The questions we get most often

Tier 1 is a single agent. Tier 2 is a team of 2-3 sharing memory and brand voice. Tier 3 is a full pre-bundled department. Tier 4 is a full company workforce deployed on your infrastructure.

Tiers 1-3 run on our managed infrastructure (subscription). Tier 4 is a build-and-handover (one-off project plus optional support retainer).

Tier 1-3 are monthly or annual subscriptions plus a one-time implementation fee. Tier 4 is a project fee plus an optional support retainer. Specific numbers are in the pricing page and confirmed in your proposal.

No. Every cost is in the proposal. The only thing not in our pricing is third-party API costs you might incur (e.g. paid Slack, paid CRM seats) - those stay with your providers.

Yes. Upgrades from Tier 1 to 2, or 2 to 3, are common. Downgrades are also available, with the timing aligned to your billing cycle. Tier 4 is structurally different - moving to or from it involves a fresh proposal.

Monthly for subscription tiers. We don’t lock in long contracts. Annual billing is offered at a discount if you’d prefer.

Yes, case by case. Mention it in your Hiring Brief and we’ll discuss in the proposal stage.

Both. The AI agents that inspired the catalogue are real configured agents that handle live customer work for our parent agency. The personas you see in the catalogue are productised versions of the same configurations, tuned to your brand voice and your data when you hire them.

Yes. Configuration is yours - the name, the tone, the introduction, the escalation rules. Our agents are ours. Yours is yours.

It escalates. We define the escalation path during configuration: which channel, which person, what context to include. The agent does not guess. It says “let me get the right person on this” and forwards the conversation with full context.

We work with major AI providers (Anthropic, OpenAI, and others). The specific model used for your deployment is disclosed in your proposal. We pick the best fit per role rather than locking everyone to one provider.

Six material differences:

Always-on, system-integrated. A Custom GPT lives inside ChatGPT - you open it, talk, close the tab. Our agents are deployed into your operations. They receive incoming emails, reply within configured guardrails, write to your databases, coordinate with other agents.

Configured to your data, not just your instructions. A Custom GPT is shaped by a system prompt and an optional knowledge file. Our agents are configured against your brand voice (tested against your existing copy), live data sources (CRM, helpdesk, calendar), and your team’s escalation paths. The configuration is a multi-day workflow, not a 200-word prompt.

Multiple agents working as a team. A Custom GPT is one agent. A Solutions department is 5 to 6 agents sharing memory, coordinating handoffs, escalating to each other and to humans. None of that exists in Custom GPT land - it’s all human glue work.

Provider-agnostic. Custom GPTs run on OpenAI’s infrastructure with OpenAI’s models. Our agents run on Anthropic, OpenAI, or other providers depending on the role. Best fit per role, not the only option.

Data residency choice. Custom GPTs send conversation data to OpenAI under their default terms. Tier 4 deployments run on your infrastructure - data never leaves your servers. For regulated practices, this is the only viable shape.

Owned, not borrowed. A Custom GPT works only as long as your customers have ChatGPT accounts. Our agents are deployed under your brand, on channels your customers already use. Part of your operation, not a chat tool you’ve taught some tricks.

Yes - and we’re honest about that. AI is probabilistic. We mitigate with brand voice configuration, escalation rules, and review cycles. Most mistakes are caught before they reach a customer. When something does slip through, we adjust the configuration and tell you what we changed.

Yes. Agents introduce themselves as AI assistants from the first message. Disclosure is the default - we don’t deploy “covert” agents. In our experience, customers prefer fast accurate AI replies to slow human ones, especially out of hours.

Tier 1-3: data flows through agents we operate on our managed infrastructure (Hostinger KVM4, Jakarta). We act as your Data Processor.

Tier 4: data stays on your infrastructure. You’re the sole Data Controller. We have no access unless you grant it.

Yes. Our Privacy Policy covers all three. We sign Data Processing Agreements as standard for Tier 1-3. For regulated practices, we recommend Tier 4 because the controller-processor split moves entirely to your side.

No. We use AI providers offering enterprise data agreements - no training on customer data, limited retention windows. The specific provider terms apply to your deployment are disclosed in the proposal.

Conversation logs handled by Tier 1-3 agents we operate: typically 90 days unless you ask for a different window. Hiring Brief and prospect data: 24 months from last interaction. Customer engagement data: duration of engagement plus 7 years for tax records. Full breakdown in the Privacy Policy.

Yes. DPA is part of the standard Tier 1-3 engagement. For Tier 4 we don’t need one because we’re not your processor - you are.

Common integrations include email (Gmail, Outlook), helpdesks (Chatwoot, Zendesk, Intercom), CRMs (HubSpot, Pipedrive, Odoo), messaging (WhatsApp Business, Instagram, Messenger, Slack), calendars (Google, Outlook), and knowledge bases (Notion, Confluence). Custom integrations are possible - mention what you need in the brief.

We can almost always integrate. If your tool has an API, we can connect to it. If it doesn’t, we’ll suggest alternatives or work around the gap. The proposal will be honest about which integrations are clean and which involve workarounds.

We need scoped credentials, not full admin. Read-only where possible, write access only for the specific actions the agent needs to perform. We document what permissions we asked for and why.

Yes. Some Tier 1 deployments run as a chat-only experience with no system integrations. The agent answers, escalates to humans, and that’s it. Adding integrations comes later as you find them useful.

A real human. We assign one point of contact per engagement so you don’t get bounced between people. For Tier 4 builds, you’ll meet the technical lead too. No client portals to log into.

Small changes are absorbed within scope. Anything that affects price or timeline goes into a written change order before we proceed. No surprise charges.

Tell us, in writing. We don’t proceed past kick-off, the first agent shift, or major milestones without sign-off. If something’s off after launch, we adjust the configuration and prove the fix. Subscriptions are cancellable - we’d rather lose a customer than keep one we can’t make happy.

Email and your preferred chat tool (Slack, WhatsApp). Status updates in writing at agreed cadence (usually weekly for Tier 4 builds, on-demand for Tier 1-3 deployments). All commitments in writing.

Standard support is during Bangkok business hours. Tier 4 support retainers can include extended hours - if that matters, mention it in the brief.

Hosting, monitoring, model updates, and support. Configuration tweaks and minor adjustments are included. Major scope changes go through a change order.

Yes. Subscriptions can be paused (e.g. seasonal businesses) or cancelled with 30 days notice in line with proposal terms. Tier 4 deployments are yours to operate, including switching them off entirely.

We notify you in advance and either absorb the change, propose an adjustment, or move you to a different model. Material changes to underlying providers are flagged in writing.

Yes. Most Tier 2-3 subscriptions include monthly review of agent performance with adjustments. If you’d like deeper optimisation - new use cases, additional channels - that’s typically a small expansion of the engagement.