Frequently asked questions

Answers to the questions we hear most.

What is WKIL?

WKIL is a platform for building, connecting, launching and managing AI Agents for business. You create an agent, give it your company knowledge, connect the apps it needs, then publish it and monitor its runs from one workspace.

What is an AI Agent?

An agent is a digital representative with a defined goal. It combines a model, instructions and personality, knowledge, tools and workflows to read, analyse, execute and produce results.

How do I create an AI Agent?

Choose a model, write the agent instructions and tone, connect a knowledge base and the apps it needs, then test it and publish. You can start from scratch in the Agent Builder or begin with a ready-made agent from the Marketplace.

Do I need coding skills?

No. The Agent Builder creates an agent without code, and starting from a ready-made agent shortens setup further. An API is available for technical teams that prefer to run agents programmatically.

Does WKIL provide ready-made Agents?

Yes. The Agent Marketplace includes more than 60 ready-made agents across sales, support, marketing, content, operations, HR, finance and industry-specific roles. Each one can be customised before you launch it.

Can multiple Agents work as a team?

Yes. Agent Teams group specialised agents around one business goal, each with a defined role. Five teams are ready to run: sales, customer support, HR, marketing and operations.

How does an Agent use my company knowledge?

Knowledge connects an agent to your source of truth through RAG. Documents are uploaded, chunked and indexed, and the agent retrieves the most relevant passages to ground its answers instead of guessing.

Can Agents connect to my business tools?

Yes. Agents connect to business applications such as messaging channels, CRM systems, stores, email and storage. WKIL also supports MCP, and custom tools let an agent call an internal API when no ready-made app exists.

How do I test an Agent before launch?

While you build, the agent stays a draft you can test. Deployment then captures a stable snapshot, so testing never affects what your customers are using.

How is an Agent published?

Deployment captures an immutable numbered version. One version is live at a time, and you can publish an update or promote an earlier version to roll back. Editing the draft does not change the live version.

Where can customers interact with an Agent?

An agent can be published as a shareable conversation page, called through a REST API, or exposed as an MCP tool for another system to call.

Can an Agent run automatically?

Yes. A workflow can start on a schedule you set, when a webhook arrives, on a table change, or on workspace events. Every scheduled run is logged like any other run.

Can I require human approval?

Yes. The Human in the Loop block pauses a run until a person approves, rejects or edits, which suits sensitive actions such as issuing a refund or sending a contract.

How do I control what an Agent may say or do?

The Guardrails block checks input and output against rules such as prohibited topics or sensitive data and blocks or redirects violations. The Evaluator block scores output against criteria you define and can retry when quality falls short.

Can I see what an Agent did?

Yes. Logs record every run step by step: the time, trigger and result, what each block received and returned, and the path the run took. Alerts can notify you by email or a team channel such as Slack or Teams when a run fails or a threshold is passed.

Can my team work in WKIL together?

Yes. A workspace holds your agents, workflows, knowledge, tables and members. Roles set who can do what: Owner has full control including billing, Editor can build and publish, Member can use agents within granted permissions, and Viewer is read-only.

Is my data separated from other companies?

Yes. Each workspace is isolated, so its data, agents and secrets cannot leak into another workspace. API keys and access tokens are stored as named secrets and are never exposed in logs or agent output.

How does pricing work?

WKIL bills in AI Credits across three plans. There is a free plan to start on, a Pro plan from $50 per month for 200 AI Credits with larger credit tiers and on-demand top-ups, and Enterprise with volume-based pricing. Usage depends on the number of runs, the model used, and the size of input and output.

Can I control spending?

Yes. A workspace credit limit sets a spending cap, and a threshold alert notifies you when a percentage of that limit is reached.

What does the Enterprise plan include?

Enterprise includes everything in Pro plus custom AI Credits, unlimited AI Agents, unlimited and custom integrations, audit logs, sensitive data detection and control, a custom SLA, dedicated onboarding, training and deployment support, and a dedicated account manager.

Does WKIL support Arabic?

Yes. WKIL is built Arabic-first with a right-to-left experience, and the product documentation is available in Arabic and English. The conversation page you publish for customers ships with a ready-made interface that supports Arabic and right-to-left.

Didn't find your answer?

Start from the login page and choose the right agent or team for your needs.

Ready to build your first AI Agent?

Start building an agent on the platform, or browse the ready-made Agent Marketplace.