Launching AIP Now Customer Service Engine | Lowe's + Palantir

Launching AIP Now Customer Service Engine | Lowe's + Palantir - 雙語字幕

Okay.

Great.

Good afternoon, everyone.

It's quite exciting to see.

I think we have hundreds of participants joining from across the globe.

So, to perhaps kick us off, my name's Anurag Bal.

I had the AIPE platform implementation teams at Palantir.

Today we're launching the first of our AIP now applications which are intended to provide

AI ready applications with a few clicks to both existing customers as well as our hopefully new partners in the future.

Today, we're launching the AIP Customer Service Engine.

And I'm joined by my colleagues, Anirud and Yusef, from as well as our partner from Lowe's, Nishant.

And I'll hand it over to Anirud to perhaps kick us off.

Awesome.

Thanks, Anurag.

My name is Anurag and I lead some of our work in retail and consumer globally.

I'm also one of the AIP leads here at Palantir and I'm so excited to launch the AIP customer service

engine today because customer service is a thing that we've all felt pain and we all viscerally relate to.

So while people trickle in, I'd like to share a personal as a customer of a brand that I love.

So I'm a member of a fitness club here in London.

They're a medium-sized business and they pride themselves in their customer service.

And the customer service is actually amazing when you go into the club and the team members are friendly,

the trainers are great, but when you have to get anything operational or administrative done, it is a bit of a pain.

The only way to communicate with them is through email, and every time you send an email, you get a can response back.

So, a few months ago I was travelling and I had to put my membership on freeze, so I sent them an email.

And I got a response immediately by an email which was like, is, you know, we're getting a large volume of customer queries.

And wait 72 hours.

We'll get back to you.

I was like, OK.

A weeks ago, I forgot my umbrella there.

I live in London.

You need your umbrella here.

So I wrote them another query.

I was like, oh, did you find this?

And the same cameras, once again, we're receiving a large volume of customer queries, you please wait, send me two hours.

I didn't find that umbrella.

And it's happened a couple more times,

and you can say I, you know, even on my gym more than the average person, but that's beside the point.

So a few days ago, I wrote a very thoughtful email to them, inviting them to this webinar.

And I want them to improve their customer service, and we think we have a product to help them.

And no points for guessing the response I got from them.

We're getting a high volume of customer queries.

Please wait 72 hours.

So, if there's anyone here from this club who has joined us today, awesome job, can't wait to start building with you.

If not, well, you're all at least 72 hours ahead of the curve.

So, let's get started.

Lowe's has joined us today.

They're an AI leader in the retail industry, and they're one of the biggest home improvement retailers out there.

They're a Fortune 50 company, and they're a first mover when it comes to customer service AI.

Mm-Shant is an engineering and AI leader at Lowe's who's driving a lot of this work,

and he's going to talk about the story and the transformation that he's driving with his team at Lowe's.

Then we're going to peel back the layers of the AIP Customer Service Engine.

This is the application that we're launching today.

Use of an eye and we go through a demo of the product and show you what it looks like.

Then finally, you leave with the keys to your own instance of the Customer Service Engine.

So, with that, over to Ms.

Shand, take it away.

Everyone, I hope everyone is doing great.

So, so as an I lead the AI and data products for our merchandising organization

here at Lowe's and today I'm going to walk you through the work which we have been

doing with Palantir when it comes to one of our customer service centers.

So let's move on to the next slide.

Cool.

So So, right of the bat, right?

So, before we get into what the business problem is and, you what the solution is,

we just wanted to kind of share how, like, what are the different things which we have observed so far.

So, solution, you know, we were kind of lagging in our ability to handle the call volume, which

was coming in, and because of that, our overdue act.

Because, you know, we started like building a lag and it kind of was becoming impossible to get around that.

And at that point in time, we were working in Palantir on some other use case as well.

And we were like, let's let's try a POC on this one.

problem.

So right off the bat, we kind of identified a pilot region where we wanted to run this.

And after completing the POC, we immediately saw like a 75% reduction in our overdue activities, which is a huge number.

And as these, like it has a ripple effect, right?

So you can do more with the same number of agents.

So that was great.

And on the right,

you will see like this,

which kind of shows how quickly we were able to move with Palantir as a tool,

so that initial graph in January, that's our pilot, we brought a few more people, and within four weeks, we were nationally launched.

And in general,

the kind of feedback which we have received from the people who are actually using the tool,

they don't want to go back, so you will observe that again post February, there's a huge stickiness to the whole thing.

They love the tool, and it has definitely improved how they kind of go about their day.

So other than that,

you know,

the old tool,

it wasn't like super transparent and,

you know, once we get to the demo, you will see all the data is kind of available at an agent's fingertip.

So that was the agent story.

Now, the people who manage these agents,

they also never had this kind of

visibility and transparency that how an agent is performing or let's say somebody calls and sick at the last moment,

but they still have 25 tasks assigned to them.

So at that point in time, it has become like super convenient for managers to even go and manage the resources.

in a good way so we can kind of attack the most important problems in a very prioritized way.

So over to the next slide and I'll kind of explain

what kind of a business unit for which

and hopefully in this slide I'll try to build like an intuition that why this is a hard problem.

So stay with me.

So I'm gonna take an example here.

Let's say that you're trying to do a kitchen remodel in your house and you want to work with Lowe's, right?

So there are kind of three ways which you can go about it.

One is DIY, which we call our do it yourself.

So let's say you're super handy and like You just know how to kind of build a Now,

very, very less number of people will build a kitchen, maybe smaller things.

We do see we have like very active DIY community.

But the other two options are that you go and hire which is basically what we, at Lowe's, we call them pros.

So go and hire a company which does kitchen remodel,

and then they can come at Lowe's and buy everything which you eventually want in your kitchen.

The third way to do it is, and that's what we're going to be talking about today, is Lowe's installation services.

So and and and why like this is such a hard problem right and I just want everybody to kind of think through it like if you are doing a kitchen remodel end to end right.

You will have multiple specialized people coming into your home so example is like at some point in time a plumber will come in at some point.

like an electrician will come in at some point in time,

a cabinet installer,

or stone installer, or somebody needs to come in and put the flooring in, somebody needs to come in and do very different things.

And before any of these installers come into your house,

we also need to ensure that the product, which they're coming to install has also been delivered to you.

So, so that's the second complexity.

Third complexity is for any time these people are coming in or any time these products are getting delivered,

you need to be at home as well.

So, so that's why we have set up this very specialized customer service.

system or agents to kind of, you know, lead a customer through the entire journey as their kitchen has been.

And it's very important for us to be super efficient while we are doing this because you are out of a kitchen while this project is going on.

And, know, like there's a limit to how much food you can reheat in a microwave.

So, so, so I think that's why like this was a really important problem for us.

It has direct implications to our customer satisfaction.

And, you know, we are kind of like kind of like dispersing how you live in your own home while these projects are going on.

And that's why we kind of started working with Palantir to see what we can do in this space.

So, so moving on to the next slide.

So, so this was like the problem in real life.

Now, if you go to the next slide, I like, like, what's the equivalent.

So, yeah, so what's the equivalent implication on the technology side.

So, so, so this is an old business which we have been in for a super long time so of course our systems have been aging.

And, and, and, and as I said like initially right like in general the call volume which we have been

has like kind of gone out of control and not having a super efficient system.

Basically is leading to a poor customer performance and projects getting delayed and it affects basically everybody who is in this life cycle or system.

So, and the main problems with the technology stack, which we had, you know, a customer agent was going through multiple screens.

You know,

like when you're calling a customer,

you don't know what all has already been delivered to their house, which all providers have already visited their house.

So, so there was like, it will take you like 17, 20 clicks to kind of.

even a state where you know where the project stands.

So at this point in time,

so you know, like all of the things which I mentioned here, there was a lot of manual for naggling which was going on.

There was not a cool way to manage how the work will get assigned to agents.

So these were the problems which we put in front of the Palantir team and we started looking into it.

So overall,

like if I have to talk about the new solution which we have built in Palantir, it stitches the data really, really well together.

We the ontology which Palantir offers to kind of of piece all the different information together.

So quickly,

a customer agent can see that,

hey, this customer three projects or three work orders have already happened, seven need to happen, these items are delivered.

So a very, very screens, they can quickly see where this customer is in their project journey.

And to manage their work, we have a concept where we are smartly able to.

to Q what is the next best task which an agent should be picking up and there's a list right

in the morning you know when the agent is logging in they know exactly how their day is going to go through.

So even if they have to self optimize a little bit they can totally do that and if you remember

the first graph which you saw adoption has been great we have seen a lot of stickiness from

our agents they are absolutely loving the solution.

So cool.

Next slide, this is like a true quote from one of our business partners, Todd, you know, he's his director for our installation team operations.

And he, like this is a true quote, like all this was achieved in four months.

You know,

we came up with like we have trained the managers on how to train their teams,

and we onboarded like 1000 users in just three weeks.

almost magical like is you know when I've talked to him so he was very very

impressed at how quickly we were able to move you know in this amazing platform.

So, so, so, so what, what, what are we planning now right that hey cool looks like you guys are doing really great you we have unlocked new levels of productivity but of course we don't

want to stop there.

There's a whole another area where you know how the expense requests are created by our providers,

you know, sometimes the providers not even showing up but we end up paying them.

We are looking at the new AIP platform inside Palantir Foundry itself to kind of see if we can take things to the next level and the demo which we have planned for you today.

We'll kind of touch on.

And then in general, we have a couple of other use cases which are going on in Palindir already.

So we are trying to see one of them is our supply chain project.

So we want to see if we can bring like even more data.

So think in your head that If you are a customer and for some reason, your cabinets are not getting delivered.

Now, the agent will probably be able to go into

the supply chain ontology and kind of see that,

hey, this thing is stuck on a ship right now, and can escalate to get that to you.

So are the future looking things which are on the horizon here.

Cool.

I will request,

use of now to come in,

help me out with the demo and hopefully we'll give you a quick idea on what is coming in future and how we are planning to leverage AIP.

So the screen which you see here,

I just want to make sure everybody knows this is very real,

but with fake data because we don't want to share real customer data so cool.

So on the left,

you are seeing the activity queue and that's what I was mentioning that agent is able to see

that these are the different things which they have to go through today.

So at the top now, if we open the first, Yeah, the first one.

So now the AIP we are going to the next step.

What I mean by that is like the screens were showing that, hey, these are the four or five work orders which have happened.

These are the two future work orders which are going to be happening.

And remember, these agents are closing their existing tasks, putting in some notes in there.

So now we are able to kind of get the power of gen AI,

we are taking all the notes and all the interactions which have happened for a project so far.

And we give a very, very clean.

to the agent that send an email to X, Y, Z confirm the installation date with them.

And then once that confirmation will come,

somebody from our team needs to call our customer, Cedric Abort in this case, and kind of confirm that that install it.

Like, what will be the installation date?

When will they be at home?

So, so, so it's like really, really powerful because to just come to this message,

it might have taken an agent over,

even with the new intuitive screens,

just to kind of piece together the whole information,

it might have still taken them 15 to 20 minutes because these are very, very complex projects in general.

So yeah,

this is really powerful then of course once we started integrating Jenny I we started thinking where else we can do it so if if we click on email installer now.

Now, a beautiful email, a pre-written email kind of comes to you.

You can, with the customer name mentioned there, with work order mentioned there.

So, so again, a very, very intuitive case to be tried in JNI.

But because we had all the other data stitched, we were able to quickly bring it to this experience.

The second thing, which I will like to, so, so if you get out of this now and create follow up task.

If you, if you go on.

and basically view details step by step, so if we, yeah, so yeah, click on that, perfect.

So, so here you can see that we took one of the SOPs which we want our agents to be using.

And we kind of personalized it to this particular task.

And at the end,

let's say agent also wants to see that,

hey, you know, what's the SOP documentation they can do that as well, but we use Janai to read our SOP.

We used GenAI a second action to read where the task is right now and kind of fuse together to give a very personalized recommendation that how you need to be creating a follow-up task.

So that's how we are using GenAI to kind of think the timing and ensuring that our agents are following the SOPs as well.

So it's, I think, a really cool way to use GenAI.

Cool.

So if you can get out of this and just kind of open the second task as well because that's more of a customer focusing.

Perfect.

Yeah.

So here it's just again like different flavor of it.

Here we are saying.

and discuss,

you know,

when the installation will happen and,

and, and, you know, so, so we can, so on both the sides agents are calling providers as well agents are calling customers as well.

So how are we, how can we use like Jenny I to make them go faster.

That's the main intuition behind building the solution.

way.

So once we were done with all of this, we were like, why stop here, right?

So one of the other complaints which we hear from our providers is that you guys end up calling us three times,

four times in a day.

So we decided to do this in a bulk way.

So yeah, if we go there.

So you can see that email installer.

So we know that today we have to, this agent has to reach out to six different installers.

And for different installers, they are addressing maybe more than one customer at a time, right?

So instead of calling them six times or sending them emails,

you know, six times, so what we are doing is can we just send a combined email?

So, so if you open the send combine email, so here instead, so now it's like that, hey, you know, today you have these three jobs.

So we kind of use whatever we had built so far, which was, you know, helping agent kind of move faster.

So, so it's like these mini boosters.

Like a Lego,

we are building more and more stuff with Jenny and we are hoping that this will unleash a completely different level of productivity and we'll keep on moving through these projects

much faster.

We'll be able to keep our customers way more informed.

And I think overall this lead to a much better customer satisfaction ratings for these projects.

So that was the demo for today.

Hopefully you guys enjoyed it.

And now over to Annie.

I'll take any questions, any which have been asked so far and we can kind of tag team for that.

Great.

Thank you so much.

We'll be now moving into a short Q&A before proceeding into how you can actually request your own customer service engine application.

For those who are interested in asking a please go ahead and submit.

your questions through the chat, we're going to try to answer them live, but if not, we'll also follow up offline after.

So perhaps in the interest of time, Nishant, one question for you.

What did you think was differentiated in building with AIP as you went from, you know, the first iteration, and then on to production.

Yeah, so see the biggest call out has to be the speed.

So it was like, you know, once we got our data, the way we wanted it once we got that ontology piece built.

and we were able to move super fast.

So going from POC to pilot to I think it is one of the easiest launch which I have seen and I think it's it's just because there's no like once once the data is available

in Palantir, there's no like no more middleware involved, you know infrastructure is already available.

So, to interrupt that's at least what I would say that you know like and in general right like you're taking feedback from the

quickly bringing it to the screens directly, getting them super comfortable with that, because we need their buy-in to do a national launch.

So I think that was at least my favorite part of the whole thing, that how quickly we were able to get through this.

That's awesome.

Thank you so much in the shot and thanks for your time.

I know we have a few more questions, a lot of interest of folks wanting to see under the hood of the customer service.

And so perhaps with that,

I'd like to hand it over to my colleague Yusev,

who actually give a demo, what the AIP customer service engine, and what does one get, they were to request this application?

Before we get into the demo,

quick couple of slides on what the customer service engine actually is, and thank you Nishant for sharing your experience.

We're also excited to see what comes next for Lowe's and for everyone else on the call,

you know,

having heard from Lowe's being the first Uber in their industry with AI and building the customer service transformation that you saw,

hopefully everyone's feeling energized and inspired by the potential that AI can bring in this space.

And I want to tell you that this impact and energy is not just confined to the retail industry.

Those are the first movers in that space, but equally our partners across a range of industries are building their own customer.

use cases.

And in building this and putting these in production,

we've been able to find common patterns that are required to put such a use case in a production ready setting.

And we've distilled that into the product that we're launching today, which is the AIT customer service.

engine.

It's packed with features, but to name five that I think are completely differentiated from anything else that's out there is number one.

The AIP customer service engine is built to learn the context of your business and the needs of your customers.

And it does that through interactions and feedback from your expert operators, your people.

Secondly, it's not a black box.

It comes with the best AI engineering toolkit that's out there today.

So engineers can pull the hood, look underneath it and touch the engine itself.

Thirdly, it's interoperable with any kind of target architecture that you might already be.

You can plug this into a custom React app and use it as a chatbot or you can build something

more complex and operationally relevant like what Nishant showed.

Fourth, it's not going to be complex to connect this with your data sources.

The AI is only as good as the data that you plug into it.

We think that's really seriously here.

So we've shipped this with over 100 different connectors to pull data from any source at any scale.

And finally,

it comes with a built-in analytics engine to make it easy for you to track KPIs and set targets on things like ticket resolution.

times, customer satisfaction, and do that in real-time fashion.

And speaking of business value,

the sky is the limit of what's possible here,

but some of the immediate value drivers are efficiency gains through reduced ticket resolution time,

a reduction in repeat just responding faster, it's responding more accurately with the entire context of your business.

And similarly, it's responding in a more tailored fashion to the customer that's creating the query by having access to the customer 360 data asset.

And finally, you can even plug this into your cross-selling upselling machines and drive the customer lifetime value up.

So now, you might be thinking, okay, AI produces business value, what's new?

The thing that's new here is it's not business value that you get a quarter from now.

This is business value that's available to you immediately.

And that's what AIP now is.

So, the first thing you get out of this is it completely bootstraps your first production

version of a customer service AI, and there are three key building blocks to that.

The first one is an out-of-the-box customer service on PolyG,

which comes with concepts like a customer, their transactions, their orders, and much more.

This is your enterprise truth.

The second building block here is a pre-configured AI agent that is tailor-made to triage and

resolve cases and it is deeply embedded into the ontology,

so it's embedded and anchored into your enterprise knowledge and updates as that knowledge graph updates.

And then finally it puts your people front and center of the workflow.

So the expert operators who've been doing this for years and years and know your business and your customers well,

they become the people who check the AI's homework.

They approve the responses that the AI is generating.

And when the responses are known, accurate, they can give it feedback so the AI can learn and improve over time.

That feedback is also captured and stored in the on-pology.

So you get that on day one, but you don't have to stop there.

This is a framework that you can use to expand the and plug it into your own organization.

Firstly, you can extend the ontology by connecting to data sources and bringing in more data.

Secondly, you can orchestrate actions.

It comes with about 20 pre-built actions, but it also comes with a blank canvas that you can use to build your organization.

For example,

you might want the AI agent to have the power to update orders in your OMS system,

or send diagnostics to your ERP such as SAP.

And really, the sky is the limit here.

And finally, it comes with a AI engineering workbench that allows your engineers to fine-tune and configure the AI agent itself.

It's prompting which LLM is using.

and much more.

I'll hand over to Yusef now to make all of this real through a demo.

Yusef, take it away.

Hi everyone, thank you for being here.

I'm part of the AIP Customer Service Engine team here at Palantir,

and today I am super proud to introduce you to the Customer Service Engine on AIP now.

What I'm about to show you is a suite of capabilities, including the operator app as we saw Nishan show, but also everything behind it.

It's important to understand these are configurable for your business.

So let's get into the operator app.

Great.

So on the left here, you can see our customer queries.

These can come in through phone, emails, text, web submissions.

On the right, you can see our organizational data, information available.

and in the middle, you can see our AI agent generating these actions.

So for this one, we're being asked to give a realistic ETA.

Here's the query.

And our AI agent has been able to interpret that,

understand that we need to respond with an email, search the organizational data to provide a meaningful, detailed and up-to-date information.

And this is going to really help our L1 customer service,

but we can move past better responses into actually solving the customer solution, solving the customer query.

So let's go to our next one.

Here we can see a request to postpone the order date.

This is asking us to do something.

Our AI agent is able to interpret the request, look through the library of actions and select the modified delivery date action.

And again,

using the request is able to use that context to pre-fill the modified delivery date with the correct date for the next available

order slot.

And next thing I want to do and update my order,

I can see here, that's written back to my ontology, my data, written back to my systems.

The next thing I want to do is update the customer.

So again, this personalized email using the ontology, using the context, we're able to approve and send that and communicate with our customer.

Update them on the most up-to-date view of their order.

But that's not the whole story.

service engine.

Let's go under the hood and see what's driving it.

So this is the engine room.

These are the data and actions feeding our operator app.

But it's important to know, whilst it's feeding our operator app here, it can also feed an external web application.

So you can take the power of the customer service engine with you wherever you need it.

Let's go to our next query.

As you can see,

as I'm clicking through the application, I can see how the, the operator app is drawing upon the organizational data stored in my ontology.

That's because the answer to the customer query isn't blowing in the wind, it's actually stored in the ontology.

Let's have a closer look.

So the customer service engine gets shipped with a customer service ontology.

This includes alerts, order events, customer, If we click here, we can see that the order is being backed by a data set.

This could be any data set in your organization.

As Annie mentioned,

we have hundreds of data connectors with a connector and we have pipeline builder which allows you to integrate your data in minutes and we can cover

that in future sessions.

But for now, let's go back to the operator.

This is an operational application.

That means our users are making decisions.

To make those decisions real, we need to give them actions.

In this case, our customers asked to cancel the order.

This is the opportunity to use the cancel order action.

So here, I can select cancel order, confirm it, it and see that right back to my systems actually update something.

This is an action.

An action can be anything from deleting your order to updating the date to writing back to external systems to generating maintenance supports.

And orders aren't fixed.

We can add, sorry, our actions aren't fixed.

We And that's because our customers queries aren't fixed either.

So this customer service engine allows you to define your own actions for your AI agent to use.

But we can also have AI enabled actions.

So this is our logic.

This is where we're defining our email agent action.

Here we can see the prompt.

We're giving it access to our data and we're giving it the context.

And when it's running, we can see this chain of thought, how it's working through the customer query, the alert up here.

And it's using these tools, using these actions to generate the next response.

And here is our output.

Right, so we have this symphony of data and actions all being orchestrated by our AI agent, but we're missing something.

We're missing it because it's not your customer service engine yet.

We're missing the sheet music to our organization.

Now, this is held in many ways, but often it's some process documents, so we can move

here into our ontology, and we can see some SOPs, and what's backing that are PDFs.

And these PDFs can store our SOPs or process documents.

Pass those PDFs and store them in the ontology and make them available for our AI agents to use.

But not all process knowledge is able to be written down.

There's a certain kind of knowledge,

call it tacit knowledge,

tricks of the trade, tribal Things that you only learn speaking to customers on the ground solving their problems day in, day out.

And this kind of knowledge is hard to learn but even harder to write down.

But it's important to making the customer service engine ours.

So we can see how we capture that in the customer service engine.

So let me back to our application into this request.

a customer's asking about their shipment.

So I can look at the email,

looks pretty good,

I'm ready to send,

but I just catch that actually this order is for an L-shaped desk,

which is a piece of furniture, and I know in my organization, that needs a signature.

So I'm going to edit the response here.

I'm going to say, please.

make sure someone is in to sign and what's going on here is I'm editing my email to make it a better response.

I'm using that experience but actually my feedback agent is also capturing that.

It's And I can tag this because this is relevant to the product category and submit that feedback,

submit my change, and send that email to the customer.

But what's going on here is we're capturing that as tribal knowledge and storing it in our tribal knowledge store.

Here we can see that exact piece of feedback I gave to customer service.

And for any future furniture requests, I will have this context.

So here's a pre-existing one to remind our loyal customers about their loyalty points.

If I move to our next query, I can see that this customer, John Smith, is a loyal customer.

And the email we've generated has included the point about loyalty points.

This is how feedback works in action.

So that's a brief overview of the customer service agent.

It's ready to use out of the box and you can incrementally make it your own.

It fits in with your existing systems, it can ingest data, and it can also write back to that data.

And you can also deploy these capabilities on an external web app.

I'm super excited to see how you guys use it.

And it feels like a really good day to be a customer.

Thank you, Yusuf.

So for our attendees, hopefully everyone got a flavor of what the customer service engine is and what it can drive for your organization.

But we also recognize that the true potential of enterprise AI can only be engaged if you get hands on keyboard.

So today we're offering exclusive access to anyone on this call.

all to get access to the customer service engine.

Just scan that QR code on the slide and register interest.

Once you have access, you can make it your own.

Expand the backing data, the ontology.

Configure your AI agent.

Look under the hood.

Create actions.

and build confidence that this is real AI for your enterprise.

Then deploy the AI solution.

Bring this to prod in your enterprise like those have and generate outside impact.

We'll go to Q&A.

I we have about five minutes left.

So have to take any questions and a hug.

All yours.

Great.

So it seems like we have quite a few questions coming in.

So I'm going to try to make sure we capture as many of those.

perhaps one of the most more common ones that I'm seeing is a question around,

how does the AIP customer service engine compare with other CRM forms like Salesforce and Dynamics?

I can take that one.

In most implementations it extends those platforms where those platforms become the

backing data source and this becomes the intelligence layer and the AI layer on top of it.

We ship this with connectors to both of those systems and more,

like I said over 100 systems to pull data and write back data as operations are happening on the different customer cases

and tickets and so on.

Awesome and then perhaps one last question and then maybe we should put back the slide with with this sign up QR code.

How can we attend a bootcamp?

There's multiple ways of attending a boot camp.

You could follow the link behind this QR code and request a boot camp.

You could also go to pangleture.com and sign up for a boot camp.

And someone from Pangleture will be in touch with you.

Awesome, thank you so much everybody.

I know there are quite a few questions that we still have an answered.

One thing I'll take away is make sure we follow up offline with anyone who has submitted a

question we haven't had a chance to answer.

But once again, thank you so much for joining.

I hope everyone is able to get a chance hands-on access to the AIP customer service engine.

翻譯語言

選擇翻譯語言

視頻總結導出字幕口語評測

解鎖更多功能

安裝 Trancy 擴展，可以解鎖更多功能，包括AI字幕、AI單詞釋義、AI語法分析、AI口語等

兼容主流視頻平台

Trancy 不僅提供對 YouTube、Netflix、Udemy、Disney+、TED、edX、Kehan、Coursera 等平台的雙語字幕支持，還能實現對普通網頁的 AI 劃詞/劃句翻譯、全文沉浸翻譯等功能，真正的語言學習全能助手。

支持全平臺瀏覽器

Trancy 支持全平臺使用，包括iOS Safari瀏覽器擴展

多種觀影模式

支持劇場、閱讀、混合等多種觀影模式，全方位雙語體驗

多種練習模式

支持句子精聽、口語測評、選擇填空、默寫等多種練習方式

AI 視頻總結

使用 OpenAI 對視頻總結，快速視頻概要，掌握關鍵內容

AI 字幕

只需3-5分鐘，即可生成 YouTube AI 字幕，精準且快速

AI 單詞釋義

輕點字幕中的單詞，即可查詢釋義，並有AI釋義賦能

AI 語法分析

對句子進行語法分析，快速理解句子含義，掌握難點語法

更多網頁功能

Trancy 支持視頻雙語字幕同時，還可提供網頁的單詞翻譯和全文翻譯功能