Tools
Give an agent something to do beyond talking, and know how a tool behaves when it fails.
A tool is a callable action: an HTTP request to your API, an action on a connected app, ending the call, transferring it, a calculation. Tools are defined on the Tools page, independently of any workflow — one definition can be attached to many workflows and many nodes.
Tool types
| Type | What it does |
|---|---|
| External HTTP API | Sends an HTTP request to an endpoint you host |
| Integration | Runs an action on an app you have connected, such as creating a calendar event |
| End Call | Hangs up when the agent judges the conversation finished |
| Transfer Call | Moves the caller to another phone number |
| Calculator | Evaluates arithmetic, including parentheses and exponents |
Picking a type on creation fills in a starting definition for it; each type then has its own configuration form.
The transfer tool is implemented against Twilio's call control. On any other carrier the tool can be defined but the transfer will not complete. Check which provider your number runs on before you build a call flow that depends on a warm transfer.
Where a tool runs
Two placements, and the difference is when the work happens.
- Attached to a node. Tools on an Agent Node or Start Call node run mid-call, invoked by the model as part of the conversation. The caller is waiting, so this is the placement for anything the agent needs in order to say its next sentence.
- On an Integration Action node. That node runs its tool once, after the workflow completes, durably, with retry and dead-letter handling. Use it for work the caller does not need to wait for — writing the outcome to a CRM, sending a follow-up.
Reads and writes
Whether a call is a read or a write decides what is safe to retry. A read can be repeated freely; asking for the same record twice costs nothing. A write cannot — if a create appointment call succeeds but its response is lost, a blind retry books the appointment twice.
Woise protects writes with an idempotency store scoped to a single run. Within one call, a write with the same identifying arguments executes once, however many times the model asks for it. Reads are deliberately never deduplicated: a stale answer is worse than a repeated lookup.
On top of that, a failure flagged as transient is retried exactly once after a short pause. A 4xx or a business rejection is not retried at all — it would only fail again.
When a tool call fails mid-call
A failed tool call does not end the call.
The error is caught and handed back to the agent as a structured result — a status and a message — rather than raised as an exception that would tear down the run. The agent sees the failure and decides what to say: try again, apologise, or take a different branch. Write the node's prompt so it has somewhere to go when a lookup comes back empty.
Work on an Integration Action node cannot affect the caller at all, because it runs after the call has ended. A transient failure there is retried with backoff; a permanent one is dead-lettered with the error recorded against the delivery rather than retried forever.
Next steps
| Goal | Guide |
|---|---|
| Connect an app before building a tool on it | Integrations |
| Attach a tool to the right step | Nodes |
| Tell the agent when to reach for a tool | Prompts |
Common questions
What is Woise?
Woise lets you build AI agents that answer and make phone calls and chat on your website. You lay out the conversation on a visual canvas, try it in your browser, then connect it to a phone number.
Do I need a phone number to start?
No. You can build and test an agent in the browser without one. When you are ready, connect a phone number and choose which agent answers it.
How does billing work?
You pay for the minutes your agents use, from your credit balance. There are no seat fees and no per-agent fees, so an agent that is not taking calls costs nothing.
Which languages can an agent speak?
Agents can listen and speak in more than 40 languages, including English, Spanish, Hindi, Tamil and Arabic. You pick the language and the voice for each agent.
Can an agent connect to my CRM or calendar?
Yes. An agent can look things up and make changes in 26 apps, such as Salesforce, HubSpot, Google Calendar, Slack and WhatsApp, while the caller is still on the line. You connect each app once.
Do I need engineers to build an agent?
No. Everything is built in a visual editor, so you can create and change an agent without writing code. If your team prefers code, there is also a REST API and an MCP server.
What happens to call recordings and transcripts?
Every call keeps a transcript and its outcome, and audio is only recorded if you turn it on. Only people in your account can see your calls, and your account lives in the region you chose when you signed up.
What do I have to set up or run?
None. We run the speech, the language models, the recordings and the phone connections, and scale them with your call volume. There is nothing for you to set up or keep running.
How do I reach the team?
Email contact@woise.ai, call or WhatsApp +91 83412 34080, or use the form on our contact page. A person reads every message.
Still stuck? Talk to our team.