 Thanks, GavTech, to the value of food, speaker, and slides and manager. Thank you. Alright, so my name is Leonard. I'm a product manager here in GavTech. But this presentation is more about this. So let's talk more about FormSG and what it is. Alright, so FormSG, what are we? We're basically like a Google Forms for government. And our tagline is the ability to give agencies the power to build government forms in minutes. So when it first came out of this, it was actually a bit ironic for some of the users because when you think of government in minutes, it sounds odd. Normally you don't think build government forms in minutes. You might think of like more of a negative months maybe. But we managed to do it. And let me first dive into what was the problem that we're trying to solve in this blue screen. Paper forms. Today in government, we are still using a lot of paper forms. Not sure if any of you have filled in a paper form from government in the past year. Maybe some of you might have. Okay, so that should be on FormSG if it's not already. So when we ask agencies how many paper forms are there? What is the size of the problem that we're looking at? We've got all kinds of answers. But we decided to search for an answer on our own. So we went to Google, our friend, and we tried to estimate the number of paper forms on government websites. So we just did a search for form within government sites and of file type PDF. Because PDF forms aren't actually digital forms. They're just paper forms converted to PDF. And we saw 60,000. Pretty large number. Some of you guys might argue that this number is obviously inflated. Because in RP 60,000, how many of these are actually forms? Maybe the query form is not that accurate. But it's a rough order of magnitude, tens of thousands. Even if you divide this by tens, it's still thousands of paper forms on government sites. It's still a big problem. And if you were to imagine if all of these were to be digitized without FormSG, there would be a nightmare. Because imagine if this form, let's say it's an IRAS form. Without a self-service form-builder tool, the agency would have to contact an external developer to digitize it. It would cause the engine to digitize forms. That's pretty expensive and really slow. Which is why we built a form-builder tool. It knows how to digitize all of these forms. There's thousands of forms. And that's FormSG. Alright, so how does FormSG actually work? It works in a slightly different way compared to Google. All the form-builder tools out there. Four steps. So the first step, you know in government we have both internet and intranet. Internet is what everyone in the world uses. Internet owns a secure intranet. An internal internet. But for FormSG, you can build your form on the internet at form.gov.sg. We took an extra effort to make this mobile optimized. Because in government, most agencies do not have an internet-enabled laptop. In government, the internet section, most users can only access the intranet. But we're not winning the whole thing on intranet because intranet is quite expensive. So if you're hosting on internet, we better make it mobile optimized. Because they'll be building their forms on their form. So the first step, you build your form on the internet. The second step, you share the form with a link. And the citizen can fill in this form. We're very integrated with government authentication. Some of you might have filled in a sync pass before. My info card pass, you might have heard of all these things. We actually integrate with them. Which means that it's easy to enable such government authentication with a click of a button on your form. And the third part is an interesting part. So we don't store form responses. That's how we're different from all the other forms with the tools out there. So for every response, you get an email. Which means that our server doesn't store any data at all. So a complete lightweight. It's basically just like a collect and forward kind of model. You can build a beautiful form using our tool. But we won't host that data for you. We'll just email the responses to you. But of course, you get this nightmare where if you have a survey of like 30,000 people, you get like 30,000 emails. No one wants to receive that number of emails. Which is why we have Step 4. We developed this compilation tool. A hack for you guys to aggregate the emails back into an Excel. I'll talk more about this later. So we've been around for only like a bit above a year. But as of today, we've had quite a bit of damage. The 90 government agencies in Singapore, 84 of them are aware that at least one officer signed up. Many of them are active, having created forms with thousands of submissions. But I'll say our penetration is about 20%. Everything are really active users. Not quite there yet. But the speed is picking up pace. So we're seeing this nice looking graph finally in terms of user sign-ups. These users are government officers signing up. You can see in the last month's loan, there are almost 900 officers that signed up in one month. So given this speed of growth, actually in another year, we're hitting 100% penetration. The whole government. But of course, to the end, we expect a slowdown in the take-up because of the laggards. But people are not just signing up. People are actually using FormHG to build forms because you can see the usage graph as well. In green, it's picking up one month as well. You can see at the beginning, we struggled to find this product market kind of fit. I mean, just looking at sign-ups alone, you can tell there's a market fit that there's a problem worth solving. But if people download and will not download, it will sign up and do not use that it might not be the right product to solve it. So at the start, we're kind of struggling from that. You can see usage, it does increase, but it's kind of flatlining. But in the past few months, we seem to have a bit of traction that's finally happening as usage. And there are about a few, like a thousand other forms of various types that are being used across government. Some of these are public-facing. You can find these forms on agency websites, such as, you know, application, the one in Brown, MOE, you know, this old-level music cost application that's on FormHG. And even for research polls as well, such as LTA Transport Master Plan E-Poll, these are done through FormHG. All right, so at the end of all our forms as well, we have these toilet smiley questions, you know, the five smiley faces, turns out a lot. Around 20% of people actually fill in this additional question. And 86% of the respondents are actually satisfied with filling in FormHG forms. All right, so all that, it's about where we are today. It's been like, I guess, okay, 15 months since the very, very, very beginning. And that's where we are. Let's rewind back and talk a bit about the journey and some of the, I guess, product issues and product thinking that went into designing this. All right, so this is where we are. But actually, it really didn't start out that way. So it started more like this. So we actually didn't start out with a form-builder tool. It started off with me manually building HTML forms for our users. And then the citizen will fill in your ugly form. I'm not a UX designer. I try to be knowledgeable about UX design. But I'm obviously not an aesthetic person. The responses still go to your outlook. That was there from the beginning. But we also had to spend 10 hours setting up this macro. The earliest version was some macro written in DBA to aggregate responses from email into Excel. Even worse is that it was basically this output of code that had to pass around to agency offices. And you know, agency offices, many of them are not the most tech savvy. So I have to explain to them what is code. And this is DBA. And you've got to change this line of code to put in your email and all those set up. You spend a long time sitting with them to set it up. But the whole point is that at the beginning with something so ridiculous, we still had users. Some people that are willing to sit down with us to solve their pain point. Because without this, what they were doing is every single week, they have a stack of paper forms in front of them and the poor dude had to manually digitize that into an Excel spreadsheet. So they were willing to sit down with us to talk about the solution with us at the beginning and spend that time because that would save them hours every single week from their workload. So at the beginning, we had our first two use cases of this. They were for very obscure things. They were not for fancy MOM, M-O-E applications, large volume, no. For things that many of you might not have think of pigeon inspections, with tracking pigeon inspections, the poor officer was doing it to paper. That's horrible. So we decided to digitize them. For house visits, the previous process for the officer to take down notes on paper, go back and manually fill it in. There was an observation in HGP who were manually collated into Excel. A lot of manual work being done here. And why can't you use Google Forms? Actually, the first thing I told them was, hey, just use Google Forms for this. It's a lot of work to build a whole form on your own. And their reply at that time, that was 17 months ago, was classified census data. I can't use Google Forms for this. And their proposal was really, hey, why don't you build a Google Forms for us? Some secure government data centers and how much does it cost? We'll pay you. They wanted to pay us, but being mindful, civil servants, I guess I would call it, we thought that there could be a better solution than just implementing the whole thing ground up, hosting our expensive government data centers. So what I've really learned in red at the bottom, it's really to... it's not always a good idea to listen to your users for their solution. You have to talk to users, obviously. We talk to our users to understand their pain points and their problems. But sometimes when you ask a user for, let's say, they're riding a horse, you can ask them, hey, so your problem is that you're not traveling fast enough. What is your proposal solution? They'll say, okay, I think we need a faster horse. Maybe feed the horse better food, whatever. But they would probably have thought of a car. So it's it's sometimes up to someone external to the problem to bring some a better solution to that problem. And that's what I learned so far. So the version zero was the email method that we came up with. And the main reason is to avoid hosting on expensive government servers. We are hosting on AWS. That's more low cost and it's all the good stuff that comes along with it. It's pretty good reliability, availability and etc. And so for that, the very first version was really just each response sends one email to you. But that doesn't really solve the whole problem because the problem is still about manually filling it in. So how do we go from emails to the Excel? That was the second challenge that we had to face, which is kind of like version 0.5. Which you thought about that VBS script that aggregates emails from look into Excel. So that works. But of course with VBA, it's a whole nightmare. I wasn't well versed in VBA. So I mean, I'd pick it up as well. And I guess this is where I learned that sometimes it's okay to even pick up the old stuff. If you find it useful to code out a quick solution, it might not be the best solution at the beginning. Of course deploying the VBS script was a nightmare. At least it helped solve that one user's problem. So maybe we can start from that and clean up later. That's what a user wanted. I unskided the user for two hours going back and forth on how to install the script. There's a lot of interesting problems here. So the script I could distribute the script through an email. It's not like I could package it inside some executable or something like that and push it down centrally. The government system worked that way. So I had to email the script over to them and the user was so happy that they actually saw the response. This is a dummy response by the way in the Excel spreadsheet. This is an actual email. As you can see the title is called VBSession from Hell and the user said, this is what Trion looks like. This is pure joy from solving a user even though it's very inefficient to solve at the beginning with this VBA script. So from there the user actually was like, hey, this works. What if they want to change fields on the fly? Just call me no problem. I can add in those extra fields for you on the HTML form. Just not VBA. But then I realized this was obviously stupid because more and more requests started coming in. I want this field to add an extra s word smithing. I want the extra field. How about a new form? I decided to ask them to do the work to build a form builder tool. So we borrowed two people. There was Abby and Nam. Sorry, I couldn't find a picture for Nam, my colleague. So we borrowed this guy. So it's a basic backend engineer and front-end engineer to start building a form builder tool. And what did we do? We looked at open source tools out there. Of course you wouldn't want to reinvent the wheel. Probably not trivial to build your own form builder tool. So we found a few. We found JQuery as some of you guys know. Maybe a bit dated. So we didn't like this because we weren't designers but we felt it was ugly as you can see. Form HG looks better, right? And there was also no user login implemented. Which is why we watered in the early days. Because obviously you create your forms. You want to maintain them yourselves. You need some kind of login to prove that those forms belong to you. So you don't end up messing other people's forms. So my colleagues wanted me to use something else called line survey. Actually some of the words are off at the bottom. But this was another alternative we didn't use it as well. It's written in PHP and no hate for PHP. I think parts of Facebook probably still use it. But I personally worked with it. Don't really like it. It was also kind of ugly. There was a free version that was open source. But then there was also this paid version. I think I wasn't a big fan of that because I would imagine that if our company offering a premium version of something and a free version if you put a business hat you'll probably say okay. Maybe I'll put some traps in the free version such that you will use the free version at a certain point and realize that your desires are not satisfied and you don't pay. So I didn't want to fall into such a trap. So I didn't want this either. But the main issue that we had was that there were actually millions of lines of code in this repository. There was a lot to navigate through and it was quite hard to find the piece of code that wanted to change in order to send emails over. There's all these form builder tools. They all store data. So I don't remember what... I mean it's PHP but I don't remember what was the actual backend database language that they used. Probably some kind of SQL I'm not sure. But we basically wanted to change the part of the code to instead of storing data and using SQL, no SQL, whatever, we wanted to change that so that responses would be emailed over and that part of the code was so hard to find. We just couldn't find it. You had to probably change maybe 20% of the code in order to add in the functionality and that was millions of lines of code 20%, hundreds of thousands of lines of code to code. No way. So we didn't stick this. Finally we found something. It's called teleforms. If you guys are aware, another popular form is called typeforms. I call it the beautiful Google forms. But teleforms is an open source version of typeforms. So we found that and we loved it because it looks beautiful. The user account is done first. It was in JavaScript. So no PHP, yes. Minstack, Mongo Express, Angular, Node.js. Angular 1 fine but still at least Angular and not vanilla for example. Although maybe some people out there do that. And the most beautiful thing is a point from a form because it basically took us a minute to find where the code we had to change to email responses over. Which means we're going to take this code and we just modify that part and we're good to go. So that was why we chose Telephone and at the same time I knew that two engineers that we borrowed, we only borrowed them for a limited time. It was a very hackish setup where we borrowed them for another team for two weeks just to set this up. And after we set this up we can move on and do something else. It wasn't really like a real project at the beginning. So I knew that time was limited but this project had a lot more potential. So we needed some permanent help. Which is why we got our good friend here, Asher. We said it's seated over there as well to join the team. So at that point in time, Asher just joined and I asked him to join FormSG. Because what I learned is that at the very beginning, when you have no users yet, the litmus test is to convince capable people to join you. Because you can't convince someone to follow along with your vision it's probably even harder to get users and their bosses to come to that vision. So thankfully we have him. Now that we have a team, we can do all the team stuff. All the code stuff so we go to the beginning, think of a name. Let's start with thought of like FormR. It's basically like the Spanish translation of forms I think. At least from Google translate. But then I wanted to buy FormR.SG but realized that it was taken. I was so sad. But then I removed the A and realized that Form.SG wasn't taken. So obviously call it Form.SG. Can't believe FormR was taken but not form. So yeah, we renamed because of that and we created a logo. Form.SG that was the earliest logo. It was not done by a professional designer. It was done by one of us. If you realize it's actually like a form but tilted 45 degrees. I don't know if you all realize that. Form.SG. We chose what I call developer colors. Because all of us are developer. That shade of blue and that shade of green. It's not the air B&B kind of colors. Sorry, the beginning team did not really have designers in it. So we set up Pivotal. It's one of the task management tools. There are many Trello, Asana, etc. We just use Pivotal because we're comfortable with it. There isn't a lot of processes at the beginning. We deliberately kept it lightweight because we're a small team. It's just a GitHub repo. You send a code review before PRS merged into master. Simple as that. Some daily stand-ups at the beginning because we wanted to move fast. We still wanted to make the team in sync. We had daily stand-ups, Slack channel for communications, and also a weekly sync up with our boss. That was all. At the beginning, it was all about the product. It was all about solving the user's problem at the beginning. It wasn't about spending two months thinking of a team name, what's the best team name, what's the best domain we can buy, all those things at the beginning. I learned that at the start we probably shouldn't spend too much time even developing full-blown agile or scrum, especially if it's a small team. Because throughout the whole process of forming into the 15 months, we really carefully think about how much time we're spending on processes and how much time we're spending on the actual work. If we end up spending more time on processes than the actual work, then this is drop those processes. For our small team, we found this, this would be sufficient in terms of all the processes they would have to do. So we didn't really do a lot of those things that you might do in scrum. Like all those retrospectives, sprint planning, story-pointing, all those things we would go that deep into all these things. Okay, so we're excited. We've got a team. We solved the user's problem. We're going to get our forms deployed. We want it to be actually used by our team. Because although the worker, all the way at the bottom in pink, that we talked to liked it, actually there are all these layers in red, purple and all that that had to approve it first before you can finally see it live. How do we solve this? We dog-fooded. We said, okay, this external agency agencies, whatever. Understandably, it takes them time to get the series of approvals, especially for a prototype project. So we did it internal. Maybe it could work for our team first. We did some productivity gains to our team. So we had a colleague who was basically thinking of a checklist form before open-sourcing. And we had another one, another colleague, Lee Wei. He was thinking of using forms for judging a data challenge event. So we asked him for help and great things came out of them. On the left, we discovered some bugs. We also realized a very critical feature that we missed out. The ability to allow new lines on your form. Can't believe we forgot about that. And on the other hand, for the judging event, our then CEO at the time, Jacqueline, was a judge. She was a judge. And the past year, she went to the same event and she found a paper form. This time it was a digital form. She was elated about it. And she asked, where does this form come from? And we told her, this was created by one of your departments. Which is why she became our first senior leader ambassador to actually help promote form SG to other senior leaders. That's how we got a bit more traction, a bit of the buy-in from the top. And that was something that we obviously didn't expect at the beginning. This whole dog-fooding thing was something that we didn't think was such fruits of labour, I guess. But unexpectedly, Lena some users from the senior leadership level and then we made a video because we think that form SG if I have to go down and present all these senior leaders, it wouldn't be as impactful as if I just made a video to carefully explain how form SG is. Because form SG in the early days is quite complicated. A video works. So I just got together learn some iMovie and all that. I haven't made a video in my life before this. So I'll play the video. It's a bit ancient. It was done like 14 months ago. And it's 4 minutes. Let's see if I can play it. Video! They set music at the start because the current process came through. The music switches. Officer had to do manual data entry for all the responses. This was the early UI. This was sent to senior leaders by the way. This was the VBA script. So they had to go into the code and change a few things. And then press the tiny run button over there. And then you get the responses in Excel. So that was the early version of form SG. There was the marketing video that we made. It wasn't the most professional but it did create a lot of buy-in. So senior leaders understood what the product was about through this video. And that is the result. We got a 10 times number of sign-ups. September 4th, October 53. We're excited seeing this. By that time we were euphoric over this. And we started getting invited to present at various agencies. And now we have a first world problem. We have a bit of attention but we have all kinds of feature requests. We have people wanting to build all kinds of form fields. And how do we tackle that? For example, we just look at two dimensions really. Intensity of problem and frequency of appearance across agencies. Just like that. So the ones in red are pretty insane problems. So those are blockers for those agencies. So without those features, they will not deploy something on form SG. So you can imagine like four agencies, A, B, C, D, right? And just by eyeballing it, we'll probably work on something like F1 because it's like red and all the agencies A, B, C wants it. We'll work on F2 maybe because it's a blocker for agency C and really want them to use form SG. And F3 maybe, F3 is maybe not a huge blocker but it's a common enough problem that occurs so we might want to tackle that. But the point is that we have to walk away for something like agency D because we'll build a lot of engineering resource to code out for features F9 to 12 years for agency there are blockers. You have to convince them to use it. Why not focus a bit more on ABC especially at the beginning which is the main thing that I learned at this stage was that we had to learn to actively walk away from difficult customers at the beginning. I mean, often times you think about customer centricity it's all about satisfying the customer but especially when you're lean and it's early not every customer might be a perfect customer to begin with. There were a lot of big issues. Chief among them is this VBA script the agency that have to sit down with those agency officers for 2 hours to 10 hours to figure out what's going on. So I rewrote it from an Outlook macro to an Excel macro big improvement there still in VBA but now there's UI now it's a script with a run button it's basically an app in Excel so what users can fill in now they don't have to do code anymore they just fill in different fields here such as what is their form title and all those things and they press one script and that will transform their emails into an Excel spreadsheet with those email responses collated So this saved me a lot of time because I don't have to spend 2 to 10 hours each time 10 then I'll set this up but I still have to accompany this with around a 10 page Microsoft Word document explaining to them how to install this but it's still not that straightforward to install you can generate like a self signing service and all these kinds of issues I have to explain what script means users are saying could I check out what script is is it a document file, is it code, what is it and agencies have to ask the IT department to whitelist our script for running I guess because of security issues it's a macro after all there's no email around so it's challenging for us to centrally push updates can you imagine if you're like installing uber app and then you have to email uber hey can you send me the uber app by email can you be sent our way just in case they got blocked and dropped I wanted to check if you send me the uber app imagine that but we had to do it like this there was no other simple way to deploy our script it had to be to email which was painful we were thinking of version control and all these things so we thought of a solution joined one of these innovation challenges bought a server on the secure government data centers just one server and we coded a tool so how this tool works pretty interesting so the first step out of the vba world the first step is to export your emails to a .psd file so on outlook you might archive emails on outlook actually I'm not sure how many people here use outlook maybe a lot use gmail I use gmail more than outlook but if you're out of outlook you can archive emails and the file extension of the archive file is something like maybe archive.psd so everyone's familiar with archiving emails at least in the public service so they receive all these emails they can receive the formage gmail responses to .psd second step is to visit this link on the government intranet and then you can dump the pst file there and we have this parser that parses the .psd file extension to a .xls an excel file it's basically a parser that we actually we didn't write it ourselves we found it open source I have no idea why someone wanted to pass .psd to .xls but I can find all the interesting things open source online so we found it how to pass something from .psd.xls and we implemented it we made some tweaks and this became our new way to collate email responses into excel and people started using this even more because what we realized before this was that the script was so hard to set up that people just didn't collate the emails in fact they bought email inboxes of larger and larger size to store all the email responses and worked off the emails they didn't rely on excel but after this we realized that we were actually aggregating and using this feature alright so and then through the months after using that framework that we talked about intensity and frequency we implemented many features such as collaborative forms auto acknowledgement email when you finish your form you can receive a confirmation email that can be customized by the agency various types of fields attachment fields where you can upload a picture to send over as an email attachment check box video about other fields even some kind of logic on your form where you can edit one field which decides whether to show another field or not and even with all these features we still have pretty stagnant growth in terms of both sign ups and usage as you can see from this graph especially the sign up graph I think for the first 6 months we jumped from September to October and I was really happy about that 10 times growth but obviously for the next few months there was this small growth people were signing up right it's just not signing up at a faster pace so what happened we needed something better we had to join our team as well we wanted to build a better user experience and we talked to users even more when I mean talk to users we talk to end users we talk to citizens we wanted to build beautiful forms it wasn't just government forms they can collect classified data but beautiful forms how do you do that actually at the back on this floor we have an eye tracker kind of room we basically invite the user down in a form and there are these cameras that actually track where the user's eyes are looking at so we invited all kinds of users we invited the young and tech savvy we even invited someone who is in the 60s as well who has just taken one computer class and even she can fill in a form as you form actually and also at the end of all our forms we added in that feedback to try to gather different forms and we even got some features suggestions from there as well and we found some interesting things actually I want to share a bit about some of the things that we found that were actually not so necessary so there were these four things there were many others but I'll just share these four progress tracker on some forms you see 0 or 4 answered sometimes it might be a percentage 20% of the form done we realized that people were actually not staring at that part of the screen when we use eye trackers we can see that they'll put a heat map of where people are staring at and realize that people are not staring at that so how are they measuring the progress to a form found something interesting people were staring at this so it's the height of the scroll bar and where it is relative to the screen so all the way the bottom people are almost done and if the scroll bar is really thin it means it's really long form so people are paying a lot of attention to that which means that actually we can remove the progress tracker and then there were these multiple pages forms sometimes we had a form with like 100 questions and then agencies would come to us and say hey we have 100 questions we need to separate our form into like 5 pages we have to do that that's like the best web design standards so we actually thought about this experimented if you talk to users we actually realized that end users aren't the most satisfied at multiple pages because number one I mean based on the previous feature they decided to remove with multiple pages you can't really track progress through the scroll bar because you can't see the scroll bar of future pages and secondly people are just most of our users playing the forms on their phone and on the phone it's more optimized to scroll through things imagine if you are using Facebook and you got to click on some next button in order to view the story of the next person the feet of the next person it's pretty annoying to use the scrolling on the mobile phones which is why we said no multiple pages mobile first we'll put all your questions on one page and you can hit us to separate them and users will scroll then there's this other back to talk button right my thing is a good idea some websites have it what we realized is that back to talk button actually causes users to skip questions because they think they can go back easily you can imagine someone feeling the form like this you feel in question 3 and then go which questions are easy to fill in I can always go back to talk there's that innate desire to not fill in the heart to fill in questions first because it's an easy way to go back so actually by removing this this actually saves user frustration at the end because they start filling in questions sequentially and we even add an error message that appears along the way to prompt them to fill in the questions as they go not leave it to the end this actually makes the user's life a lot better imagine filling in the whole form and because it's back to talk button because they don't warn you they haven't filled in questions all the way to the end it tells you 6 questions not filled in that would be quite frustrating for quite a few users and save as draft as well so everything save as draft is a great idea but we realise that this makes users dive into your form when you have a 1 minute toilet break horrible so the simple fix to this to prevent users from diving into a form without having the prerequisite materials available and enough time to fill in the form is really mandating this mandating agency officers to put this on their form estimated time to complete 30 minutes but this along with also section on instructions prerequisites they need fill in your form make users set aside that amount of time so that they can fill it in in one sitting that's a much better solution than coding a save draft functionality and along the way we also remove the passwords earlier I was talking that we had this user login system we realised that actually managing our own user login system is not that straightforward logins are everywhere you might think they are quite non-trivial to implement but not necessary you know with logins there are a lot of things to think about there's like forget password, there's change password when you first sign up there's this verifying link they generate there are a lot of things to consider it's quite a lot of code complexity so we removed it what we did was email OTPs because forms are one of these things where we hope a user doesn't login every day actually it's probably you login once and maybe three months later you create another form then you login again so people are not going to remember their passwords which is why we have temporary passwords they are sent into their government email accounts and this means goodbye lines of complicated code and has a security side benefit as well because now our form builder is linked to the secure government email server to login you're going to get the email OTP from your secure government email also we designed a whole user interface in the video you saw that it was there now it's this for home page new logo by the way as well looks a bit like okay fine clipboard maybe a couple other companies but it's more design type not those developer colors so build your form tab very developer kind of style you can see even the buttons over here and all these icons they're from an icon site that developers often use I can't remember what it was called I wiped it from my memory my designers are happy I did that so sorry yes babycon, gliffycon, gliffycon yes gliffycon and fevycon those are the developer ones the designer ones are I think box icons pretty good icons over there like the plus over there see you see the L, E and X plus the plus is from box icons and the I also the preview the I you can see it's more sleek you know okay the form used to look like this looks a lot better now and it also created a beautiful landing page today you can go on form.gap.su on your phones and you can see this landing page and this was a very powerful feature that led to the recent traction because instead of passing around that video okay fine the truth is I'm lazy to make another video but instead of passing around that video or maybe a deck of slides to government agencies it's better to pass than the product itself so a landing page that describes what the product is about especially if you're accessing on your phone form.gap.su we put all the information there what features do we have how do you get started and of course at the top as well the numbers are one part because you know there are many people they work with that are quite wrist of us so if no one's using it they probably wouldn't dive in here first so these numbers are proof that other people are already using it which gets them started and that's really how we are getting the recent traction because of this beautiful landing page and this new refresh user experience what's this oh yeah I remember the point here is yes that at the beginning we were not really focused on selling that hard because we don't really want to sell a product that doesn't work you probably will spend a lot of time banging on doors and people will sign up but people will probably not use your forms because it's a product that doesn't solve their problem so at the beginning we spent a lot of time building those features prioritizing features making sure it works before even getting a designer on the team to think about a better user experience to think about a beautiful landing page we're not concerned with that at the beginning at the beginning it was all about product features does it solve your manual data entry problem does it solve your manual data entry problem yes okay we'll code that feature and then when the time is right which is a few months ago when we knew that we roughly had a good product that can solve quite a lot of form related problems digitized paper forms then we started pushing the pedal and here's what we add and really to be what I call future proof we are moving away from a form builder tool as well as of today we like to call ourselves an eService Builder so we're more than just building a HTML form you can also enable SyncPass Mindful and CogPass just like that from your form do you think of forms as basically like apps at least for government because a lot of the apps that we have are just taking in information so they're basically forms so if we can incorporate more of those app-like features in a form builder we can move towards a simple app kind of builder such as integrating with digits ePayments and all these things we're not there yet but we could consider such things and we're also thinking of a way to move away from emails so one of the possible ways is end to end encryption but the server never sees the private key the private key could be emailed to the person that creates the form and of course once we start storing data it opens up a whole world of opportunities where it's more convenient the script was bad, the data collection tool was a bit better but it's still not as convenient if you see the responses there on the same side and once we start storing these encrypted responses provided the private key from the user it can link up to analytics and visualization results and also to integrate directly with back-end agency workflow systems, all those case management systems client, CRM, CMS, whatever you call it those they can supply the private key and they could fetch data from a form builder tool as well and we are going to open source back again after doing a bit more cleanup and the goal of that of course is to contribute to open source community but also for open agencies to perhaps get their own developers and all that to quote niche features if they want to that's how we could augment our engineering team because you know there could be a certain agency such as let's say NEA for example they might need like a dual fencing kind of feel or something like that so it's a very niche feature that only caters to them so as a platform servicing so many different agencies at this point to provide different agencies we might not necessarily want to focus on that feature but they will want that feature it might be a blocker for them what they can do is they could you know our code is open source they could hire maybe a developer to quote that's the feature and send a pull request back and then all of a sudden this feature can be available for everyone as well although it serves their niche need so I think I've come to the end of the presentation what do you guys think? it's a tricky question in front of my boss at the back just kidding so for everyone's sanity I just joined my job yesterday so I wanted to go home but I think normally it goes to 30% to 9% so I hope if Gotet has no issues maybe folks could just ask let it any questions and go around to maybe 825-9 yeah any questions? always ready to question you thank you so my name is once I'm from Germany so my question is I forgot my question for people it always happens to me first question is how did you even end up looking for was this like on a road map somewhere was this something that your boss has said they got a problem or did you just have a request one day on email and decided to solve it and then you kind of walked this whole thing second question is I don't know how many months 15-16 months or so with the benefit of retrospect if you could reverse time what would you have done differently if anything and by the way my wife works for PA I asked her yesterday she said yes it's a true market test yeah so we don't know how to start as well so basically our team has had a library of ideas that we've been trying to come up with like we have a drive ball and a poster of ideas you can have and this is one of them somebody did a form development what if you build a form build security hard so it's like don't have a database that was pretty much like the one paragraph when I joined the team there was a couple of projects forms was not on we were working on data development like parking like this this was a CCA no one told us to do it this seems pretty deep we have some cycles and I told him and he got some users and once you've got enough users it seems like a real project we get some funding for it we have to do the thing first and then get the funding out it really was just between other projects that's it I think that's basically it so let me answer the second question what would I change probably start thinking about this end to end encryption thing a bit earlier the whole email data collation tool thing wasn't the most user friendly for users I think if you probably ask people in PA they'll probably say oh it's really easy to create a form but people always say listen to the thing after the but pay special attention to that but it was so hard to collate the data back emails was the great hack at the beginning that was off the ground because we didn't have to think about hosting on our data centers but it would probably help us if you think about the end to end encryption solution a bit earlier what's up you have to fill out a constant form or in Singapore like the entry exit form and there's still paper can you use your tool for that? yeah of course if you can tell us the form if you have the form link or something or in the name of the form we could approach anything in Singapore I think people normally fill in the form of the plane with just no wifi connection maybe connection of the plane maybe I'll try yeah I agree I think some countries even made that online not on the plane I think some countries have been sorry yeah yeah yeah yeah yeah yeah yeah yeah yeah yeah yeah I'm concerned that this is something that actually counts for a lot the email is very powerful we can circulate it people think that the government I can tell you that as a government worker our internal opinions we spend months on citizens email like civil servants or email ministers last few things things happen really quick if there's stuff you want us to do to try emails have been good it really counts so my email is landed at data.caftersg any other questions yeah yeah so technically we are still a pilot program technically so KPIs are mainly defined we actually define the KPIs and we discuss it with our bosses then they give us the funding KPIs revolve around usage we get KPIs on forms created KPIs on users sign up and stuff like that so we've hit those KPIs in terms of how much we actually find that on the email so if we can actually share it so basically the way to think about it is that one of the way to think is that when we present these products a lot of the work goes for government systems are actually quite large compared to what governments spend on verification systems the amount we spend running forms I would say less than 0.1% more fraction and the reason this is the case is because if you think about the system there is no answer because you wouldn't destroy any data and you just send it through a email you usually have a server which sends a form which can be visited cash technically and it's faster if you know your government email address so our total infrastructure cost for running this is virtually no the main funding is pretty much like approving or manpower to work on this and that's it I believe proportion of amount we spend on infrastructure compared to salary working three people involved in this people for this project it's probably 90 to 10 manpower 90 infrastructure 10 that's one of the the fact that you're doing a project basically you want to type it up and think how this is going to be the biggest thing everyone's going to buy it it doesn't look like you're trying to do the opposite you're trying to punch it down it's just a small project you know what is up that's kind of what we don't do that's what we hope that you don't cause any disease because if you have to charge if you have to use the link for you it just wastes everyone's time the whole point is that someone can get up from literally 30 minutes one of the cool demos we did was Sir Asha we gave a talk at one of our conferences about phone messaging and in the talk itself he on the spot created a feedback form for the talk live from everyone the fact that he didn't have to like use the link for this so from the point of how we did this so getting that small project and no giving cycle understanding like in the government agencies there's a lot not very tech literate they probably have a lot of problems like one of the we have problems understanding was so like cause what I do is and customers are SMEs I face a lot of problems all these customers they don't know like how to even access application so what are other than eye tracking thing that you have done is there anything else that you have done to that is like to help this slightly older in the agencies to your application so the best way to educate and support users right now we have 65 agencies to identify ambassadors in each agency so a lot of these are actually volunteers so I mean some of them are actually so I guess the way gavtech works is in this building we have all the practitioners project managers, engineers, designers in this building but a large part of gavtech is actually situated across different government agencies for example at MOM we have a CIO over there and his team that helps out with the IT in MOM so what we did was actually work with people across various agencies and they will be the ones that hand hold public officers to digitize their forms let's say you know if there's a less tech savvy public officer in these agencies we could contact the CIO team which will help them build formagy forms help them understand what the script is and some cases also organize workshops we've organized quite a few some for MOM, MND etc what we do is that our team of let's say 4 or 5 will go down spend maybe 4 hours and they will invite even hundreds of their public officers to come down and we will digitize the forms for them on the spot on that day yeah Hi, my name is William, I design I have one question about open sourcing the project what are some thoughts behind open sourcing this project and like are there any special government if you decide this project it's going to be open sourced you have to fill in all the check forms open sourcing with the team okay let's open sourcing I think I think so thankfully in the government we are in sort of weird crux here because we personally believe that this public money insurance will be public it's about the same argument for like public dependent research that we opinion publish and if we spend taxpayer money doing this there's no reason and we don't make profit of it there's really no reason to not share this with people that's unsolved on the practical level we are in this nice juncture where like our senior leadership doesn't care too much about this stuff but they practically don't find the side with the sourcing so we don't have to bring the people to the tech community so go ahead that's pretty much it there is like we are sufficiently ahead of our time that there are no regulations around open sourcing things yet so we pretty much do you see that our bosses are like hey we're going to open sourcing we have dozens of things open sourcing already so if those guys follow the audios and things and they're nothing super major just yet but we have a few things like the line is partially open sourced forms from the open source parking can be open sourced once you clean up the whole little bit but yeah nothing the major concern open sourcing is that someone will basically build the sort of scoop replica fishing site which should be a concern because it's a little bit easier but really if you have that kind of vicious act that you can do that anyway because it's not what the HDNL man so realistically the risks are very low benefits are pretty high it brings in good will everyone starts using this thing that works a lot better also you're completely right so we actually haven't thought of this process yet so our thinking is as in the beginning maybe work with a few agencies to figure out how this process works I was I mean just on the comment here I was thinking that we will want to conform to our design guidelines at least so it's either a choice between just giving the agencies I guess something like you know you know just giving the agencies I guess some components to work with or letting them make the logic based changes and then for us to do a bit of work to make sure it fits in with the user experience so it could be either or perhaps any one who has the left and right this is a process that we still have yet to work on right last question send me those paper forms do you have a question? I have a question so like are you open to contribution from external actually you can say like if I'm just a random insider maybe it's like for an issue or that we're coming forward yeah definitely I mean in fact right now we are iterative, same class, all those things but our plan is to like shut them out before open sourcing as well to become this like generic form of the tool as well when we first thought from telephones we realised that there were quite a bit of messiness in the telephones code as well we spent a good like 13 months cleaning up a lot of messiness so we really wanted to contribute back to the community we saw that there were a few thousand stars a lot of dead fox so we wanted to contribute back to it that would be part of the team too that would be part of the team too that would be part of the team too that would be part of the team too that would also be there you guys can hang around the network for things to come big thanks to Michael the friend said we're not able to make it because in the YouTube channel there's a lot of cool stuff from other videos as well so thanks for coming please don't finish the pizza if you don't finish we don't get to go home see you guys next time thank you