About this listing
We create focused AI agents for teams requiring a single working journey instead of an open-ended support hire. This service includes one FAQ or support agent constructed on tools you currently control.
You pick the journey at the beginning, whether that means answering FAQs, managing simple intake, or directing a support case to a human. We agree on the limits together, then build using your own accounts and API keys, ensuring you maintain ownership of the runtime and billing rather than depending on our infrastructure permanently.
Supported platforms: OpenAI Assistants, LangChain, n8n, Voiceflow, or custom Python built against your APIs. One primary platform is selected per order; this is not a parallel build across every stack simultaneously.
Working on your own tools and keys means you provide, or configure, the runtime account and API access. Usually this is an OpenAI API key, or login credentials for the chosen platform, plus entry to the FAQ, documentation, or ticketing system the agent will access. If any of these are absent, work pauses on the private page until access is verified, rather than moving forward on assumptions.
What is delivered:
- One named FAQ or support journey, confirmed in writing at the start
- Defined boundaries so the agent remains on topic and does not invent answers
- A clear escalation route for cases requiring a person
- The prompt or flow pack built for your selected platform
- A short test checklist and runbook so your team can run it after handover
- A private job-status page, a unique link for this order only, used for scope notes, questions, updates and handoff
How it proceeds: once ordered, we confirm the journey and access through the private page and work strictly within that agreed scope. Where the build can be finished, we deliver the agent and its runbook there. Where it cannot, due to missing access, an unclear journey, or another blocker, we note the reason clearly and stop, preventing silent modifications to the brief.
Delivery takes 2 to 4 business days from the moment access is granted, with free fixes available for 7 days following handover.
For ongoing accuracy, a monthly care plan is available from US$49, covering a monthly test run, FAQ updates and fixes.
Not covered by this fee:
- Building a complete multi-team support operation in one project
- Permanent hosting or runtime costs billed to our accounts
- Continuous on-call cover within this fixed price (available separately as a retainer)
- Unrestricted changes once the journey has been locked
Tiers: Basic Agent, Standard Agent and Deep Agent. Higher tiers provide a more thorough test checklist, more detailed run notes, and, with Deep Agent, a second review of your edge cases plus stricter handoff notes.
Each order has its own private page; there is no shared portal across clients.
To begin, choose a package, specify the journey and platform you require, and we will confirm scope on your private page prior to building starting.