before writes or outbound actions.
We use it. We build it. We support it.
EdgeKeeper runs Automated Edge today. We developed the software for our own business and bring the same platform to our clients. Our team stays responsible for the system we deliver, from setup through ongoing support.
The promises we put in writing.
If we don't find ROI, you get the fee back.
Edgerton measures how your team's work really happens (with your team's knowledge and consent) and puts a dollar figure on what it finds. If there's no clear ROI path, the verdict is 'don't build', and the Assessment fee comes back. The report is yours to keep either way.
Production by a fixed date.
Set during the Build. If we miss it, additional weeks of build are on us.
You keep your data, whatever happens.
Your records, everything the brain learned about your business, the log: yours in the contract, with full export rights. The server and the software are ours; the knowledge is yours. Leaving is a documented handover, not a hostage negotiation.
Your rules decide what the AI sees.
Only what you approve goes in. Nothing you don't approve comes out. And every access is written down, so "what did it see?" always has an answer.
A named engagement lead.
Not a queue, not a rotation. You know their name, they know your business, and they pick up the phone.
Before you decide
Questions we get.
How does this help with shadow AI?
By giving people a useful approved alternative to personal chatbot accounts. Switchboard can answer about your business, clients and work using the systems we connect for you, and your team can make as many read-only agents as it wants. EdgeKeeper runs on a dedicated server we manage, with no public internet address and restricted outbound access. Permissions still apply: a person approves specific changes and access is recorded for review.
What is EdgeKeeper, in plain English?
It is our software running on a server dedicated to your business, with our team managing it. The software connects your tools, controls access, and coordinates AI tasks. The managed service means you have help with what we build.
Do we need to buy a server or make space for one?
No. We provide and manage the dedicated server on Microsoft Azure. There is no hardware to maintain in your office. Its size and running cost are worked out in the proposal.
Does all our data have to move?
No. We connect the systems agreed in your proposal. Your tools can stay where they are. We agree what information the system can use, where it is stored, and what may be shared with any outside service it uses. The database arrangement and access rules are documented in your proposal.
Can an AI assistant talk to the outside world?
Yes, where that is part of the agreed work. We can connect an assistant to outside services, with rules about what it may access, share, and do. Connecting a service does not give it unrestricted access to your business.
How is access to our systems controlled?
The server has no public internet address and restricted outbound access. Each question is scoped read-only and runs in a temporary sandbox. System credentials stay server-side. A person separately approves each specific change, and access is recorded for review. Your full security review explains the controls and their limits.
What should our IT provider review?
Review the server controls, database standard, connected systems, permissions, data processing, retention, recovery arrangements, and handover terms with us before deployment. The full security page explains the controls and their scope.
Who helps when something breaks?
We do. We support the software, connections, and automations we build. Your team has a named contact and can reach us when it needs help. Support coverage and response times are set out in your service agreement.
What does an AI managed service provider do?
We build and run your AI system as an ongoing service. That includes the server, the connections, the automations in your proposal, and support for your team. You do not have to keep it running yourself.
Can we add more later?
Yes. In Switchboard, describe the job and create as many read-only AI agents as you want. New connections, apps and automations that make changes are built with you, with an agreed scope, cost and business case. Changes stay held back until a person switches them on.
What happens if we leave?
Your business data, records, and knowledge are yours, with full export rights in the contract. We provide a documented handover. The server and EdgeKeeper software are ours; the information we built up for your business is yours.
Related guides.
Your next step
Start with an Assessment.
We install Edgerton on the computers you choose, with your team's knowledge and consent. After ten business days, you get a report and a proposal to help you decide what to do next.