Friday In-Focus · 24 January 2025
Cognition's Dale Gallaher and Ansgar Liening argue for data-centric design control over document-centric DHFs, with a live Compass.Med demo.
Summary
The session opened with introductions and catching up before Joe Hage handed the floor to Dale Gallaher, president of Cognition. Dale described his route into medtech — mechanical engineering at Bucknell, propulsion systems for ships and submarines at GE, a master's at MIT — and how Cognition grew out of thesis work and a Design for Six Sigma tool. The company has just under 50 employees, and its current product, Compass.Med, is aimed mainly at medical device companies. Dale told the story of an orthopedic implant manufacturer that collected 483s because its paper-based design history file made change impact hard to track. Nellie Bushman pushed back: carefully structured Word and Excel logs with a master index have carried her clients through repeated FDA and notified-body audits without problems, and she advised building logs that could later be imported into a real system. Dale agreed the manual route holds for single-device companies and said the switch comes with complexity. He named Jama as the main competitor and Greenlight Guru as mainly serving smaller firms, and cited a Cognition-sponsored survey in which half of respondents were unhappy with their process. Dale then walked through Compass.Med: pre-configured templates, a data-centric model where requirements, tests and risks are single objects, configurable risk management, reuse libraries and automatic trace documents. Ansgar Liening closed with a live demo showing document sections held as items in a database, one-click PDF export, item-level change tracking and tracing a requirement up to a user need.
Chapters
Times run from the start of the recording.
The slides
Take these with you
Transcript
Edited for readability: filler words, false starts and Zoom housekeeping removed. Times run from the start of the recording.
Hey Sean!
Morning.
Hello, Anskar!
Hello there. So I'm I'm a little confused because I've shown Smith twice here, but there's different faces.
Is, is it meant to? I will. I will change my name confusing Shawn, and I share an account. And when I'm on, and don't see that he's on yet. I'm like I better be sean So I tried to be Joe, but he didn't let me in yet. So I was like, Okay. I panicked and became Sean. Hello, jail.
Hi! How are you doing.
Good. Thank you.
Joe. I think I have met you in London once, if I'm not mistaken in
You are not mistaken.
Probably I don't know. 4 years ago, maybe.
I was there twice. I don't think no, I'm sure you weren't there, but I was last there the week after the Queen died.
Oh, I don't remember. Well, I remember that, but I don't know whether.
No, you weren't there and I remember the week leading up to it. There had been a threatened strike of the British rail. And I was like, Yes. okay, what did? What am I doing? No one's gonna come, there's a strike on the rail. And then it was like, Oh, there's no strike on the rail because the Queen is dead. I'm like, Okay, this is not ideal timing. But you.
You. You had nothing to do with her death. The way.
I had nothing directly to do with it. The reason I even had that event was because I was coming out to support my partner, who was to present a paper at a conference, and I'm like, Well, if you're going, I'll go, and if I'm going I may as well have a conference day before the event. She listened to her 3 young daughters, did a cannonball into a body of water that was far too shallow for a cannonball, shattered her ankle and ended up going by myself. Which was okay. I just. I still haven't gotten over the 3 West End tickets that I blocked for her. They just went empty, and I don't like throwing away food. That's not.
you know. It's not sure on bad yet. So throwing away West End tickets was. It's tolerable for me.
Yes, I can imagine.
Yeah. So I'm sure that no, we didn't meet that time. It was the time before. and good to see you again. Now, Dale. I've never been in person. I do remember his street white hair.
But it's not always been that way.
Well, how long ago did you? How, when, how old were you when you began to go white.
Oh, I mean, I would say it's probably been 15 years, but it's definitely been a gradual thing over 15 years the last 5 years, pretty pretty white, though.
And now I'm thinking you're not more than in your fifties at most.
No, I'm definitely, yeah.
Hey!
Have a bad day.
Take Dale, take it, take it.
Crazy. Yeah. Jerry says that I started the company when I was like 6 years old. So and maybe that would add up to about the right age. So.
Hey, Maya, that you're able to regularly join us now.
But but yeah, I try and stay young looking. So I appreciate your your comment.
Do what I can. Oh, there's Ben, I hope, Ben, you'll you'll come on as a panelist. You're gonna teach us a thing or 2 today.
How we doing, Joe?
Are you going? Are you going with Ben to the Amsterdam Conference?
We are. We are still debating. I'm not sure where we stand with this, but it would be nice, because we need to catch up right.
Yeah, we sure do. And I'm bringing count, too.
We definitely have to catch up with count.
Are you bringing? Or that's.
Like Morty can't go.
I think he's too fat to fit in one of those little carry bags that has to go under the seat. You know.
I don't think he's too fat. How heavy is he?
He's 16 pounds, but it's it's more of a dimensional issue.
That's what.
Versus this. Yeah.
He's Comp. He's compact. That's what you're saying.
Yeah, yeah, he's compact, but not compact enough or squishy enough to shove underneath everyone's seat.
Nelly, do you have enough snacks for all of us, or.
Hey? I've been on a road show the last 2 weeks have been brutal. I've been in the middle of an Fdb. Virtual inspection on top of a 5 day, FDA on site inspection for 2 separate clients. So I'm enjoying my little bit of snack this morning because I've now done, and I get to breathe for a minute.
Glad you're joining us.
Yeah, I won't believe I was also hosting a FDA audit just one day remote this week, and towards the end of the audit one of the young engineers who had had baby this year. She's like, hold on, I need to go get something. She went to the other room and brought the baby into the conference room with the FDA inspector I was like about. I'm like, you've got to be kidding me, because, like you guys know, I have that. Never have I ever Youtube shorts. And then so the baby started getting fussy, and she's like, I think I need to find somewhere to feed the baby and she they were. Then they started debating on what was the best office that didn't have a window to do it, and I was like, I am going to die if she breastfeeds that baby.
Oh! In this.
On it. And I was like, I'm just, gonna is that.
Are you? Are you afraid of breath?
Have I ever said all? Never have I ever. So I just have to stop the series right here right now.
Afraid of breast because I I used to breastfeed my child. Whatever I was like the kid needs to eat. So to me it was not a big deal I would have, and if I had one now I would bring it to. If I had an important meeting I would bring my child.
Don't problem that she brought the baby I was just talking about. If would you breastfeed it in front of the FDA inspector, though.
Yeah, it's a food. They don't have to look at my breast. You can put a blanket over there that it's like, what's the difference between a natural food and a bottle it only if you sexualize it.
I think in my personal opinion. I actually want to sit back for a second. When we do auditor training. You don't bring personal items. You don't bring your pets, your kids, into the because you're trying to create an environment of professionalism and keep the inspectors focused. The minute you add distractions. You create frustrations, and then auditors dig in and audit harder, and they want to dig because they're irritated with you. So the last thing you want to do is bring unnecessary noise or distractions into and frustrated auditor. So that's I think, where Michelle's going. I don't care where you breastfeed. I don't want anything but the people they're interviewing in that interview room. And that's it.
I understand that. And I know that that's that's the the go to. But since we're moving into this more New Age era era, I'm thinking to myself, had I been a young parent at this stage now, I'm an old parent. I don't think I would. I think that it seems like something that I might have done. I'm too old to help to reproduce, but I understand the concept of oh, it has to be professional, but the way that things are in the new age that we're in I don't know. Maybe maybe the auditor doesn't mind. There's other things that have become distractions. But anyway, I would be shocked too, but at the same time I could see myself doing that if I was a young one.
Speaking of parenting Jose, you it would leverage the birth of your daughter for Linkedin. Content! Shame on you, and I congratulate you at the same time.
I learned French congratulations, congratulations.
Absolutely works. People love that shit, appreciate.
She's doing great. She's doing great. Yeah.
Now did I understand that she needed a special device? Was she in Nicu, or something.
No, no, not not my last son, ended up in Nicu. But no, it was just. It was more kind of on the observation of all the equipment in the room. and you know my wife was tied to like multiple Iv infusion pumps. And you kind of think about like one of those things. Malfunctions like that could kill somebody right so. But no, the baby. The baby was great.
Okay, good. That's good news. Well, special guests on Scar and Dale here to teach us. We are here to learn, please, please kick it off.
Okay, so I'm gonna share my screen here and share. And everybody should see the 1st like the cover slide. Everybody see that. Okay, cool.
Coming to you.
Alright. So just you know real quick about who am I? So so I'm Dale Gallagher. I'm the President cognition. But from the standpoint of what a primary responsibilities there it's on the technical side of the company. I grew up in Western Pennsylvania and pretty much enjoy anything that's outdoor related, whether it's hiking, kayaking, hunting, all of those kinds of things. And I have climbed all 48 4,000 foot peaks in New Hampshire. So that's kind of one of my achievements that I'm really proud of outside of work. From a professional standpoint I went to Bucknell University, got my Bachelor's Science Degree in mechanical engineering, and after Bucknell I worked at ge designing propulsion systems for ships and submarines, which was a lot different than medical devices for sure. And then, while I was at Ge. I went to Mit from master's in mechanical engineering, and then helped start cognition, you know, using a thesis that I did, and a thesis another person at Mit did.
And so since then cognition initially had a product that was an analysis product. And we sold that off to another company way back in 2,006. The current thing we're talking about today. We started way back in 2,001. I wrote a prototype. We went to a show. Everybody liked it. It was for design for 6 sigma and to help people that we're getting trained on implement, I don't know. Is anybody on this call like had a design for 6 sigma training? I don't think it's common now. But back then it was really big and that's kind of why we wrote the product. Because people needed something to kind of implement that
methodology. And it really turns out to just be a rigorous design process that aligns really well with the design control process for medical devices. So that's why I think we've kind of worked our way into that. that area of medical devices being our primary focus for for our product.
Yeah, how many employees do you have? Now.
So we're just under 50 employees.
50.
Yeah.
I I congratulate you. You're you're responsible for at least 50 livelihoods. I think that's fantastic.
Yeah. And and we've been, you know, like, we've been around for enough years that we've had a have a really good base, you know customer base with some of the larger company mid device companies. And you know, we're really trying to get into that medium sized market, and you'll see kind of where that fits in here as we go through the list of the slides we have here today. But that's pretty much kind of where I in cognition is. And so I kind of want to get into like, why, we're here. And you know what we think the industry needs and why, we're probably gonna be able to help out there. All right. So on this slide, I'm going to let you guys look at the bullets a little bit. But I'm just going to tell you a brief story that I'm sure a lot of people on this call can relate to. So a customer came to us, probably about 8 to 10 years ago. That, you know, was a orthopedic, implant design, you know, medical device manufacturer, and they were, had.
you know, started with some 483 warning letters, and the FDA came in, and he found other problems. And it just grew and was spiraling out of control because they had a paper based system. So everything was, you know, in word and excel and printed, and in a huge room full of file cabinets, and that was their, you know, design history file. And when they were making changes to these devices they were really having trouble, keeping track of what was really impacted by the change and getting that right. And that was what was causing all of these warning letters. They decided, you know, we can't do it this way anymore. We need a system where we put all of this data in. And it keeps track of all of that connectivity. And we know what the impact of any of those changes are. And so that's kind of what this slide is getting at. You know, it's like, if you do in a manual process. It's hard to stay, you know, on. Keep track of everything that's changing. They think some items can be in lots of different places. The change management becomes hard.
You know the the review process for anything, both new and everything is hard. So everything gets simpler because a lot of the information is automated when it's all, when one item is the same item everywhere you use it. And so that's really what our system is, you know, designed to do is to make it so. You, you know, if you have a requirement or you have a risk element, or you have a test. All of those things they're only going to show up. There are only one object that shows up in lots of places. So if you change it, it updates everywhere. So I'm curious around that story. I mean, I'm sure you guys on the on the call have experienced something similar where you know you have somebody that's having these problems with their manual process and trying to get the automated process. Does that resonate with anybody here.
No.
So. I I do virtual quality management systems and you know I do leverage the Excel logs and the Microsoft words. And but the thing about it is, we've actually really thought it through. and I have been been through at least in 15 years. I've been through, at least for FDA, at least 6 or 7 Fdb, and I've had 0 problems, the clients I've had problems with with legacy documents. So if you're 100% Pushman, Qc. And you have the master index in, and you have the Dhf. And you have the Dco logs, and you have. You have established it. You don't have any of those paper problems that you're talking about, and that is to prevent that validation cost right? Because a lot of your small startups don't have the validation cost.
and then, if you do have the extra funds, I always suggest you go into matrix requirements to do your tracing. because that creates a nice traceability from your user requirements to your design requirements, to your risk, to your validation. If you do that in pair in parallel to a Vqms. You're you're usually pretty good, I mean, I literally had the Ftb. Inspector. I'm not lying. Say, they're really lucky they put your system in because it saved them so they could get a license, and he's actually the 3rd time he's audited me. So we've kind of like he's like, Oh, I know you. And it was really kind of like kind of cool. Yesterday.
Sure. Yeah. And I think that you know, especially, you know, for the, you know. So the smaller companies that might only have one device, and they don't have families of devices. I think the you can definitely do a process like that, and it can work out. So the the question really comes. When does it become. you know, a complex enough process. You know, with enough information that it gets harder to maintain that manual way, and and that's going to be different for everybody, because in like, if they have a process like you said, where you have a pretty good index of tying everything together, then you can keep things together if it's less organized, and maybe it's sooner. So I think it really depends on when
somebody gets to that point where they really need something that automates all of this connectivity and and the production of the Dhf content.
And then Dale, another contributing factor is the legacy employee in the legacy location, and, like you said, they have a file cabinet. But I can tell you. Everybody I've ever seen with a paper system. It's always well, hey? Joe has a portion of that on his desktop, and we he's in, you know, Baja, until this time, because he's, you know, he's the one who created the device. Right? And so there's always when you have that paper system. Even though you have file cabinets, there's always somebody who has that record. And so it creates a problem.
Right? Yeah. And and like, who had the last one and who made the last change, that kind of problem and and document management systems can help that a little bit. I think, too.
Yep.
Gail, do you consider Greenlight Guru a competitor.
So to some extent, we don't typically run into them head to head as much more because they really focus on that smaller Medical Device Company, where they're also providing the expertise of getting them started. And so we don't run into them nearly as much as we might other competitors. you know. I think that green light Guru still has a a thing in the mix. But I think when people begin to get to that complexity where they really want a data centric model with some of the functionality that we have. That's when we would see somebody maybe want to move to a product like ours.
Yeah. But I also like many users on Greenlight Guru, say, how complicated it is like the logs that I've designed. I've designed them specifically with the concept of in 5 years. I know 5 to 10 years depending on the company's ready. I've set my data to where one field has 1 point. So like I could meet with Dale, we could create, you know, imports. And he could literally move our data in. And we should have the ability to move anywhere from 80 90% of the data into his system quite smoothly, because a lot of people when they're creating their logs, they'll say this column has 5 points. Well, with Bqc.
That column has 1 point, and that's it. And then that makes life easier.
Right? Yeah. And so we do have like, when people are coming up into our product. You know, they have the data that will import in and and kind of mirror it all into our data model.
Who do you consider to be your chief competitors?
The number one competitor that we run into is drama.
Is.
Jama, software.
Oh!
They're they're more they don't. They have certainly a a market in the medical device, but they market themselves more as a general design tool. So they're in in the other markets where we focus just in medical device space. And and because of that, we have some unique capabilities for the medical device part. That drama doesn't have.
Yeah, deal we have Jama at the moment. Can you elaborate a little bit more where you stand out above Jama?
Yeah, absolutely. So the probably the I believe the the number one area that we stand out is our ability to produce the your Dhf documents and exactly the format that you want to match your quality management system drama is very tabular centric. As far as the way you put the data in. And once you get everything in, you have to work to get the format out you want. And that's what you know. We've heard from people that have migrated from Jama to us. That's 1 of their biggest issues. Another big area is that we have a very deep risk management functionality. And if you guys have, I'm sure you've run into risk management people. Everybody has a different way of doing it.
and we have our our product is very configurable to match those risk management processes. So we, the tying between the risk management and the requirements, then, is really tight in our system. and then I think the the last piece is, you know, our ability to configure generally the kind of what you want to see, and so on in our product is definitely superior to those are probably the 3 big areas that we see the differences that make sense. All right. So so this next slide, I'm not going to spend any time talking about the data. The the punchline here is that we did sponsor a survey last year to medical device companies and 50% of them said they really aren't happy with their process.
And only a 3rd of them were really happy with their process. So there's definitely people out there that are looking to improve. And that's really what this slide's about. If you want to request the slide set at the end of the after the webinar. The link here at the bottom will give you the access to the the details of that.
Maybe I'm Scar, or or Ben or Gerald. You could put the link in the chat now for our guests.
For the down, for the report or for the slide set.
Both.
I'll let on Scar. Figure that because I don't actually have I was gonna send it afterwards. I don't have a link ready right now myself.
I don't have the link either, so I'm I'm sorry.
We'll make sure we get it to everybody.
Thank you.
Okay, so this is our pro. Our product is called Compass. Med, I'm not gonna read all of the details here, but really the one of the big things for medical devices. It's a it's a saas solution, you know. It's it's pre-configured for you to design medical devices. If you don't have any opinions on exactly how you you do your design process. You can use compass as is out of the box, and you can get through your requirements. Test and your risk process. It does have like, I said, though it has configurable options, so you can also tweak it to match places where you might have different opinions than what compass provides. So it's that's that's probably the biggest thing for medical. This is like, if somebody wants to get started. They can just start if they don't have to do a whole bunch of configuration effort upfront.
We do have as part of that pre-configured environment. We have these off shelf templates. You get that, you know, will create your design, inputs, documents, your design output documents, your, you know, test documents, your your master traces and and on scar. I'll show you some of this in the quick, demo we're going to do. Towards the end of the the call here. And all of those are really available. At the click of the button you just click your Pdf, you get it out. You can do all of the review and approval process inside compass, or if you already have a Qms. Where you're managing your documents, you can take that Pdf over there and you can manage it over there. So there's you can do that either way. We also have a lot of functionality. So you can. You can integrate those 2 things together. Well.
so that's kind of the the key part of compass. It does. The requirements risk and test management ties everything together. So this is just a slide that'll show you how things are connected. So again, it's a it's not document centric. The documents are the output. You get out of this out of it. But the idea is that you really want the data centric model. So you know every requirement, like what its parent requirements are and what its child requirements. So if you look, we have a design input here in the middle and above it, the user need that it satisfies. And below it are, you know, your lower level requirements that that satisfy the design input.
And then on the left side of the of this trace you can see test documents that might be referring to that. So you have test protocols and test methods that it can be connected to. And then on the right side, you see the risk information that the requirement is connected to you can navigate through this. So like, if you pick. you know, the user need, it'll put the user need in the center and show it's it's so it's kind of you can just walk through the data model using this this page. So this is the trace document that you get automatically when you're after you build all of this. So one thing you do, you know, and I think we're talking about the manual effort, you know, if you do all of these requirements documents and you
create all of these, you know sub requirements and parent requirements, and so on you still, then have to go to an Excel document typically, and you'll put manually put all these things in, and you need to make sure all of those connections are there, so that you don't get anything wrong in in, because everything's already connected in compass. This just comes out automatically. There's no manual effort for whatever in in on Scar again. I'll show you a little bit of this during the the demo portion of the call. Go ahead.
No, I will. That's right.
So so this tray, this automatic, tracing functionality, we, we feel, is one of the most significant reasons you would want to go from a more manual system into something like compass. Med. And I hopefully, that resonates with this group, I think, especially as it gets more complicated. But that's probably the most significant power you get is, you have all of that traceability, and you don't make those mistakes when you make a change. You know it's everything that is connected to.
Are you guys coming to 10 x.
In May. Are you guys coming to California in May.
Sorry we weren't gonna talk about that after today's call. I wanna introduce them to the group before.
Ben and Ben and hans found all the group already. But sure, okay.
Okay.
Yeah, I think we. So yeah, I I do remember mention something there, Shawn, but I forgot about the details of it. All right. So this is a lot of information on the slide. We tried to highlight in blue. You know what the real key point was of all of these items. And you know the left side is like who we think can benefit from this. And and then the right side is, why do you want to use this approach versus, you know, a manual method, or maybe a document centric application.
Are you?
And you know I talked about like an earlier about, you know, if you're having a lot of FDA warnings because you don't have that traceability and you're having trouble maintaining it. I think that's a good reason. If you're having product recalls or product quality problems because you're missing that traceability and you're getting some of that design information wrong. We provide that traceability. One other thing that I'm not sure you'll see much in the demo, you'd see in a more detailed demo. But we have a lot of data reuse. So like, if you have a family of products, you can have like a product that
you have a library for this product, and then all the product families can derive from that. So when you make changes to common elements, all of those, all of the products in the family can pick up those changes. We also have like risk libraries. And you can have requirements, standard libraries. So things like that that you can reuse in all of your projects. We have a risk, a very rich risk grant management piece. You won't. You'll just get a very cursory look for that at that today with Onskar. But again, we have a lot of details of a prior, you know, a fairly longer a longer demo than what we have today.
Everything is tracked and audited. So whatever change you make on anything, you can go and see exactly who made it when they made it and that that helps the any kind of change management you might have. And then the it helps you be compliant. And so we we guide you through a way that we believe you know, and worked with. You know, other regulatory people to make sure that we're providing you a compliant process that meets the standards. We've talked about the templates and and the product lines, the central libraries. I think we've covered everything here the last piece is I talked about earlier is, you can just get all of your Dhf documents, you know, with the click of a button.
Any questions here on these bullets. There's a lot of information on this slide.
Good to see the demo.
So it might be. I'm going getting through these faster than I thought, so I think we might have more time for the demo on scar like if his questions come up.
Yeah. Well, I expect people to to ask questions while we while we go on here. Right? So that's fine. We're not in a rush.
Right. This is just a an roi potential way. You could calculate it for a given customer. This is gonna vary, depending on exactly what their process is before and what their process is after. But the whole point here is like, if I have a lot of stuff that's connected together. and I had to go through and do all of this manually, and assess everything manually. It could be really expensive. And then, when you but if you have everything already connected together. You already have a lot of those answers, and you know your changes. In one place. Every place that uses them is up to date. So this is in here mostly as a sample. I wasn't going to spend a lot of time talking
on this, but hopefully the the gist of it makes sense, as far as how you would reduce this from your manual process, going into something like a data centric model any questions on this. Alright. I'm gonna go on to the next part here, which is, gonna I'm gonna give it back to Onskar so he can show you what that looks like.
Oscar, Jerry Oscar! Excuse me, Jerry Wessell here just came by to say, Hi, everybody! I was chatting with Kristen, and she will put the link to the slides as well as the market survey study in the chat.
Already there! Whoops!
Perfect in Italy, alright, enjoy the enjoy the food and wine. Great thanks, Kristen.
Okay, so I took control. Can everybody see my screen? Okay? So I I usually Demos take a long time, especially when we go into a lot of detail, so I will not bore everybody on a Friday. What is it lunchtime for, you guys? And Friday evening for me with with all the nitty, gritty details, although I'm a techie guy, and I like this sort of stuff right? So so you know, stop me if I'm getting too detailed. But the aim here for me is to show you. you know why compass is a little different, and how it actually works. You know what's what's the the basics behind that? And hopefully, that helps the gentleman who asked about
Jama earlier. I think Loyk, is that the right name like. I hope I pronounce that correctly.
Yeah, it's a. It's a ditch way of writing. Luke.
Okay, Luke. Okay, look, I can hopefully remember that. So good to meet you, Luke.
Hi! How's going.
So compass is web based. I'm here in chrome. And just for general navigation or orientation we have a couple of sort of navigation windows here on the left hand side, and we have a main window on the right, and whatever you select here on the left will be displayed on the right. That is sort of the the general introduction, and that is all I want to talk about here. You see, there's nice graphs and pictures, and all this sort of stuff. That is really something for a longer demo. But in terms of principle, let me go straight into the guts of this. I work with a lot of customers. And just to give you
one example, I have a customer. They design Ct scanners, and they have something like 100,000 requirements, probably more and 60,000 tests and risk items. I don't even know how many they have. So that is for us a pretty typical project scale, you know. It's not even a very large company or very large product, but they have product lines, different offshoots of the main product, and they all need to be managed, and their issue was that they have a lot of documentation, of course, and generating these documents was one thing. Of course they can do it all in word, but they needed to manage and and literally micromanage. All these different requirements, and all the different connections that these requirements with the tests and the risks might have.
And this is where we come in. So our data model is is, of course, based on a database, and literally everything is represented in the database as its own object type, its own item. So if you look at this one here, this is what we would call a document. It's, of course, an item in the database. But if you look through that, it has a purpose, scope, definitions. All these sort of sections that are here which are in their own right, also items in the database right? And if you have a section like this purpose one here, you can simply click here, make some changes or additions. Yeah, so you can simply work in these sort of sorry. Try to avoid spelling mistakes, but I can't promise that I'll be very successful, right? But you can work in these sections in these, with these items directly with in in this web-based environment and
You can, of course, change the layout here. What people generally do is they create the document set that they need based on their sops, and they'll become part of the compass template. We are using templates a lot. You can create documents, additional documents based on the templates that the customers might have. You can also create brand new projects based on templates. That is very popular, because if people make a syringe again after they made syringes for the last 20 years. There's a good chance that many of the requirements and tests, even and probably risks, are the same or very, very similar. Right? So you can generate new projects based on project templates. And you have a running start with with stuff that's already there.
But once they start filling in these documents, here's a section that is somewhat special. And you see here, this is now having some design inputs right here. This is where you sort of zoom into the microcosmos of the requirements, and how people generally work in these is very similar to words. They might say, there's a new input, let me just check what I wanted to write here, they might have an audible alarm. Yeah. And they create a new item right here. Okay, audible alarm. And they might fill in more data about this, for instance, there might be an acceptance criteria should they'll be louder than 90 decibelts.
Yeah, so you can sort of create this item, and you can then fill it with data. And this is a general way of how people work. They just fill these sort of documents in. And what Dale was saying is at the push of a button. We can now create a Pdf. Or a word version of this, and I have a button right here. So if I just generate a document from this. just has to export it real quick. There we go. I put it into my downloads and save it, and just to show you here it is, and this is now our document that we just generated. Right? It has a cover page. It might have headers and footers. In this case I don't have a header, but I have a footer here. It creates a table of contents
and based on the work we did you see here we made some changes or additions right here we did that. We also created new requirements, and here were the requirements that were there already. So people generally work in this database environment in this document, for example, fill it in, and then the document can be exported. Of course you don't need to just take the raw data here. What these documents provide also is a way to review them. So you can do reviews right here. I'm not going into that, because that takes a little while. Happy to show it another time, maybe, but you can also submit it directly for approval, and approve and lock these documents, but also the data that's in them. Right? So you have a full control over the bucket, which is the document here, and also the sub items, or these little items, like the requirements that are part of that.
Anskar, how do you track changes? I mean, you have a big project, a lot of design inputs. And now I want to see what somebody has done.
Yes, so so we. I show you this one here to start with. We show, modify, and we track modification on each item. So in this case I make the audible alarm, and you see here that it tracks the changes that have been made here right. And we we don't have a report like this out of the box, but you could generate a report that says, Show me everything between these 2 different versions of the doc of the project or the document. For example. yeah, generally, people look when we talk about the documents, and I don't have a good example here right now. But the documents themselves have a history, of course, and you would be able to see all the changes that happen to this document between different versions. You can have a differencing tool here. That shows you all the differences between versions of a document. For example, right? Because I haven't prepared this. I cannot demo this. But I could show you that in a proper demo.
Right. You could see it did pull in at the document level. The changes that you made for the audible alarm.
Correct. and then you can see these changes. For instance, in line, you would be able to see a document. Sorry I have to sort of describe it verbally right now, but you would be able to see the the differences in line within the document, with the additions in green and the deletions in red, and this sort of thing, or you could compare them side by side as well. This is both possible. Okay. all right. So you saw already. Here I have changed the view slightly. You know this. The default view is the one that we would export as a as the submission document. But there's working views as well like this one here where we show a bit more information about these requirements.
And so let me just change to the one here that is in green, and if I open this you have different options here. It shows you modifications. I showed you this, but, for example, you can trace to a higher level directly from here to a lower level. You might want to model interactions between requirements that is sort of based on the 6 Sigma methodology, where, if you optimize one requirement that might have a positive or negative effect on another one, so we can model these, you can generate or link to test methods right here, failure modes and so on. But I want to just show you the trace to the higher level. And you see this one here. This requirement here is already tracing to user needs in light blue. The reason why I highlighted this is I want to show you 1st of all, a more formal trace.
Here's a design trace matrix. So if I click to this on this, you see, this is the the user need number 18, the one that I highlighted in blue. And currently, this one traces only to this particular user need the one in in green. The formal traces are updated constantly based on whatever work you do in your in your environment. So in this case, I'm going to the next requirement down here. And you see here, I can also link it to a high level requirement. So in this case I want to link it also to Number 18, which is this one here, which is exactly the same that we used for the other requirement. This reuse idea is very big for us. Okay? And I don't know whether you saw that when I click on this button here, it shows me a whole bunch of requirements that are currently in the system. Right? So you see here the user needs
and it takes these, the user needs particularly, it takes these from the user needs area. Right here. yeah, so we have a user needs.
Is that a a list of items for the world, or is it for this user.
This is for the whole project. So any user who is in the role of a designer basically working in the design space can link requirements to each other. Yeah. So this is available to everybody. We do. However I might. I might know where you're going with this. We have libraries, for instance, for hazards and harms as an example that are above the projects, and they can be used by by any project. Yeah. And we're this, this library use, and and the idea of reuse reuse from libraries is very big with us. And I think Dale talked briefly about the reuse of components. That is something we will go to in the future, where you can reuse whole whole items and reuse them in your project with everything that comes with it
also from libraries. And you, when when I show you the the risk a little later, I can briefly show you how the libraries work, it's really like just a picker that looks into a standard library that is managed centrally. But in this case, you know, these requirements are coming from here. And so if I added another user need, for instance, here and I'll just say device could be very attractive. I'm adding this into the document. That is, the user needs document. It's right here. If I go back to my design inputs. And remember, I
I added this requirement already here. But if I select the parent requirement and just search for attractive. It's now here, right? So these pickers are dynamic, you know. They look into the database at any given time and any changes, and any additions that are made elsewhere will be reflected here. And you can then start using these components immediately. Right? So there's reuse on this small scale with stuff that happens right here, but also reuse from libraries if you want to. And I'm showing you all this because if I go back to my design. Trace matrix. And I'm going back to my requirement that I highlighted here. The user need rather, that was that are highlighted in blue. Here. You see now that this trace matrix updated automatically because initially, it had only this link. But now it also has a link to design input 14.
Yeah, so these traces can come very large. We know that we have some customers, and they told us that that you they have, like one person working on a on a big trace like this for 6 weeks. and they're still not certain that it's accurate, because after 6 weeks things might have changed, and and they are not 100% certain that the trace is up to date. Yeah, of course, this was a very, very large trace, but you see the complexity, and how how quickly it it sort of gets overwhelming if you have like, let's say, 100,000 requirements or so, and they all need to trace with tests and each other ripes.
So let me just zoom in, because that was another thing that I mentioned before. We have the bucket here with the document that we want to export. But of course you can zoom into an item like this one here. So if I just click on it, I go directly to this item in the database. and you see there you have a lot of options here. I go straight into the trace view that Dale also showed you earlier, and this is a live trace. You know. This is an alternative trace view, if you like, from the official deliverable that you have here with the trace matrix. But this one is this trace is up to date as well, just as much as the other one.
And you see here, it now reflects that you're linking to user need number 18, this one that was in blue. It also links or or shows the link to the new user need that I just created. Yeah, it shows you that it's satisfied by some lower level design outputs. But it also shows you that it's part of a design of a of a test protocol right here. And the reason why I wanted to show you the test protocol is if I just open this. So there's a test protocol right here. and the requirement that this one is picking up is here. So there's a test in the protocol that is called alarm test. It has a certain number of steps, and it's testing these 2 requirements. There's requirement 14. The handle device shall produce an alert. When the transmitter battery has 12% charge, and the acceptance criteria is also 12%. Right here.
The reason why I show you this is when I move this out of the way. And I say, I make a change to this requirement here. I might say, here it's not 12%. But maybe I say 15%. And I change it in both places here and here. Yeah, in the name of the requirement and the acceptance criteria. If I go back to my test protocol, and I just need to refresh this because it's you know it has been called just a minute ago, and it needs a just to refresh the view. If I go back to my requirement, you'll see here that this has changed here and here. Yeah, the the point being that we have a single point of truth, we. We have this requirement one time in the database, and we reuse it now in this test protocol. And if I make a change here as long as the changes are allowed, of course, because it might get locked, based on all the workflows and the processes we have. But if I make a change. This change is automatically reflected here in in the test protocol.
So you don't need to worry about these updates.
Can I quickly ask a question.
Absolutely. Yes.
And I apologize. I had to log off for a minute, so if you already said it, I apologize. Do you change? Once the document is approved. Is there a change history that you can see? So version one versus version 2. Okay.
Yes, yes, you can. That was that was a question earlier. It was a little too much to show in this demo, but this is the sort of stuff we can show in a in a larger demo when we have a bit.
Yeah, okay, this looks really great. That's I was. I'm really excited about it. I just wanted to be sure. Thank you.
Yeah, no. Problem.
I've got another question. If my ask.
Absolutely.
So that being that this is web based, have you guys had any consideration to integrating some sort of some sort of guidance systems as you're going through these documentation like maybe integrating some sort of AI to maybe help with trying to guide what is right or within regulation, as you are making these requirements, as you're going through these design. as you're defining design parameters and making in regulation, so that even if you have some certain designer going through. They may not be completely completely versed in what the regulatory processes may be, and so just maybe as kind of like barriers and rails for your team. As you move forward.
Yes, so we don't have an AI in here right now, and I believe the correct answer for us would be to link to something that's already there. I don't think we would reinvent the wheel, but we would reuse something that's already there, and I think what you described there is. It's a very tempting thing. Help people, for instance, formulate well, formulate good requirements, because there's rules around how you write a good requirement. I can see that that would work quite nicely at the moment. We don't have anything. But there's a discussion internally about this.
Yeah, it's a beautiful template to integrate that into. Yeah.
Yes.
Yeah, what? Just what I wanted to say in in the past year we've also been having discussions more on the you know, the complaint side, and so on. To maybe like for the device you're doing, what are the possible things you might need to be concerned about with your device and make Kevin some kind of AI connection to that. That would give you some. I. The things you need to consider.
Absolute.
Can I move on to Amir's question? It's because I think Amir is referring to you. Make a new product, and you have some guidance. But now you have a standing product, and there's a slight change or addition, or whatever. And you know you only have to change small parts of your whole documentation. But you don't know which parts, because when you have a new product. You go through all the regulations, all the standards, everything, but for the for the the small parts. You don't want to do that. You don't wanna have the the 12 pages checklist. You only want to have the parts that are interesting for you, and maybe an AI guiding you to where the interesting parts would be, would be very helpful.
Absolutely. Yeah. I think the whole. It sort of goes for me, at least into the area of impact analysis. Where you say we know, we need to make some changes, but we need to understand what impacts that will have. And then I will definitely help with this sort of stuff.
Yeah, because all of the connections are there.
Yes.
And you can follow them in what's there today. But I, you know, automating that to just give you a better, you know. Report that shows all of the places without kind of navigating around them is something we're looking into.
Yeah, I mean that that's talk about saving time and and.
Exactly.
Shit. Yeah.
Your traceability matrix is very charming. It's it's very helpful. But if you have something new, you don't know which part of. But you know which part of the current matrix is affected, but know what should be added.
Yeah. So can I ask a question.
Sure.
What database is this platform use.
So I give that to Dale. We we are in the process of offer getting on a different platform. I think that's fair to say right. But Dale.
Yeah. So we right now, we're on an object database and that object database maps really well into our data schema that we have. We are looking at the potential of other databases that might be more ones that people are familiar with. You know some of the SQL. Databases out there so. But right now it's on an object database that runs on the server.
Okay, thank you. Because it's more of like text based. Database is not just a lot of text inside of your data set. So.
Yeah, so, and the one advantage of the database we have is the because the text very length is very variable. So it doesn't have any extra overhead, no matter whether it's, you know, a small piece of text or a large piece of text. So it makes it very easy to manipulate that data. But most of even the the SQL databases these days have solve that problem because you don't want to have that fixed text size.
Right, because I think it's wonderful to have a platform to fill all those information, but I think the search part will be what I will be most interested, because, like, when you have to go back, look for something. If your search portion is not as robust, then it will make it a little difficult to use. Does that make sense.
Yeah, we do have a general search thing. And and that search. If you go to a project, it basically will search everything in the project in by default, so it'll look in your documents. Look in your requirements, look in your risk, so it'll find everywhere. Something's is used inside that project. But we offer a lot of more advanced options of Hey, if I just want to find a requirement with this text in this item, you can also narrow it down that way. The other thing that we've done in the past. We don't have this encompass right now, but is to do more of a a full document, level search kind of. So you have an index search and be able to just get to all the documents that made use of something.
Thank you. Very good. Thank you.
Right. There's really only one more thing I want to show, and then we run out of time. So you see the trace matrix here of this particular web, my mouse of this requirement here, number 14 in yellow. There's all sorts of stuff. I showed you the protocol, but there's 1 thing here where we do the link or create the link between the risk world and the requirements world. Really, there's there's 2 connections, you know. This requirement might be affected by failure mode which then eventually might turn into a harm right so, or cause a harm. We don't have a failure mode, effect here. I'm not going into an Fmea for this, but I want to show you that this requirement is used as the risk control measure for and the risk control measure is called alarm when batteries at 10% charge. If I click on this, I change the view here, and I can see that this risk control measure actually controls these 2 risk entries in the risk analysis. So let me just go to the risk analysis. So that you
see how that looks. Here we go. You just open this right? And this is really, really high level. Now, right? We have a lot of interactions between our user analysis, the different Fmeas Dfmea. We do have a device. Fmea. We have a side effects analysis, all this sort of stuff, and they all can contribute to a risk analysis. And Fmea is not a risk tool, and according to the standard, it's a reliability tool, but it can contribute to a risk analysis. So, following the standard. In this case we have a hazardous situation right here with a. p. 1. This is now where the libraries come in, because hazards come directly from a library that is centrally managed by the risk team by the risk manager or so. And you can select one or several of these hazards. The same is true, for the harms you can select harms again from a library, and if you select a harm, it comes comes with a severity you cannot really alter this. It comes from the library, it's signed off, and when you use this harm it will come with this severity.
You can then define the p. 2. That the hazard leads to the harm. You have an overall probability that is based on a risk management plan that is, in the system that is yours that you sign off. But I want to show you the risk control, because our design input 14 is right here. And it's implementing this year when the alarm, the alarm and the battery is less than 10% charge. You might remember that we changed the requirement to 15% charge. And it's, of course, reflecting that right here. and you can select other requirements, additional requirements that might implement this. So this button here is showing us a picker again, with all design inputs and outputs that are potential candidates to implement this particular risk control measure. Yeah. And once this is implemented, you have a residual p, 1 and p. 2 and a residual overall. P. And then you can see the the residual risk here
that is really, really fast and really high level. And I spoke very fast as well, so I hope it gives you an indication that we do a risk management here as well. But there's many, many different facets. And the very, very last thing I want to show you is we have built in. We don't have a built in AI, I'm afraid, but what we do have is built work instructions for all these different areas. So in this case, there's the work instructions for the risk workspace. and it will give you guidance. How you use this tool here and these work instructions are not only here, I have a sort of a risk area right here, but of course you have deliverables in the risk management area. So if I go right here and I show you our risk product risk analysis, it's this one here that has all these risk assessments
that it reflect that it shows, including the risk analysis that I just showed you. But of course they all all these documents also come with work instructions, and if you deviate from the format that we have, this is also accessible as a library by the users, so they might have their own additions to the work instructions or their own updates, so that the work instructions reflect the way our users work. And that's why I leave it because it's now 5 o'clock. My time.
That's something you said in passing. That you don't have AI, and I'm thinking yet, are you actively working on integrating AI into your systems. I'll leave that for Dan.
Yeah, so we we've definitely been talking to multiple outside applications that are AI based to see see how we can integrate those into our environment. And the biggest, most of that's been more again around the post market stuff right now. But this next one we've been discussing is the the one that Monscar talked about, which is really some AI applications out there that help you write better requirements.
I'm thinking you'll need an outside group to help you with that.
Absolutely. We. We don't want to implement that ourselves. We want somebody to. That's already crossed that bridge to integrate with for those kinds of things.
I'm not looking for an absolute number, of course, in in this question, but from a budgeting perspective. how big a commitment do you need to set aside from your working budget to make that a reality is that you know part of your marketing costs. And it's 10% of what you spend on marketing? Or is it 50%? Or is it like a capital raise or something? I don't know how big is.
Get to the AI piece. You mean.
Yeah,
Oh, it's definitely not anywhere near the bigger raise, I mean, because the I mean, we're our integration capabilities are pretty flexible. And so I think if we find the right, you know, application to integrate our effort to integrate it's is pretty small on our side.
Well, I think, I think we've amply demonstrated today that we're your kind of people based on the the number and caliber of the questions that we got today. So thank you, everyone for.
Thank you.
There is in the slide set that we shared. There is a final slide that has, you know, the contact information for our, you know business manager for the Med tech space and and some other details. That So that's in the in the slide deck that we gave you the link to download.
And do go ahead and send it to to me and to Sean, so that we can.
In our library form.
For our members.
Yeah. And then Kristen will get that off to you.
Sean, some closing thoughts.
Sure you did.
Sean. Closing thoughts.
Yeah, I was just unmuting myself. Sorry. I I would just like to thank Dale and Onskar for coming. I met them recently was impressed with their solution, and really just wanted to share it with all of you and get your feedback. I guess what I would ask is, you know, from from a marketing perspective or as business owners or consultants. You know what aspect, if any, was most appealing to any of you that are remaining on the call. You know I would I would be. I would be curious to know which is it? You know the risk piece, for example, is one of the things that distinguishes this from other similar platforms. And.
Just well established risk metrics that it that the uplinks and the downlinks to to to all the requirements, are so.
I like that a lot, too. I'm interested specifically in Nelly's feedback on your question, Shawn. If she can come back on camera.
So I actually really like what I've seen, I have seen. I don't know if you guys are aware of this. Matrix requirements is a competitor of yours, because not only do they have the Qms. But they have something very similar. I think you guys have more of a visual modular based, and it would be something I'd be interested in looking at. But you know, for my clients it really is the cost point into getting to into something like this for them. A lot of them don't have the funds for it. Because and so that's why they go to that virtual paper system. But I I like everything it's like it's very intuitive. It's very. It looks very easy to use.
you know. I'd like to see a demo so I could like compare it to what I've already seen, and and kind of determine if I would suggest it to clients.
Pardon me, pardon me.
I was. Gonna say, Oscar, one of the things we didn't show is the workspace for requirements. I guess we didn't have time.
Right.
Yeah, there's different ways to to enter requirements, of course, and data generally. But it's it's correct. You know. You can work in a more tabular way, if you like, and just work like this, and and you don't. You're not married only to the documents.
So that we give.
The flexibility.
It. It kind of feels like it's more of. So like one of the things I've seen in different systems is like, you have this data demo, right? And then you go to a second module where you actually pull that data into a report or to a document for approval. And so what I'm wondering is the electronic approval of these things to create those revision requirements in a design history file concept.
Yeah, I mean, I think that, like Jerry mentioned, we do have this Workspace view, which is more just the data and not inside the tab, the document space. So you can still work in that space as far as matrix requirements. I think when it comes to requirement linkages, we're pretty comparable with what you would do with that, I think, where we stand out is like. We also linked all of the test information, your and your failure modes and your risk control measures. So those other outside links outside of your requirement space that, you know, is an additional.
Well.
Here.
I would challenge you to go look at matrix because they actually allow you to write the test script, approve the test script and execute the test script right within matrix requirements. And so then you're instead of actually doing this monstrous report, you're doing test cases that you link it to different requirements, and then you bundle the results of all those test cases to a document so that.
Yeah, that's similar. That's very similar to our test space. I guess I wasn't familiar that they had that piece. But yeah, so then they would have the same linkages to the test kind of stuff that we would have.
So I see you guys on the same competition plane. And so that you know, that's my feedback to you, cause I've been using. I've been using matrix for oh, my God! At least 4 years with my different clients, I'm not the expert I'm usually kind of. I set them up, and then there's somebody who comes in and really really configures them. But I really see that that's where your challenger is for the small medium space. If I'm gonna give you my feedback.
No, that's great, because, you know, we haven't. To my knowledge, we haven't run into them. But it's good to know, and we'll definitely take a look and try and understand where where we are. There.
What I want to know. Your story is how? So let me just give you these stories. A matrix was a bunch of engineers who got tired of all the Qms. Is getting it wrong. And so they designed it. Basically from their experiences. How did you guys come up with this idea? Was it just a business model? Was it your experience? You were kind of tired, or what inspired you to to start this.
So again it started back. you know, over 20 years ago, when we created the design for 6 sigma space and and the design for 6. Sigma had all of that same kind of stuff in it, with tests and requirements, and being able to link it all that together, and that's what they were training people to do when you went to the training course. But you didn't have anywhere you could do it. So that was our start. But then we started finding, like medical device companies that were doing that design for 6 sigma that worked with us, and we realized that when it especially got into the risk side that we had that niche in the medical device space. And then so working with those companies for 20 years is how we landed where we are now.
and why we focus on the medical device space.
I can say your visual is is a lot. Now, do you actually also do quality management documents in there? Do you have like a Qms element? Or are you specifically design control.
So so we don't sell ourselves as a Qms. But all the documents that you have in here can be reviewed, approved, and signed off inside compass. So we provide a lot of that functionality. But we don't try and compete against people that are like providing that electronic functionality.
Interesting, I, you know, like I said I would, you know, spend some time in. Obviously I can't to like after quarter one, because I'm I'm kind of drowning right now, but I would spend some time in there and like, go through what you're what you've just set up because one of the things I know that some of the other quality people and I are we're looking at is, how can I get? How can I set up a Qms. In something very simple. That's low cost that when my client comes in I can just turn on a button and push it out. And that's kind of what I've been trying to figure out so.
And at that level we have all of that. And and so the reason we generally try and stay away from it is a lot of. And it's probably because we've been with a little bit larger customers, and those larger customers tend to already have something in place that they want to integrate with. And so they don't really want us to replace that piece. But if you're small enough you don't have it. We certainly have, you know. the ability to review, approve, and click to get the document out.
I'm hoping that I know Nelly's coming to 10 x. I'm hoping. Dale's convinced that he should be there as well.
Please go.
I don't. Sean is leading the development of the agenda, but I would say that a conversation, a public conversation between Nelly and Dale, and perhaps others A workshop around this topic would be valuable.
Yeah, we'd love to do that.
I would have a blast.
Yeah.
We know these are blast.
Alright! Don't get me started!
Next week's guest is fan favorite Michelle Watt and Renee Franz. They're gonna talk about commercialization pitfalls. You guys, of course, are invited every week. and Shawn and I are gonna meet immediately after this to continue our conversation about developing the program and the like. So for Joe H. And Joe Enterprises, Sean and Shawn Smith Enterprises. Thank you very much for today's call. A closing thought from you, Sean.
T-bor is is talking. Next week Michelle had to move.
Oh!
It. Michelle and Rene had to move. So Tibor Zechmeister from Flynn AI in Europe. They've developed a regulatory platform. Cognition. You might want to come to this. You might.
Interesting, yeah.
You might want to see their AI based regulatory platform. And what what they're doing? Because it's very different. But it there. So bastion their see? Our co-founder and their head of regulatory are gonna be tbor Zach Meister are gonna be leading the call next week.
Well, let's get. Let's get that updated on on the webinar site. And. gentlemen, I'm very grateful for your time and contribution here today, and I look forward to seeing more of you.
Well, thank you for having us. We appreciate it.
Everyone.
Thank you.
Bye, everybody!
Bye, now.
Bye.
Bye, guys.
Thank you for your contributions, both of you.
Oh, I know. Yeah. you have to you. You have to throw me away. I guess I I'm missing a button.
Gallaher D., Liening A., Bushman N., Hage J., Gijben L. BRIDGING THE GAP BETWEEN DESIGN & REGULATORY feat. Ansgar Liening and Dale Gallaher. MedTech Leading Voice, Friday In-Focus, 24 January 2025.