FAQ
The questions MSPs ask first
34 answers, grouped by the page that owns each one. Every answer here is the same answer on that page.
Which stacks are the first launch focused on?
Starting with Autotask + Datto RMM and ConnectWise PSA + ConnectWise Automate. Claudia is invite-only for a limited supervised launch. Claudia’s AI agents resolve tickets end to end: they ask the right questions, get your technician’s approval, fix the problem, document it and close the ticket. Pilot scope and prerequisites are confirmed with your team.
What does my team actually have to change?
For the limited launch, choose the ticket types Claudia handles first and name the technicians who approve fixes for each client. Claudia works inside your existing PSA and RMM; your team reviews the approvals, outcomes and closure notes with us before the scope widens.
Can the agents act without a human approving?
The current limited launch is supervised. The broader product uses graduated autonomy, but unattended execution is not part of this launch offer. Any later change in autonomy needs a separately agreed scope and evidence.
What if a suggestion is wrong?
Your technician sees the diagnosis and the exact change before approving, and can reject it; nothing changes until they approve. If a fix does not resolve the problem, Claudia hands the ticket back with everything it found rather than trying something else on its own.
Will this work on our messy documentation?
Documentation and ticket history are part of the broader operating model. Their quality, access and relevance are reviewed when the pilot is scoped. Incomplete documentation is something we plan around, not something we promise will produce a reliable answer.
Do we have to migrate off our PSA?
The initial launch is designed around your existing Autotask + Datto RMM or ConnectWise PSA + ConnectWise Automate pairing. Confirm tenant access, permissions and the technician's workflow during scoping; catalogue coverage alone does not establish that every setup is ready.
Who does the setup work?
The published offer includes a two-hour onboarding session at no charge. Our team reviews the agreed workflow, integration prerequisites and technician walkthrough with you. The session is not a guarantee that every tenant or broader catalogue workflow will be ready in two hours.
Why not just hire another L1 technician?
A technician adds judgment, capacity and customer relationships across many tasks. Evaluate Claudia against a narrower, repeatable workflow that your team can supervise and measure. Use your own workload and cost assumptions in the calculator, then validate any savings in the pilot before making a staffing decision.
What happens when we run out of credits?
The published offer includes a notification at 80% of your pack. Usage above the pack is billed at that pack's own rate, with no separate overage penalty. A pack is a minimum commitment, not a usage cap. Compare the commitment and rate of each pack against your expected usage; a larger pack can be cheaper at higher volumes.
Do unused credits expire?
They roll forward within your annual term. Service desk volume is lumpy — a quiet August does not mean you over-bought, and you should not be paying twice for the same work when September arrives.
How do we know how many credits we need?
Use the published scenarios as a starting point: about 4 credits a ticket on Starter, 6 on Pro and 8 on Elite. Actual consumption depends on the actions performed. Confirm the included scope before your 30-day trial and use observed usage to estimate a suitable pack; a trial cannot predict every future month exactly.
Is credit pricing predictable, or a taxi meter?
The rates are published; consumption varies with the actions performed. The calculator uses scenario assumptions of 4 credits per ticket on Starter ($0.40 at the entry rate), 6 on Pro ($0.60) and 8 on Elite ($0.80). Your agreed pack sets a minimum commitment, and excess usage is billed at that pack's rate. Use your own volume inputs, then validate consumption in the agreed supervised trial. These estimates do not fix the cost of every real ticket or predict savings.
Could we not build this ourselves with an LLM?
You can build a focused workflow yourself. Compare the ongoing work as well as the model call: authenticated integrations, tenant permissions, output validation, audit evidence and API maintenance. Test the same bounded task in each option before estimating total cost or coverage.
We tried an AI tool before and it did not stick. Why is this different?
Start with one useful ticket type and an explicit acceptance check. For this launch, agree the ticket types and who approves fixes, let Claudia resolve them with your technician's approval, and review the outcomes together. Expand only after the evidence shows the workflow is useful to your team.
What happens if an agent gets something wrong?
The initial launch is supervised: your technician sees the diagnosis and the exact change, and nothing runs until they approve. If a result is incomplete or unexpected, Claudia stops and hands the ticket back. Unattended execution is outside this launch scope; any later change needs separate agreement and evidence.
How long before it is actually running?
The published offer includes a two-hour onboarding session with no onboarding fee. Readiness also depends on tenant access, endpoint permissions and the agreed workflow's prerequisites. Confirm these before the session; the session length is not a universal time-to-production guarantee.
Do we have to change PSA, or migrate anything?
Starting with Autotask + Datto RMM and ConnectWise PSA + ConnectWise Automate. We scope the supervised workflow around the existing pairing and confirm tenant access and the technician's steps before the pilot. Other catalogue entries do not establish launch availability.
Are we locked in?
Plans run on a one-year commitment, which is why every invite starts with a 30-day free trial — and why the commitment carries a 60-day money-back ROI guarantee: if it has not earned its place in 60 days, you get your money back. Agree a test or pilot environment, permitted endpoint scope and acceptance checks before the trial begins. Use the supervised workflow's observed results and credit usage to assess fit; access to production tickets or endpoints is scoped separately. The guarantee's terms are published in full on the pricing page, under The guarantee: how the return is measured, in your numbers; how to claim, in one email; and what comes back, within 30 days.
Where does our client data go?
Data is encrypted in transit and at rest, and security controls are continuously monitored and published in our Trust Center at trust.aiautopilot.ai, covering infrastructure, organizational, product and data privacy controls. Elite clients with stricter compliance requirements can discuss on-premise deployment.
Why do you not just tell me how much you will automate?
The calculator uses your entered workload, cost assumptions and our published prices. Up to 72% is an owner-set modeled coverage ceiling, not a measured customer resolution rate. The operating model counts ticket types, not ticket frequency or successful executions. Each launch workflow is scoped and supervised with your team. The Agentic MSP Operating Model classifies 508 of 600 ticket types as resolution-ready; that classification does not measure how often those tickets occur or are successfully completed in your environment. Review the model at https://mspagentic.ai/the-model/ and measure your own workflow before estimating realized savings.
Where does the hours figure come from?
A 2,080-hour working year per technician — 40 hours across 52 weeks — multiplied by your headcount and the share of time you said goes to repetitive work. Nothing else is assumed.
What counts as a repetitive ticket?
Examples include password resets, account unlocks, access and licence requests, onboarding and offboarding. Classify a representative sample of your own tickets and record the time spent; do not substitute an unsupported industry percentage for your baseline.
Can you run this against our real numbers?
We can discuss your workload and cost assumptions during a demo. Reviewing actual ticket history requires separately agreed access, scope and data handling. A calculator scenario is not measured customer savings.
Which AI providers do you use?
Anthropic Claude and OpenAI, hosted on Microsoft Azure within the United States. Both are configured so your data is not retained after processing and is never used for training.
Can we opt out of AI features entirely?
Yes. Contact us and we will disable AI processing on your account. The platform's non-AI functionality continues to work.
What happens to our data if we leave?
Customer data is deleted within 30 days of contract termination, unless a longer period is required by law or you request earlier deletion.
Do agents have access to our client credentials?
No. Credentials, API keys, and secrets are blocked or masked before any request leaves your environment. Agents operate on ticket content and documentation, never on secrets.
Are you SOC 2 certified?
Not yet — our SOC 2 Type II audit is underway, with the report expected Q1 2027. In the meantime our 73 security controls are continuously monitored by Vanta and published live in our Trust Center, covering infrastructure, internal procedures, product, organisational, and data privacy controls. Ask us for the report once it is issued.
Do we need to replace our existing automation?
No replacement is required to evaluate Claudia. Start with one agreed workflow, keep your existing runbooks available and compare the observed result and operating effort. Do not assume an AI-assisted workflow removes the need for script maintenance or technician review.
Does a chat interface mean a product cannot run actions?
No. Interface and execution capability are separate questions. Ask each vendor to demonstrate the same task on your stack, including denied permission, an execution failure and the technician's view of the result.
Should we use modeled coverage to plan a headcount reduction?
No. Up to 72% is an owner-set modeled coverage ceiling, not a measured customer resolution rate. The operating model counts ticket types, not ticket frequency or successful executions. Each launch workflow is scoped and supervised with your team. Ask for a demonstration on the stack and task you intend to use. Measure technician effort, exceptions and actual task outcomes in your own environment before making a staffing decision.
Can we compare a vendor's resolution percentage with your modeled coverage?
No. Up to 72% is an owner-set modeled coverage ceiling, not a measured customer resolution rate. The operating model counts ticket types, not ticket frequency or successful executions. Each launch workflow is scoped and supervised with your team. Ask for a demonstration on the stack and task you intend to use. A fair outcome comparison needs matching tasks, operating conditions, denominators and measurement periods.
Does a cheaper credit mean a cheaper workflow?
Not by itself. Compare the credits consumed by the same task, included allowances, add-ons, overage or upgrade rules and retained human effort. Use an observed workflow rather than treating one vendor's credit as equivalent to another's.
Does Flamingo publish pricing?
Yes. Its current pricing page publishes per-device options and AI token allowances. Review the linked page for current terms, included capabilities and any limits or additional charges; do not compare a device price directly with the price of one action credit.
Ask the one that is not here
Claudia is invite-only for a limited supervised launch. Invited MSPs receive a 30-day free trial in an agreed test or pilot environment — no card, and no onboarding fee. Scope is confirmed before access.