In bedtime story, Birbal understood the Intent .. AI is just beginning to.

𝗜𝗻 𝗯𝗲𝗱𝘁𝗶𝗺𝗲 𝘀𝘁𝗼𝗿𝘆 , 𝗕𝗶𝗿𝗯𝗮𝗹 𝘂𝗻𝗱𝗲𝗿𝘀𝘁𝗼𝗼𝗱 𝘁𝗵𝗲 𝗜𝗻𝘁𝗲𝗻𝘁 .. 𝗔𝗜 𝗶𝘀 𝗷𝘂𝘀𝘁 𝗯𝗲𝗴𝗶𝗻𝗻𝗶𝗻𝗴 𝘁𝗼.

There is an old Akbar–Birbal story I was reminded of recently.

Akbar’s wife wanted her brother to be given an important responsibility in the court.

Akbar decided to test him.

Some bullock carts were passing nearby.

Akbar asked:

“Find out where those carts are going.”

The man ran out, returned and answered.

Akbar then asked:

“What are they carrying?”

Off he went again.

He returned.

“How much grain?”

Another trip.

“Where did it come from?”

Another trip.

Every question got an answer.

Then Akbar called Birbal.

Birbal went out once.

When he returned, he knew where the carts had come from, where they were going, what they carried, the quantity, the price and the opportunity available.

In some versions of the story, he had gone even further and arranged for the grain to fill the royal godown.

Same task .. very different intelligence.

The difference?

“One person answered Akbar’s questions.
Birbal understood Akbar’s intent.”

And that made me think about AI. A lot of our interaction with AI still looks like this:

– Prompt
– Answer
– Follow-up prompt
– Answer
– Clarification
– Another answer

.. technically correct perhaps but essentially a Q&A machine.

The more interesting AI is beginning to do something different.

It asks:

“What is this person actually trying to accomplish?”

If I ask AI:

“Which hotel is closest to my meeting?”

perhaps I don’t really want a distance. I want to reach my meeting reliably and comfortably on time.

A good AI may therefore consider traffic , check-in time , transport availability , neighbourhood , price and my schedule before I individually ask about each one.

That is the shift:

Question >> Answer

becomes

Question >> Intent >> Context >> Anticipation >> Outcome

But there is a fine line here.

Anticipation shouldn’t become assumption.

Sometimes AI should infer .. sometimes it should ask .. and sometimes it should simply say:

“I think I understand what you’re trying to achieve. Shall I proceed?”

Perhaps truly intelligent AI won’t be measured by how well it answers our questions but it will be measured by how often it understands the question we were going to ask next.

Birbal figured that out centuries ago.

We are now teaching machines to do the same.

Until next time .. happy thinking!!

Author – Sumit Rajwade, Co-founder: mPrompto