questions

A few fair
questions.

What you get, what your team still does, and what happens when things change.

ask us anything

Getting started

Where do we start?

You don't need a list of AI projects. Tell us what took too long this week, what keeps slipping through the cracks, or what only one person knows how to do. We look at the actual work with you and choose a first job based on how often it happens, the time it takes, and what the agent would need to do it reliably.

What is the first call like?

Show us the spreadsheets, inboxes, and workarounds your team uses. We ask how the work gets done, where it gets stuck, and who checks it. Together, we identify a first job and the tools, information, and approvals it needs. Before service starts, we agree the scope, price, and acceptance criteria in writing. You don't need to prepare a technical brief.

Must we map every process?

No. We can start with one recurring job and document what we need to do it well. If the process lives in someone's head, we work through real examples with them. A broader review can help when several teams share the same problems, but it isn't a prerequisite. We agree the scope around the work you need and the time your team can give it.

Can we keep our software?

We start with your existing tools. If a tool works well and gives us the access the job needs, we connect to it. There is no reason to rebuild QuickBooks just to prepare an invoice. If a tool makes the job difficult or cannot connect safely, we explain the options and agree any changes with you first. We plan changes one process at a time, with the people who use it, so they can keep serving your customers during the transition. The shared company operating system brings the relevant context together so your agents can work across those tools.

What will our team do?

We need time with the people who do the job today, access to the relevant tools, and real examples to test. Your team helps us understand exceptions and checks that the result is right. We handle the build, connections, and ongoing maintenance. Before starting, we agree who needs to be involved and what we need from them. If your team is stretched, we start with a smaller job rather than ask you to redesign the whole company at once.

Do I need to understand AI?

No. You explain the work and the result you need in ordinary language. We handle the prompts, model choices, software, and maintenance. Your team's knowledge of the business matters more than knowing how AI works.

What if our data is a mess?

You don't have to organize the whole company before we start. We work with the people who do the job, look at real examples, and document the rules and workarounds they rely on. Those workarounds often keep the business running, so we need to understand them before changing anything. We bring together the information the first job needs and flag gaps that require a person's answer. We simplify unnecessary steps before automating them.

The managed service

Could we build it ourselves?

You can. If someone on your team can dedicate themselves to building, running, and improving the system, that may be the right choice. With us, you get a shared company operating system, managed agents, and people dedicated to keeping the work running. You benefit from our experience and ongoing improvements, while your team puts its time and creativity into the core of your business.

  • A shared company operating system. Keep operational data, work, and decisions together so agents have the context they need.
  • Agents that monitor agents. Automated checks flag failures and exceptions, with a named person responsible for investigating and fixing them.
  • Models, interfaces, and AI improvements. We follow new developments, test them against your jobs, and adopt the changes that help.
  • Connections and maintenance. We connect your existing tools and maintain those connections as APIs, software, and your processes change.
  • Testing and approvals. We agree acceptance criteria, test real examples, and define when an agent must ask a person before acting.
  • Backup and recovery planning. We agree what gets backed up, how it can be restored, and who owns recovery, including files and connected systems.
  • Security and access. We manage permissions, credentials, updates, and incident response, with access limited to the work each agent needs.
  • Ongoing support. We investigate mistakes, adjust agents as the business changes, and review whether the work is worth its cost.
  • Experience across implementations. You benefit from what we learn about making agents work, while your company's private data stays private.

The first working version is the beginning of the job. We dedicate people to keeping it useful. You focus your creativity on the work your business does best; we focus on making sure you are supported.

Who manages the agents?

A named person from our team manages the agents. You explain the job, agree the boundaries, and make decisions that need your judgment. We handle setup, monitoring, fixes, and improvements. We also agree which actions need your approval and review how much attention the agents still need from your team.

Why use a managed service?

Our service includes running the work after we build it. You get the agents, the software, and a named person responsible for monitoring, fixing, and improving them. We agree what the job should achieve and review the results with you. If you only need a plan or a one-off build that your own team will run, a consultant or agency may be a better fit.

How is completion checked?

We agree what counts as finished and what evidence proves it. For example, a report may need a receipt from the receiving system, and an invoice may need a confirmed send to the approved recipient. An agent saying 'done' is not enough. We test those checks before it starts. If completion cannot be confirmed, the job needs investigation or a person's review. We also account for the time your team still spends correcting work and handling exceptions when we review the results.

How is this unlike ChatGPT?

No. ChatGPT waits for you to type. An agent doesn’t wait. It reads the inbox, chases the invoice and books the meeting whether anyone remembered to ask or not.

What counts as one agent?

One job, done start to finish, the way you’d describe a narrow hire. Bookkeeping is usually a few agents: one builds invoices, one chases them, one matches receipts. We agree the list with you before anything starts, so the count never surprises you.

See AI employee examples
What about mistakes?

We agree what the agent can do on its own and what needs approval, then test real examples before it starts. When something goes wrong, we investigate, correct what we can, and tell you what happened and what we changed. A person from our team stays accountable for the work.

Why not hire a person?

Sometimes you should. People bring judgment, relationships, and hands-on skills that an agent cannot replace. We take on defined, recurring work with software, AI, and people managing it. That can free your existing team to serve customers and grow the business. Compare the cost of the whole service with the cost of hiring and managing someone to do the same work. An agent doing one narrow job is not a replacement for every part of a person's role.

Cost and continuity

What if you stop operating?

For new engagements, we require a written continuity plan and client-controlled copies before the first agent starts. The plan covers your data and files, the code and instructions needed to run your systems, and a no-fee continuity license if we permanently stop providing the service. Your team or another provider can take over. These arrangements must be set up and included in your signed agreement; they are not an automatic backup or recovery guarantee for existing engagements.

  • Daily encrypted exports of your operational data and files to storage your company controls, with 30 days of copies retained. Your designated administrator has access independently of us.
  • A handover package in your company's private repository, updated after material system changes. It includes your systems' source code, dependency and configuration instructions, operating procedures, and a recovery runbook. Credentials stay in a secure store, with recovery access documented separately.
  • A restore rehearsal before the first agent starts, then at least quarterly. We check that your administrator or a replacement provider can use the copies and instructions.
  • A perpetual license, written into your agreement in advance, to use, run, and change your company's systems if we permanently cease providing the service without a successor taking responsibility. It includes the shared runtime components needed to run those systems, for your business only, with no additional buyout fee.
  • If we choose to end managed service while still operating, the existing six-month notice and transition commitment applies. A normal voluntary buyout remains a separate option.

The continuity package is a requirement for new engagements. It is not yet implemented or verified for existing clients. The backup scope, access, and license must be confirmed in writing and tested before relying on them. After a takeover, you pay the hosting, AI, maintenance, and third-party service costs.

What is included monthly?

Building it is the first part. The monthly fee covers running the agent, the software and AI it uses, and the people who monitor it, fix failures, maintain connections, and adapt it as your business changes. You also benefit from improvements we make across the shared system. If you prefer to own and operate it yourself, the buyout is a separate purchase at 24 times your monthly fee. Paying for 24 months does not automatically transfer ownership. After a buyout, you take on the running costs and maintenance.

How does the bill stay capped?

You pay for at most ten agents at your company's pricing band. After your tenth agent, we keep adding up to a maximum of one a month at no extra cost and keep improving the agents you already have. Your bill can change if your company moves to a different size band. We agree the agent list and price before work starts, so you can compare the cost with the work it takes off your team.

What if I sell my company?

Yes, dependence on any provider is something a buyer may examine. Our contract transfers, so a buyer can keep the managed service. You can also buy the systems we built for you at 24 times your monthly fee, with the source code and a license to run and change them for your business. We move them to your own accounts and provide operating instructions and training. Your team or another provider then takes on the running costs and maintenance. These options help you plan for a sale; they do not guarantee a buyer's valuation or approval.

What if we end the service?

You can end any agent, or all of them, at the end of any month. Your data and documented procedures remain yours. To keep the agents running, you can buy out the systems at 24 times your monthly fee and have your team or another provider operate them. Without a buyout, the managed agents stop, and you use the returned data and procedures to arrange another way to do the work. The separate continuity protections apply if we permanently cease service, subject to your signed agreement.

let’s find the first job

Give your team its time back.

Show us what your team keeps doing by hand. We will work out where to begin.

Book a free hour

One hour. No technical brief needed. Or email us directly .