基于大语言模型与边缘交付的敏捷站点超个性化构建

Hello. Thank you for coming. Um I'm going to talk to you about Agility Sites, how we call it as building hyper-personalized websites. I'm not going to just talk about it. I'm going to show you what we're building.

Um I've been working on on this project for for a bit now, and we'll try to show you what is possible today with with AI.

Uh I work at Adobe at a I'm a principal scientist at a product that not many people know, Adobe Experience Manager, content management. We run a lot of uh website properties for big brands, and my background is in in open source, uh contributing to to a lot of foundations and projects.

What are Agility Sites, and how are we building this thing?

So, we're looking for sites that are uh looking at the what intent the user browsing uh has. What is the user doing? What is the user trying to achieve? And the end goal is to personalize these pages for the for the current user browsing, so that eventually this uh drives uh higher engagement or uh conversions, whatever the marketing teams want to want to achieve.

And these pages are personalized in real time based on the on the user that is uh accessing the site, and what is the what is the user doing.

The stack we're using is AMH delivery. So, this is the part of the product we we have, uh where all the content is on the edge and then we have back end service that powers this experience with different LLM providers LLM services we use Cerebras for fast inference or we can use also we tried bedrock and and a bunch of others. I'll be showing Cerebras today and you will see the reason why.

The the engine that is personalizing this this bits is using the rich content and blocks. So different blocks on the site are customized depending on on what the user persona is.

We don't want the the whole site to be generated. I mean if you talk to marketing people they they have a very strict brand guidelines. You don't want to just come up with our have some hallucinations there. So the what is personalized is different sections of the site and we use the whole site as a corpus. We built a rack from the whole site. So what is generated is grounded on on the existing site.

We tried to solve the problem where one size fits all. We want hyper-personalized experiences. Also we want to help our customers to do more automatic authoring. So not having to create thousands of different variations of the site but use AI for this and then do these multiple layers of of personalization.

Some examples of what we're doing or I'll show in the demo. It's instant persona adaptation, query generation when the user search for something on the site, the page with the results is customized for them and also uh, something like recommendations where after you browse the site for a period of time, we we can create a page that recommends something based on on on what you are what we think you are looking for.

For marketers, uh, they can define this strategy on natural language, and they can use analytics to to drive the loop of personalization, and what is the end goal, and how this goes back again to change to adapt the personalization to improve that uh, whole cycle. Everybody's talking about loops in this conference, so that's that's one of the loops there.

How the architecture look like? So, it's a dynamic front end with some blocks, what I mentioned before, and with uh, edge delivery services is basically you compose these blocks, and uh, they are updated on on real time through with the AI.

The back end, uh, we we do the um, evaluation of the models and the providers, and one thing we realized is is that this is very dependent on the site. So, we have a bunch of prompts, and we look uh, we run it across a huge variety of uh, models and providers, and then we look at the accuracy, we look at the speed, but this is going to depend highly on what type of site, like how big is the site, how I don't know, what different what different um, area is the site targeting, what what type of commerce it is, and so on. So, we we run this this um, evaluation continuously. We use uh, Promptfoo. Uh, anybody heard about Promptfoo?

Okay, some people.

So, Promptfoo allows you to evaluate models um, prompts against my multiple models, providers, and you can do local models and any of the a bunch of open AI compatible uh providers and and uh a lot of them, basically.

We look for two things. Why? Accuracy. That's that's typically what people look for, but also we want the speed because we don't want the site generation to take more than 1 or 2 seconds, right? Because people uh this is already uh proven that people want the the faster the site, the more conversions it it generates or the the better the experience it is for the user.

Um yeah, what I mentioned is different sites may have different requirements. Uh so, you may have to run this uh evaluation of models depending on the site.

This is a an ex uh we we secured this some of these queries, so we have a 15 prompts for this example site.

Um we have uh at the top you can see with Cerebras on the Gemma 4 model that was announced last last week, we can get an average latency of 1.1 seconds generating a page.

You you can compare that to the second one, which is 4.6 seconds, right? So, the difference is huge. And that's why uh we use Cerebras for for this use case.

And uh you can see that different providers, different models have different um different speeds. And here is uh let me I can show you the whole thing here. Not this one, this one, right? So, at the at the bottom we have other other tasks.

Sometimes uh maybe some of them may be good. They don't need to be perfect, but they're good enough if they're fast enough. So, that's going to be the the kind of decisions that you need to make on whether the model is good enough for your use case or not.

Yeah, we're looking Yeah, average 1.1 seconds. And then the the next ones are going from 4 seconds higher. And you don't need a huge LLM to do this sort of work because you are generating text, you are deciding where to put blocks and how to organize the website, you don't need a lots of information for that.

So, this browsing and the queries uh is are being recorded. So, these are the metrics or the the the data we gather from the user, and this is fed into the LLM to personalize the site. And then in this example, we personalize the hero card, the products, the blog feeds, and and the navigation based based on the persona.

Also, what are some of the buttons like our call to action navigation, you can also we can also personalize those.

We we create and I'll show you the a for you page, which is a recommendation. And this is a interesting one because this you could pre-generate, right? As the user browses your site, you gather these signals, and you could keep generating in this. So, in this case, you wouldn't need so such a big speed.

But but that's interesting because it it would be if a user wanted to buy something, you could just say, "Okay, for you, I will recommend these three products or or something like that."

Um Yeah, and then they can see this recommendation, and if they go there, that that could be pre-fetched for them. And obviously, you have to keep updating it as the user navigates around the site and and so on. So that that's also something to consider on the cost cost implications of doing multiple generations, multiple LLM calls.

When when the user runs a query, dynamic personalized page is shown to them.

When the these queries are also grouped into personas or intent types. So what is this guy what is this guy trying to do in the site? Is trying to buy something? Is trying to just get information? So you can get marketers to decide what type of groups, how many groups you want to have, how you want to deal with with customers. And the AI will choose the the blocks and the suggestions for for those groups of people.

Um And we can adopt yes, the the different blocks, the the the sequence of the blocks and media. You could also do media. One of the things we consider is there was some a model announced today or yesterday the the nano banana light. So

could

even

generate

images

very

fast

on

the

fly.

Obviously

not

as

fast

as

text,

but

that's

also

something

that

would

be

I

don't

I

don't

know

if

it's

that

something

like

marketing

people

would

want

to

have

generated

images.

That

depends

on

on

the

quality

a

lot

if

it's

on

brand.

the

site

in

this

example

we

have

a

a

product

site

and

then

we

have

guides,

experiences,

blocks

and

the

whole

response

of

the

LLM

is

grounded

there.

And

there's

comparisons.

We

can

do

comparisons

between

products

that

are

tailor

and

the

product

pages

can

be

tailored

for

the

for

the

user.

this

is

this

is

a

bit

of

the

stack.

Um

not

going

to

spend

too

much

time

here,

but

the

browser

you

have

some

layers.

You

have

the

browser

where

the

signals

get

get

uh

from

I

got

I

got

I

got

so

I

got

from

the

from

the

user

and

then

we

have

the

back

end.

Uh

we

can

have

the

back

end.

We

run

this

some

of

these

in

in

Google.

Some

of

these

are

in

our

Cloudflare.

So,

the

back

end

is

basically

just

calling

the

LLM

and

doing

some

reasoning

using

the

rack

that

is

built

on

on

the

site

to

do

the

generation.

you

have

obviously

you

have

to

have

the

vector

database,

the

inference

uh

machinery

and

uh

that

obvious

business

manager

is

doing

the

serving

the

the

at

the

edge

is

serving

the

the

pages

and

the

static

content.

let

me

show

you

because

I

think

this

is

uh

so,

we

call

this

uh

audience

of

one

because

the

idea

of

in

marketing

they

they

always

dream

on

being

able

to

personalize

things

for

each

individual.

So,

we

call

it

yeah

audience

of

one.

So,

I

have

this

this

site.

Uh

this

is

a

site

that

is

absolutely

generated

uh

example

site.

It's

a

coffee

uh

machinery.

So,

I

can

go

and

and

read

some

stories

and

I

can

go

and

look

at

some

products.

go

and

look

at

this

product.

I

can

spend

some

time

here.

Uh

let's

go

and

click

here.

Okay,

so

I'm

I'm

browsing

around

the

site

and

I

have

this

debugging

tool

thing

uh

which

Uh

let

me

go

here,

I

think.

Let's

see.

down

there

is

the

signals

that

the

that

the

browsing

is

giving

us.

So,

I

don't

know

if

you

can

see

it

much

because

I

cannot

see

it

much.

The

So,

the

user

is

bucketed

into

the

exploring

category.

We

have

the

pages

that

have

have

visited,

and

then

we

have

how

much

time

is

spending

on

each

page.

All

of

this

data

is

now

available

for

the

LLM.

if

I

go

here,

I

already

have

a

for

you

page

that

was

generated

for

me

and

based

on

my

browser.

And

you

will

not

notice

that

it's

slightly

different

than

everything

else,

but

if

I

go

here

and

I

run

a

query

like

I

want

I'm

looking

for

a

coffee

machine

to

uh

prepare

coffee

while

camping.

site

is

this

was

just

generated

for

me.

And

then

you're

going

to

see

some

things

like

the

text

is

customized.

Camping

shouldn't

mean

compromising

on

your

uh

whatever

routine.

Uh

the

coffee

tips

for

camping

um

machinery

that

are

being

recommended

are

coffee

agile

and

um

or

the

nano,

which

are

good

for

for

the

for

a

camping

trip,

right?

you

saw

how

fast

this

was.

I'm

going

to

run

it

here

something

similar

that

I

had

here

and

I

can

run

it

on

the

debug

mode

here.

And

you

will

see,

let's

make

this

bigger.

time

164

seconds

to

generate

the

page.

So,

this

includes

a

round

trip

to

the

LLM.

This

is

using

Cerebras

Gemma

So,

the

the

Gemma

model

from

Google

running

on

Cerebras

on

their

very

fast

chips.

Uh

we

get

2,300

tokens

per

second.

Which

is

not

bad.

I

would

say.

if

I

run

it

again,

uh

probably

something

like

that.

Uh

the

LLM

time

is

1

second.

And

again,

2,200

tokens

per

second.

This

is

something

that

we

only

dreamed

about

before.

the

on

this

site

example

site,

we

have

some

other

options.

Uh

so,

because

we

we've

been

showing

this

to

customers,

so

we

have

the

the

ability

to

change

the

different

models,

temper

temperature,

tokens,

and

so

on.

And

we

can

uh

we

can

show

uh

and

try

the

different

models

and

see

how

they

behave.

Besides

the

automatic

test

with

Prompt

Full,

then

we

can

uh

manually

come

and

and

click

things

and

see

and

see

how

that

how

that

works.

uh

we

also

have

uh

OfOneLabs.

So,

we

have

we

build

this

tool

that

generates

an

agentic

site

for

any

site

we

want.

So,

if

somebody

wants

to

have

a

demo

for

a

customer,

come

here

and

enter

the

URL.

In

less

than

an

hour,

you

have

an

agentic

site.

I

did

this

last

week

with

the

AI

engineering

site.

I

got

this

site

that

is

just

a

search

box

and

a

few

things.

And

let

me

open

it

here,

the

full

page.

Not

this

one.

Yeah,

okay.

I

could

say

Europe

AI

conferences.

So,

these

suggestions

are

also

AI

generated.

And

I

get

a

page

that

is

more

focused

on

It

should

be

more

focused

on

on

the

on

this

European

conferences.

I

go

back,

did

I

go

I

can

search

for

anything

the

same

way

I

did

with

with

the

Arco.

So,

I

as

a

specific

There

was

someone

that

was

generating

a

good

comparison

side

to

side.

Let

me

see

if

this

one.

Okay,

here.

This

one.

went

and

this

generated

a

page

with

a

pretty

good

comparison.

If

I'm

looking

at

two

conferences

and

I

need

to

decide,

if

I

figure

out

that

the

user

wants

to

do

that,

this

is

great

because

that

gives

them

a

side-by-side

comparison

on

the

fly.

this

this

is

I

think

this

is

cool

already,

but

then

we

have

I

have

this

idea

that

probably

the

I'm

a

bunch

of

people

are

we

are

talking

about

is

the

web

that

is

is

the

web

the

future

still

and

so

on.

Nobody

knows.

But

we

can

also

do

something

with

this

with

this

audience

of

one,

this

generative

sites.

So,

imagine

you

have

you

have

your

personal

assistant

and

you

ask

a

query

through

in

this

case

through

Google

and

you

say

I

want

to

buy

I

don't

remember

what

the

query

said.

It

was

something

like

I

want

to

buy

a

machine

and

I

get

this

on

my

Google

TV.

Right?

So,

this

is

absolutely

personalized

to

my

query.

No,

go

back.

This

is

absolutely

personalized

to

my

query.

So,

I'm

there

in

my

living

room.

I

don't

need

a

phone,

I

don't

need

a

computer,

I

don't

need

anything,

just

my

voice

and

something

that

will

kind

of

show

me

something

that

is

absolutely

personalized

to

to

me.

Okay,

so

that

one.

what

I

was

trying

to

show

and

hopefully

you

remember

from

this

session

is

that

this

is

now

possible.

It's

only

going

to

get

better

from

here

on.

It's

only

going

to

get

cheaper,

it's

only

going

to

get

faster.

And

you

will

uh

be

able

to

have

uh

huge

personalization

options

for

sites

and

for

other

things.

And

you

can

do

this

with

intent

driven.

So,

what

is

the

what

is

my

user

trying

to

do?

What

does

my

user

want

to

buy?

These

sort

of

questions.

And

you

can

uh

assemble

a

page

just

for

them.

And

you

can

also

do

this

with

uh

multiple

models

and

and

eventually

it's

just

going

to

be

faster

and

faster,

right?

that's

it.

Um

thank

you

for

coming

and

I

hope

you

you

got

the

idea.

Thanks.