Terms, privacy & refunds.

01Who we are

Clawbak is a product of Clawbak LLC, a limited liability company registered in Wisconsin, United States. In this document “we”, “us” and “Clawbak” mean Clawbak LLC; “you” means the business that has signed up for the service.

Most of our customers are managed service providers (MSPs), and most of the information you put into Clawbak belongs to your clients rather than to you. We have written this document with that in mind: you are our customer, and your clients’ data is something we hold on your behalf and nothing more.

02What the service does

Clawbak connects to a support mailbox you nominate, reads the support requests that arrive there, classifies them, drafts replies, and builds a knowledge base from the way your technicians actually resolve tickets. Depending on how you configure it, it can also send replies, create tickets in your PSA, and run pre-approved scripts on endpoints you have connected.

Nothing happens automatically that you have not switched on. A new account starts in the most conservative mode available: the bot observes and drafts, and a human sends. Every increase in autonomy — replying without review, running a script, closing a ticket — is a setting you change deliberately, per ticket type, and can change back.

03Your account

You sign in with your Microsoft work account. We do not issue or store passwords for Clawbak, and we never see your Microsoft password.

You are responsible for who you grant access to your Clawbak workspace and for what they do with it. Tell us promptly if you believe an account has been compromised.

You must be at least 18 and authorised to enter into this agreement on behalf of the business you are signing up.

04Trial & billing

We collect card details when you sign up so that the service continues without a gap if you decide to keep it. You are not charged during the trial. On day 31 the first monthly payment is taken automatically, and monthly thereafter on the same date.

Prices are in US dollars and exclude any taxes that may apply. Founding-member pricing, where offered, is locked for 12 months from the date you subscribe.

Payments are processed by Stripe. We do not receive or store your full card number.

If a payment fails

Stripe will retry a failed payment over the following days and email you. We will also contact you. We do not cut off access the moment a payment fails — if there is a problem with a card we would rather hear from you than lock you out of your own helpdesk. If an account remains unpaid after we have made reasonable efforts to reach you, we may suspend it.

05Cancelling

You can cancel at any time from your billing portal, the link to which is in your welcome email and available on request. You can also simply email us and we will do it for you.

Cancelling during the trial means you are never charged. Cancelling after that stops future payments; your workspace stays available until the end of the period you have already paid for.

There is no cancellation fee, no notice period, and no need to explain yourself (though we would genuinely like to know what went wrong).

06Refunds

After the first 60 days, we do not offer refunds for time already elapsed, but you can cancel at any moment and will not be billed again.

If we bill you in error — a duplicate charge, a charge after a cancellation — we will refund it in full, and you do not need the 60-day window to ask.

07Your data

What we hold

• Support emails and text messages that reach the mailbox or number you connect, including their contents and attachments.
• Knowledge-base articles, whether you wrote them or the bot drafted them from a resolved ticket.
• Contact records for the people who contact your helpdesk — typically a name, an email address, sometimes a phone number, and any notes your technicians add.
• Configuration and credentials for the systems you choose to connect (your PSA, your RMM, your documentation platform).
• Operational logs of what the bot did and what your technicians did.

It is yours

You own your data. We hold it to provide the service to you and for no other purpose. We do not sell it, we do not share it with other customers, and we do not use it to train AI models — neither ours nor anybody else’s.

Each customer’s data lives in its own separate storage account and its own separate search index. This is an architectural separation, not a filter applied to a shared pool.

Getting it out, and getting rid of it

You can export your knowledge base from the dashboard at any time. If you want a full export of your ticket history, ask and we will produce one.

When you close your account, we mark it for deletion and keep it recoverable for seven days in case the closure was a mistake. After that window we permanently delete the storage account holding your data, the search index, and the associated records. Backups and logs age out on their own schedules and are not retained indefinitely.

We do not retain your data for marketing purposes after you leave.

Your clients’ data

Where the information you put into Clawbak concerns individuals — your clients’ employees, typically — you are the controller of that information and we process it on your instructions. If you need a data-processing agreement in place, email us and we will sign one.

08AI, and what it will and won’t do

Clawbak uses large language models (currently Anthropic’s Claude) to classify tickets, draft replies, and write knowledge-base articles. Two things follow from that, and we would rather say them plainly than bury them.

It can be wrong. A classification can be off; a drafted reply can be confidently incorrect. That is why the default configuration puts a human between the bot and your client, why increases in autonomy are per ticket type rather than global, and why a single bad automatic reply demotes that ticket type back to draft-only automatically. You remain responsible for what is sent to your clients under your name.

Your content is not training data. We send ticket content to our AI provider to get an answer back, under commercial terms that prohibit using it to train their models. We do not train models on your data ourselves.

When the bot sends a message to one of your clients directly, it identifies itself as an automated assistant. It does not claim to be a person.

09Who else touches your data

We use a small number of third parties to run the service. This is the full list as of the date at the top of this page:

Systems you connect yourself — your PSA, your RMM, your documentation platform — are your relationships, not ours. We use the credentials you give us to talk to them on your behalf.

If we add a subprocessor that handles customer data, we will update this list.

10Security

• Sign-in is through Microsoft; we hold no passwords.
• Data is encrypted in transit and at rest.
• Each customer’s data is in separate storage, not a shared pool with row-level filtering.
• Credentials for the systems you connect are stored in your own tenant storage, not in shared configuration.
• Where the product can act on an endpoint, it runs only scripts a human at your organisation has explicitly approved, and it records every run.

We are a small company and we do not currently hold SOC 2 or ISO 27001 certification. We would rather tell you that than imply otherwise. If you need specific assurances for your own compliance obligations, ask and we will answer honestly about what we do and do not have.

If we become aware of a breach affecting your data, we will tell you promptly and tell you what we know, including the parts we do not yet know.

11Acceptable use

Don’t use Clawbak to:

• Break the law, or help somebody else break it.
• Send unsolicited bulk messages of any kind. The messaging features exist to reply to people who contacted you first.
• Access systems or data you are not authorised to access.
• Attempt to extract another customer’s data, or to probe the service for that purpose without our written agreement.
• Resell the service as your own product without a written arrangement with us. Using it to serve your clients is exactly the point; white-labelling it as your own software is a different conversation, and one we are happy to have.

We may suspend an account that is causing harm to other customers or to the service. We will tell you why.

12Availability

We aim for the service to be available continuously, and we monitor it continuously. We do not currently offer a contractual uptime guarantee at the Starter tier, and we are not going to publish a number we have not earned the operating history to stand behind.

Clawbak depends on services we do not control — principally Microsoft 365 and Microsoft Azure. An outage at one of those may be an outage of Clawbak.

Clawbak is not an emergency system. Do not rely on it as the sole path for anything time-critical or safety-related.

13Liability

The service is provided as it is. To the extent the law allows, we disclaim implied warranties of merchantability and fitness for a particular purpose.

To the extent the law allows, our total liability to you for any claim relating to the service is limited to the amount you paid us in the twelve months before the claim arose. We are not liable for indirect or consequential losses, including lost profits or lost business.

Nothing here limits liability that cannot be limited by law, including for fraud or for death or personal injury caused by negligence.

14Changes to these terms

We may update this page. If a change materially affects your rights or what we do with your data, we will email you before it takes effect and give you at least 30 days’ notice, during which you may cancel without penalty. Minor clarifications take effect when posted, with the date at the top updated.

These terms are governed by the laws of the State of Wisconsin, USA.

15Contact

A person reads this address, usually the founder.

tony@clawbak.ai
Clawbak LLC · Wisconsin, United States