Frequently asked questions.
Getting started
How long does it take to get a workflow live?
48 hours from access. Once we have access to the relevant systems and you've walked us through the process, we can have the agent live in 48 hours. The initial scoping happens on a 25-minute call with our CEO.
Do you need access to our production systems?
Yes — usually a dedicated user role with the minimum permissions needed to do the work, the same as an outsourced human operator would receive. We never touch anything outside the agreed workflow scope, and all actions are logged and auditable.
Can we start with just one workflow?
Yes, and we recommend it. There is no minimum cohort, volume, or contract value. The first pilot is usually one workflow priced per completed task, so you can validate quality and unit economics before committing to anything broader.
What if our workflow changes after we go live?
Tell us what changed and we implement it for you. No SOW, no retraining cycle, no extra bill — and no staff to re-learn it, because there is no staff. Most changes go live the same day we hear about them.
Technical & integration
Can you work with systems that don't have an API?
Yes. We use APIs where they exist because they're faster and cleaner. Where they don't, the agent operates the UI directly — the same way a human operator would. Legacy systems, no-API tools, and custom in-house software are all in scope.
We use a legacy or custom system. Will your AI agents work with it?
Almost certainly yes. If your back-office team can log into it and use it today, the agent can too. Custom admin panels, mainframe-era ERPs, and obscure industry-specific tools are common in our customers' stacks — we built the company assuming the world is full of them.
Quality & exceptions
What happens when the AI gets something wrong?
The agent flags genuine edge cases for human review rather than processing them wrong. When a real error does ship to production, it counts against the per-task SLA — penalties are 20% of the workflow contract value, so quality is structurally aligned. Errors are also reviewed end-to-end so the fix lands the same week.
Can we audit or review the AI's work?
Yes. Every task the agent runs is logged with input data, decision trace, and final action. You can review any individual task, sample audit a percentage, or stream every action into your own observability stack.
Pricing & commitments
How is your pricing structured?
Per completed task — typically around 50% of the cost of human labour on the same task. No headcount, no labour-hour, no minimum cohort, no setup fee, no separately billed automation. A per-task price is agreed upfront for each workflow, and the invoice is exactly the number of completed tasks times that price.
Is there a minimum commitment or volume?
No. There is no minimum volume, minimum value, or minimum contract length. Pilots are typically month-to-month so you can validate quality and unit economics with no lock-in. Longer commitments are available if you want pricing certainty across a multi-year horizon.
What's the SLA penalty if you miss quality targets?
20% of the workflow contract value, paid back to you on missed quality. The industry-standard cap is 3%. We can offer more because per-outcome pricing aligns our P&L with your quality bar — if the agent gets it wrong, we lose money too.
Got a question we haven't answered?
Ask it on the call. 25 minutes with Ping, our CEO. Bring one workflow your ops team complains about most — we'll have an AI agent up and running on it within 48 hours.
Keep reading.
Human offshore teams vs AI agents — side by side →
How our AI agents handle your back-office work →