I built this app for Elation Health and it really showcases how smart and contextual onboarding tool, not only how it could be done but how it can add value.
00:13
the tool watches what a clinician does in the ehr.
00:17
It identifies teachable moments and it also communicates with support CSM and product engineering teams with those observations so they can act upon those.
00:30
lots of value be realized from a tool like this.
00:33
first of course our primary objective is to get our doctors to master the tool as quickly as possible.
00:40
we also want them to, we also want to reduce support costs and ultimately we want to accelerate time to value.
00:46
So really getting to that out of the onboarding phase and where we can actually realize value from our, from our customers.
00:52
So let's jump into it.
00:53
So I'm logged in as Dr.
00:55
Mike Teodosio.
00:56
And so I'm logging as a doctor.
00:58
Let me click on patient history here.
01:01
So this is really just the patient directory there.
01:03
I have three patients logged in and of course what I'm seeing here is their SOAP note history.
01:09
So the pre, the notes of the previous visits, there's a knowledge base.
01:13
So here's just some knowledge articles, knowledge based articles.
01:15
And this is actually integration with Confluence.
01:19
So this is real time pulling in our data or articles from Confluence to generate this list.
01:25
I can open them up if I want to see them.
01:26
That's great.
01:28
but really the action kicks off here in the visit note.
01:31
So this is, I'm Dr.
01:33
Mike.
01:33
I had a session with a patient and I'm going to record my findings here.
01:39
So let me select a patient.
01:40
Let me click James Chen, who has hypertension.
01:43
Here's the view history.
01:45
So this is James's he's been here twice before the two notes from his previous sessions.
01:52
And what I can do is I can copy to note.
01:55
So I'm going to copy all of this and put it into the note, page, the active page here.
02:01
That's great.
02:01
Now you'll notice on the right side we have something called the guidance stocks.
02:04
This is kind of the onboarding element here.
02:06
And so it's noticed my activities, what I clicked on.
02:10
So by the way you copy and pasted, there's a safety check.
02:13
Be sure you update the information.
02:15
So just don't let's, let's not forget to to update all of these fields.
02:19
And second there's a shortcut.
02:21
I can push control Enter that'll automatically submit this, note and save the note.
02:25
And I want to call out something called live signals.
02:28
Let me expand that.
02:29
So again.
02:29
And notice that I copy and paste it and it's sending that information to our CSM and our support and our product teams.
02:36
We'll get into that in a moment, but let me go ahead and collapse that.
02:38
And you'll also notice here at the top there's a, progress bar.
02:41
So this is a bit of a gamification, mechanism where I can see, as Dr.
02:47
Mike, I can see how close am I to getting 100% mastery of this tool by am I clicking on everything?
02:54
Am I, noticing all the features.
02:56
Features and all, all the, AI enhancements here.
02:59
So let's go ahead.
03:00
And I'm going to delete here what I copied over into the subjective field.
03:04
I'm going to paste something that I've already copied into my, clipboard.
03:09
You'll see the guidance doc adjusted again.
03:12
So I'm here.
03:13
I'm going to scroll down.
03:14
Most notable is that AI Assistant is now available.
03:17
So it's calling out this panel here, which is AI Insights.
03:21
And this is really the first introduction of AI.
03:23
So let me go ahead and click View.
03:25
Now.
03:25
AI generated this.
03:26
So AI watched what I was typing.
03:29
And it's also looking into the patient's note history.
03:32
It's detecting a pattern.
03:34
So pattern detected across prior visits.
03:36
Fainting and dizziness is documented for this diabetic patient.
03:40
Use the patient history to review, context.
03:43
So what, it's advice.
03:44
This is not medical advice.
03:45
What it is, it's tool advice trying to get me as Dr.
03:49
Mike to master the tool.
03:50
I should be able to access quickly access the view history.
03:53
I did click that before.
03:54
so I already know that.
03:55
But again, this is just an AI check.
03:57
By the way, be sure to look through the, note history, the patient's history.
04:02
sorry.
04:02
So great.
04:03
So let's go ahead and add.
04:04
His glucose is 110 and his heart rate is 80, 8.
04:09
Great.
04:09
Now let's just say I'm happy with that.
04:11
And I want to see, submit this.
04:13
I'm going to save it.
04:14
So I'm clicking.
04:14
This is saving to Confluence.
04:16
It's actually real time writing this article.
04:18
This, this note, SOAP note into Confluence.
04:21
And I can show you that by clicking Going back up to, patient history.
04:25
Go back to James Chen.
04:27
And I had to click this refresh button and there it is.
04:30
So this is the, the.
04:31
And I can open it so this is the SOAP note I just created.
04:35
It's again this is live, it's going to Confluence, it's pulling it in.
04:38
So this is kind of a working live demo.
04:40
So that's really cool.
04:42
I want to be able to switch roles now.
04:44
So that is let me switch over to CSM now.
04:48
I'm logged in to CSM and now the CSM dashboard is now activated.
04:51
Let me click that.
04:52
So again me as a customer success manager I want to have insights into all of my customers that I'm responsible for and see what's going on.
05:01
How can I better support them.
05:02
So here's Mike and you can see this is the mastery kind of progress bar.
05:08
And this is really just a suggestion for how I can provide training to Dr.
05:14
Mike.
05:14
So based on Dr.
05:16
Mike's behaviors it suggested three things.
05:18
One structured glucose fields, the AI summary generation and keyboard shortcuts.
05:23
So it's recommending some training.
05:25
Theoretically I would click this and it would link to my learning management system.
05:28
I'm not doing that here but I could link to my LMS and just schedule a training session.
05:33
similarly there's another tab called Suggestion modules.
05:36
So now it's looking at all of the behaviors of customer behaviors for suggested training.
05:41
How can I support all my customers, all the doctors that I'm responsible for with some training and I can deploy modules here theoretically if we were connected to lms.
05:53
So let's switch over from customer support to to to to customer support.
05:57
And now you see the support dashboard is activated.
06:00
Now this is interesting me as a support agent, I want to know how can I improve my knowledge base.
06:06
So this one is looking for, this is the content gaps again based on some behaviors.
06:11
For example 47 clinicians were affected by structured glucose fields.
06:16
I should write an article about that.
06:19
So this is suggesting new content for me to create.
06:22
There's a feature discoverability tab here.
06:25
So based on how can I add in context help.
06:29
So basically SOAP notes, templates I should be able to.
06:32
Again there's 63 of them were missed.
06:34
So how can I improve my how can I improve the support for these features?
06:42
search optimization I think is most interesting.
06:45
So it says here 23 failed searches.
06:47
So 23 people searched for how to enter glucose and they didn't find Anything.
06:52
So it's either one of two things.
06:53
One, I need to write an article for how to enter glucose, or we do have an article on that, but I need to fix the metadata, fix the taxonomy, whatever it is, so that I can, so that when people do search it they can find it.
07:07
So that's pretty cool.
07:08
So let's switch roles again to our product team.
07:11
product dashboard is now activated and again me as a product engineer, what and I'm responsible for the continued enhancement of the SOAP utility.
07:20
and what.
07:21
I have a backlog of enhancements and bug fixes and bugs that I want to fix.
07:26
So how can I look at customer feedback to influence and prioritize my work?
07:31
And these are the observations.
07:32
So there's some UX improvements that it's recommending, there's some feature requests by the way, based on some of the behaviors.
07:38
We think you can add these new features.
07:41
And also it's actually suggesting smart features.
07:45
So how can we use AI to really improve the customer experience?
07:52
So that's all great.
07:52
Now let me add one last role here, which is admin.
07:56
Admin activates all of this.
07:57
I want to go to performance and roi.
07:59
And this, I spent a lot of time on this section here.
08:04
So the question is how can this tool, this, how can an onboarding, an AI enhanced onboarding program bring value?
08:14
So the first input here is number of customers per month.
08:17
So let's set a 10.
08:18
I can adjust it here if I wanted to, but let's just set it 10 customers per month that we new customers that are being onboarded.
08:24
That's great.
08:25
Now here is the onboarding maturity.
08:27
The level 0 means that we don't really have an onboarding campaign or onboarding intelligence.
08:32
So this is really the how much it costs without onboarding.
08:37
So we can assume that There's.
08:38
It costs $85 per support ticket.
08:41
It costs $150 per hour for CSM hour to support the customers.
08:45
average customer values about $50,000 a year and churn replacement cost.
08:49
So how much does it cost to replace a lost customer is one and a half times 50, thousand.
08:53
So $75,000 to replace a customer.
08:56
All right, so there's some assumptions here, but let's move over to basic level one.
09:02
So if we initially this is called this app, the beginning of the year one of our onboarding, our smart onboarding program, Level 1 introduces savings and it's just going to give you the number I'm able to save $63,000 per month.
09:17
So let me go through that, what that means and how we got to that, value metrics.
09:21
So how are we saving money?
09:22
So first is through case deflection.
09:24
So normally it takes about, we can anticipate 18 cases per month per customer, but I'm able to reduce that to 10.
09:31
So I'm saving eight tickets.
09:33
Now there's a value with that.
09:35
CSMs normally spend about eight hours per month per customer.
09:38
I'm able to drop that to five.
09:40
So I'm saving three, hours per month per customer.
09:43
A churn reduction.
09:44
I can project that.
09:45
I can save about 5% of lost customers.
09:49
So there's a small number associated with not losing customers.
09:53
and of course customers buy more.
09:55
So there's a 3% upsell lift.
09:58
So I can save or create one, thousand five hundred dollars in new value.
10:03
That's great, but it's not all about the money.
10:05
So what's the qualitative impact?
10:07
So moving the CSAT scores customer satisfaction, we can move the needle from 3.5 to 3.8.
10:12
can increase the discovery rate of how quickly our customers are, finding features.
10:19
there's teachable moments.
10:21
Obviously using a tool like this, we're going to greatly increase the teachable moments, the contextual in the flow moments to teach and time to value.
10:29
So most onboarding phases around 90 days.
10:33
Can we reduce that to 60 days at level one?
10:36
Yes.
10:36
So that's kind of way for us to quantify the dollar and the soft value of level one.
10:44
But more what's interesting is the maturity progression tab.
10:47
So now I'm at level one, what's it going to take for me to get to level two?
10:52
So first it gives us a little bit of a summary of what level two does.
10:56
but really level two is going to save me $148,000 per month.
10:59
And here's an interesting, we'll go down, scroll down a little bit more, a kind of a breakdown of all of our levels and how much money I can save.
11:06
So at level one, we're not saving anything but level zero, we're not saving anything.
11:11
Level one, we're saving $6,000 per customer.
11:13
We have 10 customers.
11:14
That's $63,000 per month for $765,000.
11:18
But it obviously it increases as we mature our program or we start aut start really deep in with, AI, we start getting tighter team integration.
11:28
so at level three we can go from $700,000 a year to $3.3 million.
11:33
$3.3 million per year.
11:35
And you can click that, you can kind of dig into the details.
11:38
So let's just go down to levels.
11:40
Let's go to all the way to level three.
11:41
What does that look like now?
11:43
I can.
11:43
Now let's go through the value metrics of level three.
11:46
Level three, we're saving instead of 18 tickets, down to 16 instead of eight hours per CSM.
11:51
is one, we're reduced that number to six and a half hours, reduced by six and a half hours.
11:55
So you can see that we're forecasting and all of this stuff is backed by some kind of data.
12:00
So here's a Bain, that supports that Forrester.
12:05
So these are not unrealistic, projections.
12:10
So the last page I want to show is about this app.
12:14
So I built this tool using a product called replib.
12:18
Repl Dot is an app that builds apps.
12:21
And it cost me almost $100 to build what you see here today.
12:25
and if we went through the traditional developer route, it would have cost probably over $10,000.
12:30
Not to mention it would have took much longer.
12:32
I spent probably two full days doing this.
12:36
It took a while for me to get to this point.
12:38
But for a hundred bucks in a couple of days I pulled this back this app off.
12:42
I wouldn't have done it even for $10,000.
12:45
It would have been, in probably weeks, if not months it would have taken to get to this phase.
12:50
A little bit about me and a little bit about the architect, the tech stack.
12:53
Here you can see all the tools and the technology we're using.
12:56
I've already mentioned we're connecting and integrating with Confluence.
12:59
We're also integrating with AI.
13:01
In this case, I'm using Grok.
13:03
Grok is the AI, engine bot that powers the X platform, formerly Twitter.
13:08
It's very good by the way.
13:09
So I encourage you to kick the tires here to, to use this app.
13:15
The link is on the page here.
13:16
if you want me to give you a one on one demo of this, I'll be happy to do that as well.