The Teaching Accelerator, combining Power Platform and Generative AI to boost education
Skills:
AI Product Management80%
Key Takeaways
Building the Teaching Accelerator using Power Platform and Generative AI with Azure OpenAI
Full Transcript
the global Power Platform AI hack was a great success last year thousands of people joined all over the world and a grand total of 115 Solutions were submitted Elliot Fraser is here to tell us all about his grand prize winning [Music] solution hi everyone welcome to another episode of the low code Revolution today we are going to talk about the glob Power Platform AI hack during the hack we had a bunch of webinars on Power Platform and AI but participants were also creating their own Solutions based on the Power Platform and AI Elliot Fraser won the heck and he is here to show the teacher accelerator solution Elliot can you please introduce yourself to our viewers hi Daniel thanks for having me on uh nice to meet everyone I'm Elliot Fraser um I'm a Power Platform engineer based in the UK uh I work for a company called clear bank and uh yeah as Daniel said I I won the global ai ai hack competition yeah and how did you like the uh the the whole hack yeah it was great um I had so much fun building the solution I learned um a lot throughout uh building it um I'm actually trying to turn this into a real product now uh and trying to help some teachers um with their day-to-day work so it's been really exciting and a great opportunity to work with uh this new technology yeah and during the heck was that your first time that you worked with Power Platform and AI or did you already have a little bit of experience yeah so um in my day job I work with Power Platform um been working with that for around three and a half years now um and then in the last year uh I've started to work with with AI specifically open Ai and chat GPT um so yeah um kind of new to the the AI side uh and we're starting to do a lot of this at my uh company as well uh that I work for um so it was great to put this into practice during this hackathon yeah and I believe you also can show us a little bit right about what you built yeah I've got a demo if you're ready to see it ah yeah I would love to see it okay so yeah here we are in the teaching accelerator uh the whole idea behind this was to um help teachers kind of accelerate their teaching process so particularly around uh kind of building lesson plans um for their their classes uh just a bit of background on this uh I uh got quite a few friends that are teachers and these guys work really really hard um often working kind of 10 12 hour days even on weekends as as well um and a lot of this is just uh due to um I'm not having enough time to to plan all these lessons and it takes a long time if you want to do your job really well um so that's what this app is trying to do is trying to accelerate some of that with the use of um AI which we now um can use via open AI or uh and kind of the chat gbt that I'm sure everyone's familiar with um so this is a uh a model driven app uh which the users can use to uh to do that so it's a little bit different right like if you normally go into a model driven app you have the like the the view of uh of all the items yeah that's right um and I think that's okay but I think um customiz and design a little bit makes it a lot easier to use or more pleasant to use um and I just think it looks better so this is what you call a custom page and um I've used these in a couple places throughout the app just to customize the user experience a little bit further um so yeah this is just a homepage really this is just helping you kind of navigate to different parts of the app but I think uh you could definitely extend this further and add a few more things into this homepage um very cool yeah I feel free to jump in with questions to My Demo so um I'm just hit create module for now so um the way this is structured is we have a module which might have many lessons uh underneath it uh as part of it sorry uh so let's first plan out what this module is going to be so um there's quite a few Fields here um we're not going to fill them all out now but I'm just going to give a model name so let's just say like computer computer science as um module um that we're going to create so I'm going to hit save and we got two ways we can interact with with the AI here we've got the AI Wizard or we got ask GPT uh this one we're going to use first is more of a um of a a quick way of generating content whereas the um the rgpt is um a way to do this through a chat based interface so first we're going to choose uh which which field to generate so uh I'm going to create Lear learning outcomes for now I'm going leave these custom instruction blank for a minute and then I'm going to hit generate and it's it should now generate me some content based on um only the module name at this point um so there it's done a pretty good job so let's just hit save and this will save this directly back to the app um so the way I've built this um is actually building up a context for the AI to uh to draw from when it goes and creates uh more content so if I now go to the AI wizard it uh and open this again it should have awareness of the module name and the learning outcomes now so let's go and create another one let's maybe do the syllabus um and hit generate and if I wanted to I could add some custom instructions there and it's just going to follow follow those so if you want if you say you want a syllabus that contains 10 lessons um you could do that okay SK is maybe a little bit more than we need here so it's just added the mod lames even though we didn't ask for it let's just remove that and it's done eight lesons for us which I'm not going to read them now but looks pretty good uh I'm going to save these um and you already can see this literally taking a couple seconds and I've generated kind of half of my module already um yeah this is really cool I can imagine that this would really help people to get started really quickly because normally when you look at a blank page or a blank screen that's like the yeah it it it it doesn't go as fast as just having some text already and then editing it of course so I I think this is a huge time saer yeah I think um I think um it will I think the name teaching accelerator is really good because it's going to accelerate the the journey for the teachers right it's not going to replace anyone at least not yet um and hopefully um but yeah it accelerates that that process so if if you kind of stuck for ideas you can just generate some content keep keep generating it until kind of you get an idea that you're happy with and then you can expand on that um with your with your ideas I guess cool so we kind of generated half the module here I'm going to move on to start generating some lessons so if we go to this lessons tab we can see we've got this empty Sub sub grid here obviously we got no lessons we can create these manually but I'm going to use the the GPT chat window here to do this so uh by the way these are custom pages that I've been launching in the side here um just like that homepage just a canvas app that we've embedded into into the model D app here okay so um we've only got one thing we can do from this uh from the module entity that we're working in so it's create lessons so I'm going to hit hit the create lesson button it should ask us some questions so it's going to say how many lessons would you like to create I'm going to say um eight always say please do an AI right [Laughter] do you have any specific ideas or topics um I'm going to say uh please uh please um so yeah as I said that record that we've been building up um actually gets kind of added into the ai's context here so it understands what we're talking about so I don't need to go and copy the module aim and the syllabus into this chat it just knows what those are okay so now was giving me a draft of of the lessons um of I could just keep chatting away with this with this AI here and and ask for something really specific if I wanted to um but for now I'm just going to say export and I use that keyword it just knows to actually go and uh run a function that we've built into the AI and it'll go and um actually create these records now and relate them to this module so if we just give it a minute and um yeah to come back to your your earlier story about friends that are teachers did you also show it to show this uh to him or to them yeah yeah I've got them using it as well um that's cool my first test is um yeah they really really like it and they've been giving me some really good uh good feedback and um helping me rate on it really nice yeah I also like the the fact that you kind of have a chat kind of way to to to get to this uh to to all the lessons here and then also the export is really cool that you can yeah create all the records because that's normally like a really timec consuming uh thing in model driven apps when you have to add a lot of records that always um yeah cost a lot of time so uh this is way better yeah I think um I think for anyone who's thinking about um using um open AI with um with model driven apps or canvas apps or anything where you want to start creating records um i' would say look into the uh like the functions they call it functions functionality um that's how we're that's how I'm leveraging that um in to actually um get a kind of repeatable um schema of of data that I can pull into automate and actually use that to create the records um and save them into Data ver yeah veryy nice cool so if I now close this window and go back to the lessons we should see those have all been created so we've got all eight rows um I assume these will match what we were just discussing with the with the AI so let's just go into maybe the introduction to computer science one and we can see uh we even got an objective that's been added in here um okay so we can go through the same process now just filling out the different fields here so um it's the same process really we've got a few more options here we can um create an introduction lecture script activity reading list and these are all available um via these buttons here oh this is nice so maybe I won't for the sake of time go into now but um yeah what I've done is predefined a load of different functions for this uh for this Ai and let the user choose it right at the start of the conversation um and then I give the AI a little bit more context as to what we actually want it to do so um it will have slightly different instructions depending on if we're creating a lesson activity or if we're creating a reading list for example yeah I really like the fact that this is context aware so when you are in a lesson you get different uh different options than when you are in a in a module for instance yeah and we actually bring the module context into the lesson as well so it's kind of looking up to its parent record and um it knows that um this is part of the computer science module uh and all the the module aims associated with that very cool cool so um yeah so the teacher would go and create all of those and I'll show you something I'm working on it's still not finished yet but um I want to have a way for users to create additional content as part of that lesson so this might be uh things like quizzes or slideshows yeah PowerPoint slides pictures maybe uh the first one I've worked on is creating um a lecture script and an audio uh file output um for uh for the lesson so maybe they want some audio over the top of a slide or they're creating a video for their class um we can now do that through the teaching accelerator so if I go to the content Tab and for now I need to do this manually I'm going to hit create lesson content and I'm going to call this um intro slide a Qui SC this so so it knows what we're talking about and I go into the AI wizard again um and all the only thing I can do is create a lesson lesson content script so I'm going to hit generate and we'll see what we get Okay so we've got a lecture script that someone could read out if they wanted to uh I'm just going to remove a lot of this for now so we're not Lo of audio and go to the text to speech tab now uh now this is using again open AIS uh text to speech uh API so uh we can choose standard or high quality you can choose uh one of several voices I quite like this one so I'm going to leave it there and then also change the speed if we want to I'm going to leave that there as well and hit save and generate now and yeah as I said this is going to come out to open ai's uh API and it's going to generate an audio file um for the user so if we hit play now hopefully you guys can hear uh good morning everyone yeah welcome to computer science model module today we're going to dive into the exciting world of computer science and explore the fundamental concepts that form the backbone of this field oh nice there we go yeah and I think um I've tried a few text to speech um options and this is far the best one I think it sounds kind of as close to um realistic as any I've ever heard so uh really impressive what they're doing um yeah wow so yeah another addition yeah so um as I say I'm trying to turn this into a product so uh I am going to continue building on this uh and add more features so uh I'm going to rely on my teaching friends to keep giving me feedback and if anyone watching has got any ideas yeah feel free to uh reach out to me yeah I'm sure people will reach out or leave comments in the under the video uh because this is a really really cool uh cool solution and uh it totally shows what we also aimed for with the uh with the hack because um these kinds of solutions are what we really want to see of course like the the the really combining of low code and AI is uh is such a powerful combination and as you can see these kinds of solutions are really helping the world so I think that's a that's a great uh great solution well thanks Elliot um I love seeing the solution it's a it's a beautiful solution and that's also why you won of course and uh I can't wait for yeah seeing uh when what you what you do next actually with the uh with bringing out the solution and putting it on appsource uh I would love to uh uh love to see it evolve at least and um yeah to our viewers um I hope you like this video and please like And subscribe and leave comments below if you uh have some questions for Elliott I'll make sure that he uh hear them and if you have feedback on the app also let it know I'll make sure that I pass it along and uh yeah hope to see you at the next [Music] episode
Original Description
In this episode of Low Code Revolution, Elliot Fraiser joins us to show off his grand prize winning AI-powered Power Apps application, the Teaching Accelerator. This solution won first prize in the Power Platform AI HackTogether for it's integration of a broad array of Power Platform components including Model-Driven Apps, Power Automate, Custom Pages and Generative AI via Azure OpenAI to accelerate the process of creating content for teachers. Tune in to hear about Elliot's Power Platform journey and see a demo of his app!
Chapters:
00:00 - Intros
02:30 - Teaching Accelerator Solution Demo
05:10 - Using the AI Wizard for Content Generation
08:15 - Create a lesson with GPT
14:20 - Using text to speech
Watch on YouTube ↗
(saves to browser)
Sign in to unlock AI tutor explanation · ⚡30
Playlist
Uploads from Microsoft Developer · Microsoft Developer · 0 of 60
← Previous
Next →
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Prepare for the DP-300 exam & the Azure Database Administrator Associate cert | Data Exposed
Microsoft Developer
What I Wish I Knew ... about landing a job in tech
Microsoft Developer
Igniting Developer Innovation with Vector Search
Microsoft Developer
Combining the power of vector search with Azure OpenAI then revolutionize image search with vectors!
Microsoft Developer
What I Wish I Knew ... about finding your place in tech
Microsoft Developer
Fluent UI React Insights: Accessible by default
Microsoft Developer
Signing Container Images with Notary Project
Microsoft Developer
What I Wish I Knew ... about finding your place in tech
Microsoft Developer
What programming languages does GitHub Copilot support?
Microsoft Developer
What I Wish I Knew ... about how much your job can change
Microsoft Developer
What I Wish I Knew ... about how much your job can change
Microsoft Developer
How do I become more confident about AI?
Microsoft Developer
How do I become more confident about AI?
Microsoft Developer
Performance Demos of SQL’s Intelligent Query Processing Feedback capabilities | Data Exposed
Microsoft Developer
What I Wish I Knew ... about coming to Microsoft
Microsoft Developer
What I Wish I Knew ... about coming to Microsoft
Microsoft Developer
Revolutionizing Image Search with Vectors
Microsoft Developer
Igniting developer innovation with Vector search and Azure OpenAI
Microsoft Developer
Getting Started with Azure AI Studio's Prompt Flow - Part 2
Microsoft Developer
What I Wish I Knew ... about finding my career path
Microsoft Developer
What I Wish I Knew ... about finding my career path
Microsoft Developer
Windows Terminal's journey to Open Source
Microsoft Developer
Can I trust the code that GitHub Copilot generates?
Microsoft Developer
What I Wish I Knew ... about interviewing
Microsoft Developer
What I Wish I Knew ... about interviewing
Microsoft Developer
What is the Microsoft TechSpark Program?
Microsoft Developer
SQL Server 2022: Accelerate query performance while reducing query compile time - w/ no code changes
Microsoft Developer
What I Wish I Knew ... about discovering computer science
Microsoft Developer
What I Wish I Knew ... about discovering computer science
Microsoft Developer
Call center transcription and analysis using Azure AI
Microsoft Developer
How to use Text Analytics for health in Azure AI Language
Microsoft Developer
Azure OpenAI-powered summarization in Azure AI Language
Microsoft Developer
Accelerate data labeling using Azure OpenAI and Azure AI Language
Microsoft Developer
Building a Private ChatGPT with Azure OpenAI
Microsoft Developer
What I Wish I Knew ... about how to interview
Microsoft Developer
What I Wish I Knew ... about how to interview
Microsoft Developer
Getting Started with Azure AI Studio's Prompt Flow - Part 3
Microsoft Developer
Intelligent Apps with Azure Kubernetes Service (AKS)
Microsoft Developer
Getting Started with Azure Blob Storage | Data Exposed: MVP Edition
Microsoft Developer
Chat + Your Data + Plugins
Microsoft Developer
What I Wish I Knew ... about different career paths
Microsoft Developer
What I Wish I Knew ... about different career paths
Microsoft Developer
Advanced Dev Tunnels Features | OD122
Microsoft Developer
Learn Live - Manage performance and availability in Azure Cosmos DB for PostgreSQL
Microsoft Developer
Plan your SQL Migration to Azure with confidence | Data Exposed
Microsoft Developer
What I Wish I Knew ... about social skills in a tech career
Microsoft Developer
What I Wish I Knew ... about social skills in a tech career
Microsoft Developer
All About Vectors, Search, and Function Calling in Azure OpenAI - Labor Day Special
Microsoft Developer
Introduction to project ORAS
Microsoft Developer
What I Wish I Knew ... about finding the right major
Microsoft Developer
What I Wish I Knew ... about finding the right major
Microsoft Developer
What I Wish I Knew ... about how to approach programming
Microsoft Developer
What I Wish I Knew ... about how to approach programming
Microsoft Developer
Learn Live - Scale from a single node to multiple nodes with Azure Cosmos DB for PostgreSQL
Microsoft Developer
What I Wish I Knew ... about diversity in tech #1
Microsoft Developer
What I Wish I Knew ... about diversity in tech #1
Microsoft Developer
Get started with SQL Server AGs across Windows, Linux and Container Replicas | Data Exposed
Microsoft Developer
Writing LLM Apps with Azure AI and PromptFlow
Microsoft Developer
What I Wish I Knew ... about how cool working in tech could be
Microsoft Developer
Open Source foundation models in Azure Machine Learning & optimization techniques behind the scenes
Microsoft Developer
More on: AI Product Management
View skill →Related Reads
📰
📰
📰
📰
EdTech Blogs vs. Vlogs: Which Helps People Learn Better?
Medium · Deep Learning
The Mission Behind Anitha Rises: Empowering Students, Women, and Lifelong Learners
Dev.to · Anitha
Why Marks Alone Are a Terrible Measure of Progress
Dev.to · jahnavi sharma
2–2–1: On Student Accountability and Punctuality.
Medium · Deep Learning
Chapters (5)
Intros
2:30
Teaching Accelerator Solution Demo
5:10
Using the AI Wizard for Content Generation
8:15
Create a lesson with GPT
14:20
Using text to speech
🎓
Tutor Explanation
DeepCamp AI