Why she is called Siya.
Siya is another name for Sita, a princess of Mithila remembered for her courage, and for arguing that girls should be taught. We wanted a name that meant somebody who does the work. Not a product code with a vowel removed.
Our mission
To take the work nobody chose off the people who did not choose it.
Our vision
A floor where the only thing a person does is talk, and everything that follows a conversation has already happened by the time they hang up.
She is on our own line right now. Ring it and interview her.
Bhoomija. The daughter the earth had.
We all call the earth a mother. India is where that mother has a daughter with a name. Sita was found in an open furrow on a dying plain, and bhoomija - born of the earth - is what she has been called ever since.
Mithila
A kingdom on the plain where Bihar meets Nepal, ruled by Janaka, a king better remembered as a philosopher than as a soldier. The rains had failed for twelve years. The court told him to plough the first furrow himself, as an offering, and he went out to the field and did it.
The furrow
He found a child in it. He raised her as his daughter and called her Sita, which is the Sanskrit word for the furrow itself. The rains came back that year, and the region has treated her as the reason ever since. She is held to be an incarnation of Mahalakshmi, the goddess of prosperity and of what sustains a household, and she is worshipped across India today.
The bow
Janaka kept Shiva's bow in the palace, a weapon no visiting prince could lift off the ground. Sita moved it as a girl, without ceremony, while cleaning the hall it stood in. Her father made lifting it the condition of marrying her, and one man in the world managed it. That was Ram.
Ram
She married him and followed him into fourteen years of exile she was not required to share. The Ramayana is usually told as his story. Read it again and it is hers: she is the one who chooses the harder thing every time it is offered, and she is the one who refuses to be taken back on terms she did not agree to.
The argument for teaching girls
She was educated in a court that debated philosophy in public, and she held that daughters should be taught what sons were taught. It is the part of her story told least often and it is the part we thought about hardest. Thousands of years before anybody was obliged to agree with her, she was already right.
And why an AI
Because a name is a promise about behavior. We could have called her Nova or Aria and promised nothing. What we wanted to promise is that she does the work, she does not stop when it gets tedious, and she hands the difficult thing to a person instead of pretending. It is a very large name to have taken. It is supposed to be hard to live up to.
Six things, on every call, all day.
Not a feature inside the platform. SIYA is the platform doing the job you currently pay people to do, and she is on your lines the day you switch her on.
She answers
Every call, on the first ring.
At three in the morning, at lunchtime, and on the fourth simultaneous call when your three people are already talking. Nobody hears a queue, nobody leaves a voicemail that gets picked up tomorrow, and the inquiry that would have rung the practice down the road rings you instead.
She qualifies
She asks what you would ask.
Budget, timeline, whether they own the roof, which policy they are calling about, whichever four questions decide who is worth a diary slot. She listens to the answer rather than ticking a box, and she can tell a buyer from a supplier from somebody who dialed the wrong number.
She books
Into the calendar your team already opens.
She reads what is actually free, offers slots that exist, and writes the appointment back where a person will see it. Google Calendar and Microsoft 365 directly, not through a middleman, so the meeting is on the real diary before the caller has hung up.
She writes it up
Before the call has finished ringing off.
Transcript, summary, sentiment, a score, and the fields on the record updated. The thing your reps do at six in the evening from memory, done properly, at the moment it happened, by somebody who was listening.
She follows up
The text, the email, the task.
They go out from the same system that took the call, so nothing waits on a person remembering. The confirmation, the directions, the quote request that needs to reach your ops team by lunchtime. It has already gone.
She hands over
The moment it stops being her job.
Somebody asks for a person, or asks something she has not been given, and the call goes to a human with the entire conversation attached. No repeating, no starting again, no being told to hold while an AI works out that it cannot help.
The real question is what stops her saying something that costs you.
It is the objection every serious buyer arrives with, and it should be. An AI that can say anything to your customers is a liability with a voice. Four things are in place, on every plan, and all four are checkable.
She only says what you give her
Her answers come from the script, the knowledge base and the allowed answers you set up. She is not improvising from a general model and she cannot invent a price, a discount or a delivery date that is not in front of her.
She hands over rather than guesses
The moment somebody asks for a person, or asks something outside what she has been given, the call goes to a human with the whole conversation attached. She does not bluff to keep the call.
Every call she takes is recorded and transcribed
Same recording, same transcript, same summary and sentiment as any call a person takes. Nothing she says to a customer is unauditable, and you can go and read it.
She is scored like an agent
The quality pass that scores your reps runs on her too. If she is handling something badly you find out from the same report that tells you a person is, and you change what she was given.
The fastest way to test all four is to ring her and try to make her say something she should not. How we handle your data
Interview her
You do not have to take our word for any of this.
Ring the number. Ask her something awkward. Ask for a person and see how fast she gives you one. It is our own main line, and she has been answering it since August.
What the AI does on every other call →