Crayon Data Crayon Data Tangram AI
Anticipatory AI · Fully agentic · ✷
BANKING · CUSTOMER EXPERIENCE & VALUE SUITE
A N T I C I P A T O R Y   A I

Reads location, transaction, interaction and external data before the customer asks -
ranking every offer by taste, not by who paid.

38% search by voice ✦ 60% buy inside a livestream ✦ very few banks with GenAI in-app
Why Anticipatory AI exists

Customers signal constantly. Banks wait to be asked.

The problem

Acquired, then ignored

The bank signs the customer up, then waits to be asked instead of watching what they do next.

  • 38% of Indonesians already search by voice — most banking apps don’t listen
  • 60% of buyers purchase inside a livestream — most banking apps don’t watch
  • Acquire, then wait to be told what the customer wants next
Today, how it operates

Static rewards, generic pushes

The same reward goes out to everyone, regardless of what they spend, the time, or what's already booked.

  • Same push whatever they spend, the time, or the calendar
  • Pay to rank offers ordered by merchant spend, not customer taste
  • Never notices a booked flight the banking app had no part in
With Anticipatory AI

The need, anticipated

The engine reads the signal as it happens and turns it into an offer before the customer asks.

  • Fused, live transaction, time, calendar and bookings, read continuously
  • Genome-ranked offers ordered by taste, never by who paid
  • Already planned a detected flight becomes a trip, not a notification
How it works · three moves

Read it. Rank it. Plan it.

01 Read the Signal Transaction, time, calendar and bookings, fused live
  • Three kopi runs this week become a “same-day” filter
  • A Garuda booking becomes “Bali, anticipated”
  • Every card carries a match %, not a guess
02 Rank by Genome Offers ordered by taste, never by who paid
  • Taste-graph ranked, e.g. a 93% match
  • Vibes: swipeable, reel-style discovery
  • Grid view for browsing the full matched set
03 Plan the Moment From a detected trip to a booked one
  • Upcoming trip surfaced automatically
  • Add stay, activities, airport cab — one tap each
  • Book in seconds, no separate app
The capability set

Six things it does, before the customer asks.

One profile, read the same way across Now, Vibes, Travel and You.

01 Contextual signal reading Transaction, time of day, calendar and bookings, fused live
02 Trip detection A flight booking becomes a travel plan, automatically
03 Taste-graph ranking Offers ordered by genome, not by merchant spend
04 Conversational discovery Ask anything, get ranked picks back
05 Reel-style browsing Swipe Vibes, or open the full grid
06 Book in seconds Stay, activities and transfers, one tap away
Now · Putri, Jakarta

Es kopi susu o’clock. 93% match.

read from transaction, time and taste

  1. 1
    Three kopi runs this week become a filter The same-day pick surfaces automatically, no one asked
  2. 2
    A Garuda booking becomes a trip Bali, anticipated — travel mode switched on by itself
  3. 3
    4 offers, ranked by genome Not by who paid for placement
1 2 3
Vibes · Ask it anything

Show me culture picks. Ranked by you.

conversational, not a search bar

  1. 1
    Culture that fits you M Bloc Space 88%, Pasar Santa 86%, Museum MACAN 85%
  2. 2
    Every match, explained Ranked by taste, context, influence and behavior
  3. 3
    Suggested, not searched Prompts anticipate the next question
1 2 3
Vibes · Discovery

Every offer, ranked for one person.

the full grid, matched to Putri

  1. 1
    Every card, genome-matched Apple, McDonald’s, Starbucks, Zara — all ranked for Putri specifically
  2. 2
    Two apps, one ranked feed Grab’s rain discount and Gojek’s rain voucher, pulled into Putri’s single feed
  3. 3
    One tap to open Every card opens straight into the merchant’s own offer
1 2 3
Travel · Anticipated

Where to, Putri? Already answered.

detected from a Garuda booking

  1. 1
    Bali, in 5 days The whole itinerary opens pre-filled, not just a flight reminder
  2. 2
    Add stay, activities, cab One tap each, no separate app
  3. 3
    Book in seconds The whole trip, not just the flight
1 2 3
You · Profile

The genome, behind every card.

8,460 points · 93% taste match

  1. 1
    4 offers unlocked From one profile, ranked continuously
  2. 2
    93% taste match The number every other screen is built on
  3. 3
    One profile, everywhere Now, Vibes and Travel all read the same genome
1 2 3
In summary

The need, anticipated.

It reads location, transaction, interaction and external data before the ask, and ranks
every offer by taste — not by who paid, for a market where 0% already
search by voice, 0% buy inside a livestream, and very few banks do this today.

Try it first — pick a persona in the hosted sandbox.

demo.crayondata.ai/anticipatory-ai

Anticipatory AI · Crayon Data

Thank you.

Scan for the live portal

QR code to the live portal