Speaker A
Okay, well, hello everybody. The time has come where we will be going live um for this webinar. And today we are going to be discussing shifts in sourcing and procurement and focusing on Find My Factory's latest release, Spea.
Find My Factory's Matias and Morten introduce Speya, an AI chat assistant layered on their sourcing platform, and demo how it turns a sourcing request into a supplier discovery plan and list.
Key Takeaways
- Speya provides a conversational interface on top of Find My Factory's data and AI, making supplier sourcing easier for non-experts.
- A chat-based discovery plan can be refined iteratively before running searches, improving result accuracy.
- Grounding the language model in curated database context is central to reducing unreliable outputs.
- Sourcing workflows can be largely automated, with results exportable to spreadsheets for further use.
What the video covers
- Webinar hosted by Rebecca from Find My Factory introducing Speya, the company's latest release focused on sourcing and procurement.
- AI engineers Matias and Morten, who built Speya, explain its purpose: a chat-based AI layer on top of Find My Factory's AI that writes prompts, gives feedback and guides sourcing cases to a ready supplier list.
- Speya works from text, image or file input, and can be used for supplier discovery as well as enriching existing supplier data.
- Live demo: a paddle racket manufacturer search, where Speya performs quick market research and produces a discovery plan that the user can refine via chat.
- Speya adds enrichment fields (e.g., certifications, minimum order quantity) to the supplier spreadsheet, with suggestions for which data points to add.
- Filters are applied for location, such as suppliers in China that can export to Europe or Sweden.
- The presenters emphasize that the platform grounds the language model in verified database context to limit hallucination.
- Results can be exported to Excel, and the tool is positioned as a way to automate tedious sourcing work so people can focus on higher-value decisions.
- Q&A touches on how enrichment fields are suggested and how the creativity of prompts determines the range of supplier types that can be found.
Chapters
- 00:00Welcome and webinar introduction
- 00:54Meet the Speya engineers and background
- 04:13Agenda and how Speya works
- 06:24Live demo: starting a supplier search
- 12:27Building the discovery plan and refining it
- 21:40Results, enrichment fields and filtering
- 32:20Exporting lists and workflow automation
- 43:20Q&A and closing remarks
Full Transcript — Download SRT & Markdown
Speaker A
Now, I have two uh of my colleagues with me. Uh they are the ones who actually created Spea.
Speaker A
So, we have a lot of uh we have a lot of brainpower in the room right now. Um Matias and Morten, uh please go ahead and introduce yourselves.
Speaker A
Yeah, thank you very much, Rebecca. Uh yeah, Like I said, my name is Matias. Uh I'm an AI engineer at Find My Factory.
Speaker A
I've been with Find My Factory for the last 1 and 1/2 year. So, really when it's been starting to pick up speed.
Speaker A
Uh during this time where it's it's been a really exciting journey. And we're continuing to just build and revolutionize the way we see sourcing.
Speaker A
Uh and I hope that we can that we can share it with you today and maybe answer some technical questions. If you want to get down into the technical, maybe some quick tips and tricks about how we think
Speaker A
you should use the tool. Uh maybe learn some new things and just help you out.
Speaker A
And my name is Morten. I work also as an AI engineer at Find My Factory. And I'm super excited for this today to discuss about Spea and how you can use it in a in a good way in order to find
Speaker A
the right uh suppliers for you. You guys were able to actually roll out Spea pretty quickly. So, I guess before we jump into the tool and show you guys how it works, um it'd be great to get some background. So, what was the idea
Speaker A
that sparked Spea? How long did this take? What was uh the story behind it?
Speaker A
I'm going to go ahead. Yeah, I'm sure. Uh yeah, there's uh we saw that uh I mean, a lot of people are are used to working with AI in a chat format. Uh and what Spea is is that chat format but
Speaker A
within Find My Factory. Uh Find My Factory is a very powerful tool. It can do a lot of things. Uh the searching and the researching.
Speaker A
It can give you suppliers and give you information about almost any data point that's publicly available, even a bit more. But we saw that you know, sometimes it's this can be difficult to prompt an AI and manage that yourself. So, what Speaia can do is
Speaker A
that it it works as a an AI on top of the AI that we have and helps you like writing the instructions, the prompts, give you feedback, and really really taking you through the journey of one of your sourcing cases often right from the
Speaker A
start to a ready list with Speaia taking exports and do what you need to with them.
Speaker A
Yeah, I I think it I would say also that is a extremely powerful tool because Final Factory doesn't provide the data, but just the interface to the data that we can also provide with the Speaia in order to
Speaker A
get find the right suppliers for you. Uh and that creates a really powerful tool in order to um yeah, for you to find the right suppliers. Yeah. Well, we've already gotten a lot of really fantastic feedback from the customers who have
Speaker A
seen it. So, thank you guys so much. Uh especially for myself, I'm using a lot of tools day-to-day, of course, showing customers how to make sourcing autonomous or make sourcing and procurement autonomous. I think Final Factory has achieved in doing this, but
Speaker A
Speaia just allows actually the tool, I think, to be a lot more user-friendly. It's easier to use, it has this chatbot that can prompt, you know, simultaneous multiple queries at once, and it's quite intuitive. So, that's a really good
Speaker A
thing. Um so, for today's agenda, we are just going to go ahead and jump directly into the tool. I think that's going to be the best way for you guys to see it.
Speaker A
Um and of course, during this uh my uh Mathias and Morten will tell us the back end, what's going on technically. Anything else you guys think we should add to the agenda? What do you think? No, I mean uh look forward
Speaker A
to just showing you technically. I I I hope we can get the big picture. I know, you know, there's a lot of people using AI tools, but it can be difficult to understand the differences between them, what is actually the language model,
Speaker A
what is the things that are built on top of the language model. Uh just take you through with how you can utilize the latest AI. And what's making Spae so useful there?
Speaker A
In the whole Yeah, and some tips and tricks as well. And how we ambition it, I hope to at least to share. Yeah.
Speaker A
Okay, well, in that case, let's go ahead and jump in. Uh, we'll be using my computer here, so you guys feel free to also uh, grab the computer if you need to. Let's go ahead and share my screen.
Speaker A
For most of you who have seen Find My Factory, now we're sitting on, of course, the initial dashboard. And to start using Spae, you go ahead and jump to the source and then enrich. Let's go ahead and get started.
Speaker A
So, now that we're on the prompting page, uh, we're actually able to prompt the source or prompt a supply discovery with a text, a image, or even a file.
Speaker A
And this is a great way you can actually apply Spae not only to supplier discovery, but you can also ask it to do additional data enrichment to let's say existing suppliers.
Speaker A
Anything else to add? No, this initial stage is fine. I mean, we want this to be I think simple the way to start. Uh, we used to do the the chat interface, and I think just getting straight in here too and just write what
Speaker A
you want to find, and then you figure out what Spae is your assistant. And then if you have a blueprint, if you have some specifications from a colleague or something, just drop it.
Speaker A
See what happens. Yeah. Well, since I have the uh, creators of Spae with me, let's go ahead and do a search. What does uh, I'll let you guys take the wheel. What do you think we should look to source?
Speaker A
Okay we're um, I'm thinking about a paddle lifting. Okay, I know that you need an Iraq because you keep you keep borrowing it and all that. I'm so Okay, so okay, but I I was thinking about doing like a
Speaker A
drone or something. Yeah. I have it. So, the the paddle rack turn. Yeah. So, it's so the here what we record right now is just writing into the li- the language model interface. So, I mean, you don't have to be super
Speaker A
specific, uh, as long as the semantics are correct. So, I think manufacturer paddle rack it is good. Yeah. I I think I want them cheap as well.
Speaker A
So, where do you get them cheap? I guess The most try naturally. You got to really share.
Speaker A
I was um Let's say base in China. So, our manufacturers of paddle racket based in China.
Speaker A
Go ahead and start with that the source. Yeah, so I guess to take you through here what's happening.
Speaker A
We before we actually delve into the data and suppliers, we want to to understand the the area better. So, what does it do a quick little mini research? It looks up the paddle racket market, maybe some manufacturing, just understanding what
Speaker A
are the players here and what kind of search terms or manufacturing techniques they use in order to find them better.
Speaker A
So, I think maybe it produced a plan now. Let's see what it came up with.
Speaker A
So, yeah. Good paddle racket manufacturer well established with carbon fiber frames and EVA foam cores being the standard. I'll set up a broad discovery plan targeting Chinese OEM or ODM manufacturers.
Speaker A
So, Spaya did some initial initial the mini research here and it produced a plan for us.
Speaker A
And this plan you can see that it's a query on top paddle racket manufacturer and then there's the requirements.
Speaker A
You can here work with this in case you want to if you want to specify it more, you can talk to Spaya where you can manually I did like for example manually or you can write another message in the bottom. And this way you
Speaker A
can continue like to working on this discovery plan. Sometimes maybe I know a lot we have our our customers are professionals, so they know what to find. Maybe they are very quick in starting the plan. But if there's a case
Speaker A
where you're not really sure about what you're looking to find, you can talk to Spaya, chat with it, really crunch down this plan to something that's really accurate and help you become the actual expert of it. Yeah. So, try to
Speaker A
be able to ask the Spira in order to help you with the different requirements.
Speaker A
You know, actually, I agree because we work with some large customers as well who might have, let's say, sourcing managers or team members who specialize in sourcing, let's say, in India, um you know, Eastern Europe, for example, uh
Speaker A
and they're constantly getting projects to do. And even for myself, if I'm on a demo or something and I'm not really sure what I'm looking for or what they're looking for or I'm not sure what they're looking to source. Uh so, it's
Speaker A
actually quite intuitive that I mean, it can break down, let's say, how is a paddle racket manufactured? Yeah, yeah.
Speaker A
And then you kind of understand maybe what you're looking for a bit more. Yeah, do my research. I mean, I think paddle rackets Maybe it's quite easy to find a you know, if you if you search for paddle rackets, I mean, they have a
Speaker A
search in our databases and I I would get everything. Uh the search term is not difficult, paddle racket. But sometimes if you're maybe working with a fast mirror or you're working with something new that has a very unique product specification,
Speaker A
despite they could be a bit difficult to find. Uh and in that case, uh the you can work with the plan, uh with the target suppliers. I mean, uh manufacturers that work with the right techniques when they do some welding,
Speaker A
they do all of these techniques that required to produce certain part. So, you can really really work with the search plan to narrow down to something that you believe will find what you're targeting.
Speaker A
Okay, well, let's go ahead and start the search then and see what we start to get.
Speaker A
Uh something I also really like about um Find My Supplies as well is that we are uh working with so much data. Uh we're currently using 250-plus data sources.
Speaker A
So, as Spira runs and as it does this evaluation for you, any of the suppliers that popped up in this pre-vetted supplier short list, they're always going to come with a specific source.
Speaker A
That way we can at least ensure that the data accuracy is I wouldn't say 100% but as high as I I it can possibly be. It's wrong with in something. Yeah, it was that important.
Speaker A
Yeah. Um, I remember the other day we were having this conversation and you mentioned something about contact minute. Yes. And how Speia differentiates from other tools on the market.
Speaker A
It is quite obvious that we are seeing something like this. Um You want to go I can go with them.
Speaker A
Yeah, so when you're working with the language models, there's there's this concept called context and the context is basically everything that you've talked about with the agent before.
Speaker A
This is what the AI uses to produce its output and produce the results and similarly.
Speaker A
A question we often get is how do we distinguish ourselves between like other chatbots. I know Claude is becoming extremely powerful. A lot of our people we talked to are using Claude for a lot of things.
Speaker A
But uh these chatbots are kind of an engine, but they're not really they don't have the tools and they're not built for the way that you can use them in a workspace kind of way that we're doing.
Speaker A
Um, so with my my factory we we really built the strong scaffolding around a powerful language model. We make sure that the context contains only truth, so data from my factory.
Speaker A
Uh, it's always live in the in the workspace. I mean, you you see the spreadsheet. The spreadsheet is basically the context.
Speaker A
The AI has access to everything you have access to and it can read, it can understand and really work with this.
Speaker A
Um So, yeah, I think that's the answer. Anything to remember about context or Yeah, yeah, absolutely. Um, can I ask as well, uh, maybe a silly question, but I think maybe something a couple users might be interested in is how I guess do
Speaker A
you orchestrate Speia or make sure Speia is taking the correct paths and essentially is thinking like a somebody in sourcing and procurement. How do we know that it's following those intuitive intents, I guess, and has the right intention. It's probably is way to frame
Speaker A
it. Yeah. And then we we we work a lot with this given designing spell. And it comes down to a lot of testing. A lot of, you know, we we design it then we make sure that it it would
Speaker A
it behaves within the government sphere. But we test this rigorously. So all against all of We have a lot of different search categories, and our requirement, and we really go through them to beginning to the end and make
Speaker A
sure that okay, it's generating what we expect it to generate. Well, as of right now, we are sitting at 40 49 sub suppliers so far. Uh for the purpose of this webinar, I'm actually going to pause the search. But what will
Speaker A
happen, and you guys know this as well, is that the tool is going to continuously work until you ask it to stop. Uh there's So there's really no shortage of the amount of suppliers that you can get.
Speaker A
Usually for our users, I kind of say, "Okay, at the initial search, let's cast a wide net. So let's not get too into the specifics, because then of course, maybe some of that that specific information it's not accessible on any
Speaker A
of those 250 data sources. Doesn't mean that they can't do it, it just means that either our public or private data doesn't have a clear source. So again, as you can see on the screen, every company that we give you and every match
Speaker A
is going to be validated by a clear source. Again, reaffirming the fact that we only want to be providing you with valid and accurate data.
Speaker A
So let's go ahead and pause this right now. Currently at 52. Looks like there are a couple coming in at the end of the uh the chat. I think I think one one thing that I I like to explain because it's
Speaker A
really interesting how this how this spreadsheet gets created. Because um um it's a lot of data processing is happening right now. We're We're I mean, we're working with our enormous database uh to get these matches done. Uh and
Speaker A
then if you click on any match if you if you click on, you see the the reasoning and you see the sources. Right now it says manufacturer direct, quite easy to verify.
Speaker A
But you see the the the the the requirement and you see this course. And what we're doing is that we're we're leveraging AI on a scale that for a choir that I'm I'm not sure if other people are doing in the same way.
Speaker A
You can think of every every supplier here, we are we are creating a research agent and we're just grinding through our own data. We're identifying possible candidates all of these companies and then we for every all of them we have a
Speaker A
research agent that really goes into the depth and in our own data tries to extract and match this.
Speaker A
So, you can think of Spea controlling hundreds of other research agents, which I think is really cool.
Speaker A
AI you're the controller of one AI which then controls and orchestrates hundreds of other AIs. Spawns multiple of the different AIs in order to to direct.
Speaker A
Wow, it's almost shocking that the way that this was done previously was manual over doing. I know it's not a new thing.
Speaker A
It's Yeah, it's I think it's it's really impressive impressive and I'm really happy with that. You know, it's something that I'm I generally excited for. It's something that we I really would just want to continue building more of
Speaker A
these stuff like that. Yeah, it's like previously was wasn't just going on Google, searching around, looking into that way, but yeah, now with this tool Yeah, like imagine I'm going on the second page of Google.
Speaker A
Do you want the second I've Is there a second page? I don't know. I don't know. I don't know.
Speaker A
You know, they they do say that as well, but I guess the SEO engine is is only probably like going to move over to the GEO. So, I want to bring it back though to to Spea for a moment because I really
Speaker A
like what it gives us right now. So, we have this list of pre-vetted suppliers and immediately we have the reasoning behind Spea here.
Speaker A
And then we have additional enrichments that we can actually add to these suppliers. Can I ask how How are these additional enrichments being suggested to us? How does Spenel How does Spenel know this information, I guess?
Speaker A
Well, we're we're trying to when we as Spenel earlier, we're trying when we design Spenel, it was supposed to become the experts, be one step ahead when you're actually doing the search.
Speaker A
And as you can see in the in the chapter, it's starting in searching and becoming this expert for you in order to guide you on what what way or path you should take in order to find the right
Speaker A
suppliers for you. I I would say that's the the most powerful thing about it is that it's Yeah, I mean it maybe I mean it's like you saw we had some more to say some suggestions here.
Speaker A
I can continue here. I mean if I was for example I I I the more context I get getting back to the context problem, I give to the AI, the better decisions it can make.
Speaker A
So for example, one thing that I that I another advice I'd like to give when working with Spenel is to really explain your case, explain who you are, why are you sourcing or procuring this thing, uh where you're located. If you give it all
Speaker A
of this context, it will generate rules about compli- enrichments about compliance, uh about possibly important certificates to discover and all of this. So yes, we have the data and with Spenel you can access it easily.
Speaker A
Oh. Also really boundary of Yeah, I we I hope so. I mean but it depends up to the our customers to decide here. Yeah. I really believe so. I mean really of course do like the summary as well
Speaker A
because like I said for some of these larger uh companies that do have, you know, suppliers based everywhere including their own I guess boots on the ground approach, um this is a way that I guess you can import or take on hundreds of
Speaker A
projects. You don't maybe need the exact expertise for what you're looking. No, exactly. That's a really good point.
Speaker A
That's a really good point. Yeah, I mean because we I mean some people are very uh professional within their own sourcing categories. But as you say, sometimes there comes up a case that you never worked with before, and you're
Speaker A
still expected to find a supplier in that so. I would say that Find My Factory makes it really easy for yourself. Yeah.
Speaker A
So, uh should we go ahead just saying another uh micro photo here, like you. Yeah. He gives you the to the back over here. Um okay, should we go ahead and uh add on this additional enrichment? What do you think? Sure.
Speaker A
Yeah, sure. Yeah. All right. So, based on these suggestions, we have bought then. Okay, OEE. Yeah, I'm having fam. Minimum order quantity. This is like Yeah, it's it's good. Could be difficult for uh for uh Chinese, but yeah. But let's go with all
Speaker A
of them. See what we can get. Yeah, and I also want to mention as well, too, um so much of Find My Factory is understanding how to prompt. Mhm. So, let's say as well, we're giving Spae such a
Speaker A
I guess yes or no task. Yeah. So, let's see what we find in, but Spae doesn't just give you what we're looking at right now. It's not just a match. It can also give you much more extensive information. Let's say you
Speaker A
want a list of all Certificate numbers. That's right. Uh that these suppliers hold. Yeah. And I I think Yeah, exactly.
Speaker A
I just want to highlight right here, for example, if you see on the screen, right? So, with the deployment of um of managing all extreme and research, you see that Spae writes writing the the research columns. Previously, you did it
Speaker A
yourself by going into the research and clicking and writing yourself. With Spae now, it's it's writing this for you.
Speaker A
It's really cool, because we've seen some really complex uh instructions being written by Spae for for agents to use about compliance, about maybe you have a within your own compliance team, you have many rules that you have to verify. Then doing a
Speaker A
step-by-step instruction. I just imagine that a human doing all of those steps for every supplier. Yeah. I Yeah, I mean, the the time saved is Yeah, it's I think I see it. No, um really proud of that.
Speaker A
Yeah, I think we have customers as well where it's so costly. Let's say you are going through all these suppliers. You found a batch of suppliers that are fantastic. I guess I guess for what you've done this far. Then you enter the
Speaker A
RFI back and forth. And then you realize, okay, this fantastic supplier, great pricing, however, they don't hold the certificate we need. Now, what does that mean? You spent so much time on these one suppliers. You have delays. You're
Speaker A
not going to market as quicker. Um and of course that's going to be costly and timely. So I really do think I think the big value case that we have here, of course, is saving time, data accuracy, and going to market much quicker.
Speaker A
Uh which is as we know is such a competitive advantage for our customers. I think we have some results. Yeah.
Speaker A
Okay, let's see what we've got so far. I mean, we had OEM ODM as the search requirements. So I'm happy that we see that all of them are being verified.
Speaker A
I see certifications. I want to find a good one for my my own private rockets.
Speaker A
Yeah, let's see here. I guess this Yeah. They have a lot of certificates. I don't know what this certification means.
Speaker A
Wait, I know I saw as well. What I saw actually says it's an army one. I've heard of this one as well.
Speaker A
So yeah, as you see they're coming in. And then also core materials, I actually didn't I'm not sure either which one I Maybe I should have targeted a distributor rather than a manufacturer.
Speaker A
But yeah, you see the the data is coming. I think the thing to really point out in this is the the interface is just natural language.
Speaker A
One way that I like to abstract from my factory to understand really what we're building that you have three layers. When you build, you have the data. And we continuously work with data through web scraping, through partnerships, through just buying it
Speaker A
simply. So this is the first bottom layer of our my factory. On top of this, we have our agent infrastructure.
Speaker A
Uh and this is the the one that goes through all of our data to actually bring it forward. Yeah. And on top of this we have the the user.
Speaker A
Uh and the user is just communicating in uh in natural language with the agents and then the agents does everything else to interact with the the data and bring it forward to you.
Speaker A
Okay. I just want to uh take a note, you guys. I just uh hopped into the chat. Um please tell me if you guys can hear us clearly.
Speaker A
Um I'll go back into the chat and just uh give it give it a look. Um but we should be okay. We have another microphone uh with us right now, so Okay. That's uh We're working. All right. Yeah, let's go and see how it
Speaker A
does on the chat. But okay, I see uh see I see Marcin you're typing.
Speaker A
Yes, we can hear. Perfect. Thank you guys so much to you for the engagement.
Speaker A
Uh this is the this is the type of uh this is the type of audience we need.
Speaker A
They're good. Perfect. All right. Okay, so we got this following a little bit. I guess maybe I'll keep it here.
Speaker A
And then you see we Yeah, this one should be working as well. Um but okay, Marcin or uh Joachim, please keep us updated if you can't hear us.
Speaker A
But okay, back to Spire now. So, now that we have these suppliers generated, what's the next step or what would you suggest the next step to be?
Speaker A
Well, we could suggest tasks as Spire to suggest uh to move forward. I mean, we want to continue I mean, narrow down the list. Mhm. So, we could ask it to summarize. Okay. Maybe there is some additional data points that we should uh
Speaker A
we should find. Yeah. I mean, really I mean, I think I think for the following we let Spire take the lead. Yeah. And suggest what we should do. Okay, is it already you know, orchestrated to find all of these different uh suppliers for
Speaker A
the power rockets? Yeah. So, you just ask it again in order to find the right one.
Speaker A
So, maybe we can summarize it. Yeah. Okay. Summarize it. Yeah. Which one is the best one uh the best is a billion. A billion dollars. Uh an interesting one though.
Speaker A
Of course, like if you're if you're looking for new suppliers that we found them, we have the the enrichments. I guess the next step is to start sending out FRQs. For example, should we go ahead and ask for some email addresses?
Speaker A
Yeah, that's a good idea. If I find that to know for Not to write the part.
Speaker A
I guess Here you find the rackets. Uh okay. I like yours. So I might be able to find the yours.
Speaker A
For each supplier, I would say uh please include maybe sales um addresses or and phone numbers. Let's do that. Uh I You think we can Uh we will just find some Okay, let's see.
Speaker A
All right, I'm just going to jump right back into the chat and make sure.
Speaker A
Perfect. Not seeing anything. Most guys become friendly. about this a lot cuz I I use this I I use this shirts. I use this paddle rackets.
Speaker A
And I just wanted it was a better opportunity to better look for the new paddle rackets.
Speaker A
I mean that's kind of the cool thing about Find My Factory. Of course, we have like big customers using this on like an enterprise scale, but even for a single music paddle racket. I would use this. The supply intelligence part of
Speaker A
this is really really interesting. I mean sure you have to have a different faces. Yes, discovery generating good candidates and then the the intelligence and system intelligence part is just natural language.
Speaker A
Ask what you want to know and then I mean let's see if we can bring it up.
Speaker A
And I and I guarantee you if you can find it manually on the web or in any public data, I'm I'm sure we find it too. So just it's about paradise just making this data in a paradise. Yeah, okay.
Speaker A
paralyzed paralyzed the paralyzed too. And now can I ask you guys just while we wait for these um results to to come in?
Speaker A
Um is there any other I guess AI companies or any other company in general that's doing something similar to Spyai or Not that I know of. Okay, so we're pioneering this it seems. I mean yeah, I really think so. This this scale and
Speaker A
with our data it's yeah. Yeah, I would I'm so excited to see the scale of one of finding all these companies and having this nice verbal interface to anybody and don't look Yeah, I mean I think we we're seeing we're seeing
Speaker A
really a model come up in many industries with like the the workspace and the chat. Uh I think we we see we have a lot of startups as well.
Speaker A
I think Legora is one of them and I mean Labor Bull and what you do is that you you you you you you build this AI agent in the chat and then you have the workspace and then you
Speaker A
you make it so that the the AI can do everything within the workspace as you can do. And then you really you build the the tools and you make everything available for itself. I mean my vision is that we we we continue building this.
Speaker A
We have the the the the workspace right now is just a spreadsheet but it can be easy. Files it can be. I mean a lot of different things.
Speaker A
So uh Yeah. Okay, well that's good to hear for me to hear as well. I work in sales so that's a good question that I typically get quite often. Yeah. Um so that's a good one.
Speaker A
Okay, but now we have excuse me. Now we see those email addresses and phone numbers coming in. I mean we tend to find when it comes to sourcing and procurement all these feedback that I've gotten is yeah, finding accurate suppliers and then how
Speaker A
do I reach these suppliers? Yeah. That's when you go through LinkedIn or you have to go I don't know have a subscription for a ZoomInfo or or something like this. So it's actually great that we can potentially minimize somebody's tool
Speaker A
stack by implementing Speira. It can do so many different tasks and hold so much data which is amazing.
Speaker A
Um so okay, now we have a list of suppliers. We're at 55 right now. Uh of course you stopped it. Just wanted to mention that we could be up to maybe you know 200 but We didn't continue generating new search
Speaker A
plans. I mean I think we we searched for China, right? In this one. Yeah. Uh from here we could try another target area, compare the data, collect more. I mean it's I think creativity is it's just a limit you know with Speira.
Speaker A
Yeah, I was actually on the a call the other day um and the big question that we got was okay, how does this compare to what I found? Well, what you can also do as well is if you go ahead and just
Speaker A
hover to the top, you can see the import export. And this is where you can actually even add on your existing suppliers to see how they're going to compare to this new supplier list.
Speaker A
Yeah. And that's a great thing. Let's say you had this big batch of suppliers, maybe they're at risk of let's say tariff implications, maybe they're at the risk for delayed um delayed deliveries, whatever it may be, you can actually
Speaker A
alleviate that work with the right into Find My Factory and again has Bay View the work for you. Yeah. Or even find alternatives to those existing suppliers by asking Bay View to run that task for you. So that was a really interesting
Speaker A
use case um for them. That's cool. That's good to hear good to hear. So it's not just supplier discovery as well. I mean Bay View is kind of your AI orchestrator for really whatever it does. Intelligent so I can just um
Speaker A
I mean I really I really like the view we have and I we've got this it's just to just to just bring in more data, continue bringing more data and then you kind of and then you you simplify. I
Speaker A
mean I think we can do it right now. I mean just have Bay View simplify the list for us. Okay. Yeah. That's I think we just uh I'm doing gymnastics over here for fun. Um but okay, let's uh go ahead
Speaker A
and say how about say I'm based in Sweden. Yeah. Uh what are the top suppliers for me that can what you say import deliver? No, it's where export. Uh that can export to Europe.
Speaker A
I mean creativity here is the the limit. I mean if you want if there are other requirements that are more important to you, you just tell them. I mean I it's I mean pricing of course if if it's
Speaker A
public which is not often it it isn't always public. But for the maybe if some of them, you know, you can ask about get those data points as well and then MOQ. I think we have we had some data
Speaker A
from the MOQ previously. I saw it came in. So just what is important. Yeah, I saw that as well. Maybe we can go ahead and hover over that for a little bit.
Speaker A
So, okay, let's go ahead and add Yeah. So, now we're getting down to the other last data here before we can we can simplify if they export to Europe.
Speaker A
All right, and we even do have that minimum order quantity, which is actually Yeah, like you said, it's quite impressive. We're not always finding that companies are sharing this data with us. Yeah, let's see.
Speaker A
Here you see the the the I mean, if everything I just want to reiterate that every every data point that we get in is grounded in something. And you always have access to our sources. So, you see them here.
Speaker A
I mean, it's a a funny thing, too. If you contact, let's say, a supplier and they say, "No, we have a minimum of 1,000." Well, guess what? You actually don't.
Speaker A
But the size that we are the lab rats to do I haven't thought about that.
Speaker A
But I can actually import at a at a lower scale. So, new suppliers as well, even benchmarking. I think that's a It's good. interesting way. I think we've all, you know, tried to buy something and they say, "No, you need X amount."
Speaker A
And it's like, "Well, I've actually 52 items." familiar. Uh so, it's so it's a really great way to I think leverage I haven't thought about that. The data on the product side of this. And we have a couple ones here, so Yeah, maybe that's
Speaker A
one I should probably exit. Good night, Alan Turing. Yeah, yeah, we do, actually. Let's go ahead and see it.
Speaker A
I mean, they emailed us to this one. I learned but they don't they don't export to Europe, so Doesn't look like it.
Speaker A
Okay, well, now we have all these data points. So, should I just go ahead and ask for a tier one, tier two, tier threes? Who should we Who should we contact?
Speaker A
Um Give them a thank you. As Bay has been doing. It would work. It's really good. So, don't forget to send out the thank you. Um okay, uh give me the top suppliers, um top suppliers out of this short list.
Speaker A
I really like this as well, too. I mean, even if you have all this information, let's say you don't want to go through every single data point individually.
Speaker A
Yeah. You can see that visualization, you can see, "Great, I have all this data on these suppliers." Yeah. Um and then you can ask आँ of course to make this more intuitive choice for you. It's going to basically
Speaker A
evaluate the data on your behalf. Yeah. Um a great way as well, we have a couple customers using this, you know, like I said, at a team or enterprise level. And this is a great way as well, let's say
Speaker A
you're looking to boost a new project. Um maybe you're looking at some risk exposure for existing suppliers. You can also share this internally. So, you can actually reevaluate these suppliers if you have a teammate that has, let's say,
Speaker A
different um different parameters for different categories. You can use this kind of idiosyncrasy. No, no, we're definitely we're seeing that they take combining teams, I mean, from there there's so many different uh applications for for this. Yeah. I'm
Speaker A
really happy about that. That's good. Well, I think to uh to wrap up this Spe A demo, um Martin, I think you can take take back some of your paddles uh because we have found with TSOM good ones.
Speaker A
And cutting based on the based on this information, looks like Top Carbon Sports. Okay, sounds good. Yeah, I really like, of course, again, the the summary of it, the intuition of it. Did I'm not buying 20,000, right? No, that's
Speaker A
a production. I know. Yeah, good. Production paddles per month. Most of their clients are in the US and Europe. Okay. Good. Uh let's go into the website then. Do you have the I guess go in the supplier list. We do.
Speaker A
There they are. Top uh Top Sports. Top Carbon Sports. Top Carbon Sports. And I guess you got the link right there. Yeah.
Speaker A
Not sure if you guys can see that. I'm only You got the link and of course you have the good ones as well.
Speaker A
So, okay, now we have tier one, tier two, and tier three. And what we can do right now, just to kind of clean up Spe A one last time.
Speaker A
Thank you again. It's been fun to have And uh remove all remove all other suppliers.
Speaker A
Yeah. And now we've gone from that list of 55. we've either done those data enrichments and now we're trickling down to what was that? Just maybe six or six or nine.
Speaker A
So, what do you estimate is the uh typical time this process usually takes? Huh, I mean I mean, I can only give I mean, I haven't worked in procurement, but from what from what I've understood is that this could take a lot of time previously
Speaker A
and and the it's I think uh it's important to to really get into the AI tools and AI space because I think in so this is for procurement, but in many ways in your professional life, there's AI tools
Speaker A
waiting for you. Uh and they can help you save in time and energy in many areas that maybe you don't think of before. So, yeah, don't be be curious about AI.
Speaker A
Test and try things out. That's my advice at least. I want to say something as well. We actually got We got so much data in this initial search, but I've been in meetings where sometimes it says no data
Speaker A
found, no data found, no data found. What's actually so interesting is that again, you can ask Spae dive deeper.
Speaker A
Yeah. Dive deeper into this task. Why is there no data found? Find something for me.
Speaker A
And then we actually see this amazing thing on the back end where we can actually see it's prompting that search again. It's scouring different domains and it is going to find something for you. So, like Matias said, be curious,
Speaker A
you know, if you are testing the tool and you see no data found, don't let that stop you.
Speaker A
Ask Spae to work more for you. Give it additional tasks. Yeah, remember the layers. There is the the the data and the agent layer. And the agent layer is doing its best from your from what you tell it to do, but
Speaker A
Okay, and now it rained. We're moving, yeah. Uh 94 or 49, excuse me. Let me just reload and we should have the updated Aha, okay, that's the thing we have to fix.
Speaker A
Cool, and here we have the the top six ones. Top six, we have their emails and everything. So, I guess uh I will send this to you after this demo is up.
Speaker A
And they got And for example, too, like if you're a consultant as well working with Find My Factory, you can always export these lists as well. So, I mean, I know so many people in this industry they're still in Excels. No shame whatsoever.
Speaker A
Excel is is a good tool to use. And you can see that Find My Factory essentially behaves the same way as Excel. It has the columns, you know, it has the different rows, everything. So, if you need to share this internally or let's
Speaker A
say share this with a potential customer, you can export the data as well. What is next for Find My Factory?
Speaker A
Well, we're mostly looking to after you've gotten your suppliers and you have gotten the right one, what's what's the next step after that? That's So, we we've been looking to a lot of my Okay.
Speaker A
You want to also contact these suppliers. And how can we do that in the best way?
Speaker A
Well, we have Spire from the beginning. So, utilize Spire as much as possible in order to do that for you as well.
Speaker A
Which can just simplify the process even further than I just almost had to just hit one button in order to make everything work. Okay. Yeah, I mean, we have we have an engage module that This is driven This is human driven. This is
Speaker A
not AI driven. And I think that moving forward, we're already building it. So, I guess we can disclose We can disclose that.
Speaker A
And yeah, we we want to get that email and email kind of working into Spire as well, so that you will be able to Spire will recommend emails and you can send through one of our email addresses or
Speaker A
send from your own email address directly to the supplier cuz it's just about data.
Speaker A
And then Spedia just writes apps. Just writes the apps for every everyone. And then you do you click send. I mean, we we're not really sure how it would look like, but just continuing. I mean, if if there's data we cannot find
Speaker A
publicly or if there's data that it's not not available publicly, email them and let Spedia handle the emailing and the signing and you get a notification on your Spedia has found this data to list in Ireland.
Speaker A
Another data source. So, at some point too, I mean, sourcing new suppliers should take in total, let's say, an hour. Yeah. It says, let's say, weeks or months. I don't know. I really think I'm really excited for the
Speaker A
future. I just you can you can work on the more important things, the the interesting things, and delegate the manual and the tedious to agents for you. Yeah. I guess the biggest question is like for for everybody in this webinar is is okay.
Speaker A
What can you do now with all this time that you've saved? I mean, more projects can you run? How quickly can you go to market? How quickly can you expand, let's say, existing tangible data sets and ours? I mean, what do you do when
Speaker A
you get 50% of your time back potentially? Yeah, for sure. I mean, yeah. I mean, I I don't see AI replacing people. I'm skeptical with this narrative. I just make it AI allows you to do the the interesting things. Yeah.
Speaker A
The things where the the human context can can have every part of the on the entire project, the whole of it. And I I I don't With the I think we're very far away from that. Yeah. And you can with
Speaker A
Spedia and tools AI tools like that, you can focus on those parts. Yeah. Okay. Well, you guys, thank you so much.
Speaker A
We have a last 5 minutes here. So, what I'm going to do is I'm going to go ahead and move us to the mentee. So, let me share my screen.
Speaker A
I see one question in the chat. Marcin, that's you. Do we have a recorded version? We do. Everybody in this webinar, you'll get a recorded version after this call.
Speaker A
So, you can revisit revisit and if you have any questions don't hesitate reaching out to myself, Morten or Matias on LinkedIn. Okay, I do I do open up the budget free raffle.
Speaker A
Okay. Perfect. I will go ahead and share my screen really quickly just so we can present the um the Menti. Let me make sure I have this.
Speaker A
And we have four questions in already. So we're going to go down the line. Um Morten, how would you answer? Somebody asked, "Can't I just use ChatGPT?" It's a good question. It feels like a question that you Yeah, is it handy?
Speaker A
Yeah, yeah, yeah, it's good. It's good. I don't know. And it's good that you um I would say the the most important part is that when working for example with ChatGPT, it's that you don't As Matias said previously is that
Speaker A
the context around it is extremely important and just starting a chat from with ChatGPT for example, you just get that far. But you don't get to the next step.
Speaker A
And it's working for example with Speira compared to ChatGPT, doesn't really have the correct tools in order to find those suppliers that you're actually looking for. Like ChatGPT may show you something, but is that the one that you actually
Speaker A
want in the end? Yeah, I think I think yeah, there's there's many part. I think context is a very important argument cuz what what we we built is the scaffolding, the spreadsheet. Uh in the chat there is just text. I mean
Speaker A
you can extract that into Excel and you know, previous as well, but uh the the context of what you're working with is always in the spreadsheet. I can work with manually and so on. But then there's also the the
Speaker A
data and scalability point. Yes, you can you can have data in in ChatGPT too as well.
Speaker A
It's not I don't think I'm not I don't I'm not sure, but I don't believe they have current data. Uh I don't believe they have the scale of data as we do. So I mean that and as GPT so
Speaker A
it's prone to do that. So, I think there is very very big difference in ChatGPT. And I know that if you're working with compliance or things like this, you need to have correct data.
Speaker A
Yeah I know somebody who uses ChatGPT and Gemini. What I really like about Swaia is sometimes you're asking ChatGPT or Gemini to do something. And what happens next? Follow-up question. Next?
Speaker A
Follow-up question. Follow-up question. So, it's almost like I have to continuously guide these tools to make sure that it's not leading me astray. I mean, they're good great tools, don't get me wrong. I use I mean, my other employers daily for
Speaker A
tasks for the work that I do. But they're not built for that structured data. They work great as engines to power reasoning and everything. But the scale of which we're using language models for to for to certify and validate our data points
Speaker A
to do this in just in a chat in ChatGPT, it's not possible. No. So, I think yeah. I thought I hope that I'll just ask a question. Yeah, that's all the great question.
Speaker A
All right, answered. So, now we have a next supplier. Matthias, I'll let you lead this one then. Can you find indirect indirect suppliers like cleaning companies in X town or other services? Yes, so this is this is this is what kind of what's been
Speaker A
exciting me about it is that we started out, you know, looking as a Find My Factory, it's the name. About procurement and finding manufacturers.
Speaker A
But we've seen that we have almost everything in our data. So, we I mean, we have people looking for service companies, looking for distributors, I mean, manufacturers, but also indirect like I've seen cleaning companies as a query come up before. So, so yeah, yeah,
Speaker A
yeah, just try it out, I think. Test and you'll see that you will get them and I mean, the the the enrichment and the vetting stage of actually certifying the data points, it's similar from a manufacturing company to a cleaning company. Yeah, I think
Speaker A
just you mentioned earlier also of the creativity. That that's actually the true limit of this that you can find so many different types of things or the supplies or almost everything. You can just Yeah. Exactly.
Speaker A
Yeah, and I just uh All right, two more questions. Um does this get shared with my teammates?
Speaker A
I enter Uh well, I there the the the Yeah, so I can I guess we can go into this. Um I would say in terms of your any of your searches, I don't know. We are extremely We are extremely private.
Speaker A
Uh we do of course adhere to GDPR, SOC 2, ISO. Um so if you don't want this information shared, absolutely, it won't be. If not, I mean, yeah, the the the the the list I think if you want to talk about within
Speaker A
the the workspace, that your lists are shared with everyone in the team. Yeah. But the chats and your conversations with Spyai is not shared. So the way the the things you ask Spyai and the conversations you have with Spyai is not
Speaker A
shared with your team members, but the result of the spreadsheet is shared with everyone. Yeah. So.
Speaker A
Okay. Good uh And we decided to have that view because I think it makes sense, you know, you're working in your chat and you're Maybe you want to keep up you know, what you're writing and asking everything.
Speaker A
Speaking on each other's conversations might not be super fun, I don't think. Yeah. But if you are um of course looking for a team, um find my back to you for your team. Uh Spyai, soon to be Spyai, uh for your team, uh reach out to
Speaker A
us and of course we can kind of go into this a bit more. Yeah.
Speaker A
All right, answer two more questions. What's coming next for Spyai? Yeah, we teased already. We We We did image. I mean, uh I guess that's for another Morten has been talking about another idea that really uh that something but
Speaker A
about artifacts. Yeah, it's a bit of a But the baby do you want to tease with that for example.
Speaker A
Well, looking at the what's what you want like what's the next step as I previously mentioned with the emails what's the next step of it? Yeah, often that's you want to have a really nice way of exporting this to the the next step in
Speaker A
that case. And then you possibly have a lot of questions that you want to answer for example or you want to fill out a specific document based on what you actually found.
Speaker A
Well, we can solve that as well with Speia in order to utilize its full potential.
Speaker A
And that Yeah, why not let Speia fill in the compliance documents? Yeah, something like that. Yeah.
Speaker A
Not bad. All right, almost to the last question. So, what are the best prompts to use with Speia? Any tips?
Speaker A
Yeah, I think the case is really good. Coming back to context again. The more you describe about the case you're trying to solve the more information it gives Speia to help you. So, if you explain as I said your job title,
Speaker A
position, location, why why are you doing this? What's important for What is your end goal with this whole thing? All of this information makes Speia make more educated opinion.
Speaker A
educated decision I would say instead of having to guess the final parts. If you give it to it, it make the decisions.
Speaker A
Also that it helps it even more stay one step ahead so that it can help me in a third way because the more you explain about your problem the easier it will be able to help you with that essential
Speaker A
search. I want to thank you so much for joining us for this session. And again to Mathias and Morten. If you have any questions or want to see Speia live for you and your team, head to findmyfactory.eu and you can actually go ahead and book a
Speaker A
call with us directly. That will be with myself not Mathias and more than that, if necessary, I can bring them in.
Speaker A
I'm open to I can take you always. That's good. But, um stay tuned as well, you guys. We will have another webinar coming up, and I believe this next webinar will likely be probably more than again going into
Speaker A
May instead of that else. Into that new email release. Same my new other actors well. Exactly. I'd like to see that.
Speaker A
Um but, thank you so much, everybody, and we will see you next time. Yep. Bye, everyone. You were good.
Topics:SpeyaFind My Factorysourcingprocurementsupplier discoveryAI assistantsupplier enrichmentwebinarpaddle racketB2B sourcing
