( FAQ )
Frequently asked questions about UrbanHorn
How integration works, what happens when an agent cannot help, how pricing and credit are calculated, and what we do with your data — answered plainly.
How hard is this to integrate?
If your systems already expose APIs, the work is configuration rather than engineering. You connect an endpoint, describe in plain language what it does and what it returns, and the agent calls it when a conversation needs that information. You are not writing a bespoke integration for every question a customer might ask.
Do I need to change my existing systems?
No. Your systems of record stay exactly where they are. UrbanHorn calls them the way any other client would — nothing migrates into our platform, and nothing gets rebuilt. That is deliberate: the fastest deployments are the ones that do not require a data project first.
What if some of our systems don't have APIs?
Agents still work from your uploaded knowledge for anything that does not need a live lookup — policies, product detail, process questions. You can add live integrations later, system by system, rather than waiting until everything is connected.
What exactly is UrbanHorn?
An enterprise agentic-AI platform. It lets multiple models and specialist agents work together across your APIs, your tools, and your uploaded knowledge, behind a single orchestration layer — so you configure agents instead of wiring models and integrations together yourself.
Is this only for customer support?
No. Support is where most teams start, because the value shows up fastest and the work is well understood — but the same layer runs internal helpdesks, operations lookups, sales and onboarding assistance, and back-office workflows. What changes between them is the knowledge and the endpoints you connect.
Why orchestration rather than a single model?
Because no single model is best at everything, and keeping up with that is not a good use of your team’s time. We choose the model for each use case across the providers we support and route the work accordingly, behind one shared source of truth and one set of controls. You configure agents; we handle the model selection.
Which model will my agents use?
We choose it, per use case, from the providers we support — matching the model to the kind of task rather than running everything through one. You don’t need to select a model, hold accounts with model providers, or change anything when we add or move to a better one. If your procurement or compliance process needs to know which providers are involved, ask us and we’ll tell you.
Where do the answers come from?
From the knowledge you upload and the systems you connect. Uploaded material is the agent’s source of truth and answers cite it; live values come from your own endpoints at the moment they are asked for.
What happens when an agent doesn't know the answer?
It says so, instead of guessing, and escalates to a person with the transcript, the knowledge it retrieved, and any tool results already attached. An answer that is wrong but confident is worse than no answer, so the system is built to stop rather than fill the gap.
How does the handoff to a human actually work?
The conversation moves to a handoff inbox where your team picks it up with the full history in front of them. The customer does not repeat themselves and does not start again in a new channel. Growth and Scale plans include the handoff inbox.
Which channels does it work on?
WhatsApp, web chat, an interactive avatar, and a direct API. It is one agent with one set of knowledge and one set of integrations behind all of them, so an answer does not change depending on where the question was asked. Starter covers web chat and API; WhatsApp is an add-on at $20/mo per number.
Is it built for my industry?
The platform is industry-neutral. The same core serves retail, SaaS, hospitality, finance, healthcare and more. What makes an agent fit your business is the knowledge and endpoints you connect, not special code written for your sector.
What do I need to get started?
The documents you already have, and — if you want live lookups — an endpoint to point at. You do not need to write code to get a working agent, though the API is there when you want deeper integration.
How does the free trial work?
Three days, and you need a card to start — it is not charged during the trial. You get the platform to build an agent against your own knowledge and see how it behaves before you decide. When the three days are up the account moves to Starter at $49/mo unless you cancel first, and we email you the day before either way, so nothing arrives as a surprise. The avatar channel is not part of the trial.
What do the plans cost?
Starter is $49/mo and includes $15 of AI credit, 2 agents, and web chat plus API. Growth is $199/mo with $60 of credit, 5 agents, the WhatsApp add-on, and the handoff inbox. Scale is $499/mo with $175 of credit, 50 agents, avatar top-ups, and configurable retention.
What is AI credit, and what happens if I run out?
Plans are prepaid: your usage draws down a credit balance included with the plan. If you need more before the next cycle, you can top up in $10, $50, or $200 increments. There is no surprise invoice at the end of the month — you spend what you have added.
Is the avatar included in my plan?
No — the avatar channel is billed separately through its own top-ups, and it is not available during the free trial. We would rather say that plainly here than have you discover it at the end of a trial.
How is my data handled?
Every organisation’s knowledge, conversations, and tool connections are isolated from every other organisation on the platform. Your data is never used to train models for anyone else. Scale plans add configurable retention so you control how long history is kept.
Who builds UrbanHorn?
UrbanHorn is a product of Codelabs, built and supported by our team in Dubai. If you have a question this page does not answer, write to talk@codelabs.ae and a person will reply.
Three days free. A card is required to start and is not charged during the trial. When it ends you move to Starter at $49/mo unless you cancel first — we email you the day before either way. The avatar channel is not included.