Session Information:

Speeding up the proposal phase in project business by 300% using an intelligent tender document screening approach

Identifying risk extremely early in the project business processes was the key target to develop the intelligent tender document screening approach.

Currently the systematics are used in a broader and more holistic way to enhance proposal processes and get much more speed while reducing reading time of large or many documents

• Intelligent tender document screening approach
• Most interesting use cases and benefits
• How to implement it
• Results and outcomes of different usages and scenarios

Session Transcript:

I'm very excited about our next speaker, Oliver. Macko is the partner and vice-president for, a bit, for project businesses, as Siemens Advanta. He has more than 20 years of experience seamless. All in solution and project businesses today, Oliver is going to showcase for us one project where they sped up proposal phase is within the solution businesses by more than 300% using digitization solutions. So Oliver, wonderful to have you here, look forward to your presentation.

Thanks, Jose.

Thanks for giving me the chance to showcase these kinds of things.

So, For a moment.

No way to pick here. So, thanks for giving me the chance to showcase how we speeded up proposal phase in project business. So, maybe a little bit of introduction in that moment.

I'm working for ..., is the business unit, maybe, I don't want to say new kid on the block, but the new. It's related to IOT IOT services, but also any area of digitization, and digitalization of processes, and also internal processes.

My personal relationship is, our responsibility is, especially project business. So, from project sales face down the road up to project execution and project delivery. And I'm happy that I'm gonna be able to showcase how we speeded up proposal faces or ability to say request for proposals. And how we handle these kind of things by 300% using our own solution, which is called intelligent document screening.

And they will introduce today's approach to the next half an hour.

So, I don't want to be boring, but some rough numbers, as you see these kinds of things, 385 employees around the motor, but I'm pretty sure that when people, let's say in the room, will know what Siemens is doing, a glimpse of that is right, solid here, started from the right side to the left side. Yes, of course, we built these ... computer monographs and these kinds of things, for four hits me a business or health care business. But the rest is still related to projects and solutions. So, starting with Siemens Gamesa, we have Wind Power Wind Parks offshore, or onshore in the mobility area, We talk a lot about traffic. Solutions. White, rolling stock, high speed rail, metro rail systems. Also, tram train systems and just kind of things.

And coming more on the left side, we talk about buildings, bidding solutions, project and the bidding area, or maybe campuses for complex of buildings and the bidding modeling systems. And on the left side, really, how do we build, let's say, power plants and big project in that area? Some of you may have heard about the edge a project we did, where we build up three power plants in the same time. So based on that, around about 50% of our turn over is based on project business and solution business.

Saying that we really started doing projects for customers really early in the 19th century. So the first really make or maybe Giga project both the telecommunication line from London to Kolkata.

And throughout all these years really do projects in all the areas I mentioned earlier like High Speed, Mitchell's. You'll see some pictures here.

To give you a rough number, based on that, we really start roundabout.

Let us say, five projects per Workday.

So, around about that, this is the amount of project we start, and I said earlier around about 50%, is based on project film from the tone of our revenue.

Saying that, please assume, uh, dead, therefore, you will get a lot of request for proposals, ..., and this kind of thing, and, therefore, there's a huge amount of paper to be handled during the sales phase and proposals.

Just to came up with a proposal to our clients at the right moment in time, and they should also be, let's say, less risky, having a good cause position, and oil just kind of things, so project business as a place in our DNA.

Saying that are we talked about, or we thought about how? how should we handle this huge amount of paper?

So we have business units doing round about one thousand eyepiece and proposals per year.

And the other side is saying that you're doing a big metro system, Let us say, for the London Underground, the complete RFQ, RFP is around about one thousand or 10000 pages.

And you have to handle all these kind of things based on that. They are specific things in terms and conditions. We have laws and standards, and a lot of requirements that have to be handled.

And interesting questions in that moment really early in the proposal phase is, are there any social pus requested? Let us say, that the customers have a request we cannot handle.

So there could be some kind of unlimited liabilities. This could also be some technical issues. Like for power plants, we have to fulfill some kind of specific NO requirements. Maybe we can.

Maybe we cannot, or maybe also some kind of specific test cases. And the test case is documented for, let's say, our gas turbine or maybe compressor casings is on the page 999, maybe, in a small footnote. And maybe they have some problems if we don't see that.

Are additional. What's interesting for us right now are specific requirements. Like, maybe also you have it like terms on phrases like buy America.

If we, from a European standpoint, don't see this kind of things by America. Then maybe we have a lot of problems, because maybe we would buy our components in the wrong place. So finding this kind of SawStop has a huge documents is difficult. And that's why we have this kind of machine I will explain later.

So whenever you have an additional topics like know we want to try to prevent non conformance costs. Therefore, maybe we have some kind of crisis phrases, poor quality for safety, function of safety.

We have to find and maybe and this is also an interesting case for us.

Maybe some kind of request for proposal is let's say, optimized for specific competitor and maybe not really optimized for us to make the somber quite clear. We try to build some kind of a machine later on. I will show where we have specific GE voting.

So, general electric voting, anti corresponding Siemens voting and wouldn't it be good to have suggest kind of balanced situation? If you see more G E voting it, if P then Siemens voting, we will not go for the bit. So just what the storyline this, what, our expectation to our machine, we try to bid and the build it.

So we've been able to screen large amount of text elements, phrases, 11,000 pages, really easy. And most of them in the ISPs, are let us say, handled in PDF documents, so we take it ....

Additionally, we try to optimize proposal phases. So we, we will not only with use or maybe having technique, choose to identify the right phrase in this kind of documents. We just want to make the handling of the situation later on much easier.

So besides this, some of these ISPs let's say, consists of more than, let's say 20, 25 documents itself.

So based on the edge, you'll need some kind of system to navigate to all your findings.

And then it should also be integrated. In the tooling landscape we have.

So, because we don't want to have only the egg rosenfield. Just wanted to have some kind of systems. Now, there will be able to integrate it in the whole process.

Right now, we implemented the system for the last, I would say, five years with some kind of easy web access.

So based on that, you will be able to upload your PDF and screening documents on your SIQ.

You'll save a server based on data DD. The whole amount of ISPs have to be secured.

Then the machine is working.

Additionally, unable to explain it on the next pages, you need some kind of search criteria.

Maybe additionally, also some kind of OCR screening based on that, that will not get it in a PDF format.

Maybe you'll get it on the on the on the scan Format, and you'd have to do the OCR ing before You need this kind of expert knowledge. So, is it a terms unconditional related? Is it more or less functional safety related?

What does so stop us, What a G E voting.

So when we combine these, I would call it digitized or digitalized expert knowledge, in something we would call dictionaries.

And based on this dictionaries, the complete PDF is screened against this. And does not only, I have it later A, control F So trace also. We will be able to search for let us say if you have some kind of technical topic like reliability or maybe availability has to reach 95% and we will also be able to search for availability and maybe a percentage even indicates that. We don't know the percentage. So this is also implemented. So, we have some kind of rules.

Not only to go for phrases, also for specific.

... in one sentence, or maybe, which are connected in one sentence, and are not connected in one sentence, or maybe also on one page, and just kind of things.

You will be able to download your document and really stressful site search for that search results, and you will be able to work on that.

So, and just to give you a glimpse how this could look like, as you can see here as well, let's say glimpse the format.

So, on the right side, you have your original 10 document, maybe more than nine pages, let us say one thousand, and you will get the search results as these kind of mock ups. And so, if you are the functional safety experts, or maybe you are the one with the light blue elements and light blue elements are really interesting for you. And then you work on that.

But to interesting effect on that, is, that if you see the bookmarks, so maybe it is only interesting that you will have specific phrases for functional safety on page one, page two, page five, Page eight, and Pence debates nine.

Based on that, and to assume your document has an amount of more than one thousand pages, you will really guided to these kind of pages where you find your relevant content.

If you are the Functional Safety Expert Quality Expert at TAC, the Legal, whatever you kind of expert, while you have, for just kind of proposal, and based on that, you will get much more speed, because you will be guided through your specific content that you need, to your failure to your task on the proposal generation.

Additionally, you will get a hit index.

So if this kind of document really contains your content or not, And audio will get an idea on the pages, maybe, if to kind of page is cleaned or scan before. So maybe you will not get them results, but you need to information that you will not get a result.

So based on that, as human, you have more than one thousand pages, you will be able to read only these kind of topics you really just want to focus on.

And this is also based on the sentences you see into avoiding in a dictionary.

So maybe that in a function of the safety chapter, you will find a quality requirement, or maybe in a Validity chapter, you will find a functional safety requirements. And based on that, maybe OCW Crime is not guided to rewrite expert based on this.

All the pages were marked, where you'll find you interesting things if you have to roll like a lawyer, technical expert discount, this area, or that area, and may have additional area.

Saying that, uh, also, we will close the loop, just if you, think about that, you have people from former projects, or you have people with really experienced ones.

Who are, let's say, close to retirement or something like that, maybe take data specific domain knowledge, and also places in a dictionary.

Just to enlarge our search criteria and also to work with their experiences.

So people close to retirement, really the experts on that, people, let us say, hoo, hoo, If you would like to address the beauty, you are, let us say, knowledge, cabals, let us say countries. So what we do, sometimes, in the, let us say, steam turbine business, we will do proposals in Brazil. We will do proposals in the US. People do proposals for a steam turbine business also in Germany and in India.

And therefore maybe it's also interesting to spread the knowledge. What does a showstopper? What kind of thing is a really interesting criteria, which may produce huge non conformance cost at the end? It's also placed in a dictionary. And based on that, we will be able to spread the knowledge across the world for these specific proposal departments, and also, this is the case, why we will be able to speed up, wherever we use release, we reduce claims, and all those kinds of things I mentioned earlier.

So we digitize the expert knowledge, present a dictionary with, As I said earlier, the three main let us say, cases are, Do we see interesting show stoppers?

In our risks, in these kinds of proposals, all requests for proposals for that request, Volvo poets we receive. So maybe higher risk topics, things we don't want to fulfill. Things, we have any deviations and this kind of thing. So, this is the ah let us say the main use case on the left side.

one of the interesting use cases for our sales department and it's the one in the middle, which I would call competitor voting.

So, if you have some kind of specific competitor voting and just let me explain the one I mentioned earlier with the General Electric case.

It is not just averting as is more or less how you handle numbers sewed the range of gas turbines.

Normally, will deviate of T of the power output of gas turbines will deviate a little bit from from company to company.

So, if if G has let's say 400 maybe we have 380 and if you have a customer looking for 400 to 200 to 420. So, maybe you see OK. it is more or less to them. If you have a customer looking for more or less, 362 or 380, I don't matter was an unbiased, you know, a little bit more its focus to you. And if you have a huge amount of just kind of elements, you see that it may be a little bit imbalanced, and they may not go for that.

Addition, I also mentioned Indias huge documents.

This is the third topic.

If you have a huge ISP, was one thousand pages, or something like that?

one key element is that you have to slice it in smaller pieces, and chapters, and disk kinds of things, and handle it across your organization.

Just the right expert will get the piece of the DD part of the document he needs.

And based on this, he will receive really all requirements.

He really needs not only the chapter by because maybe the customer also mentioned that a specific requirement and the wrong chapter unsold E T, The apportionment of requirements also important, and taking all these things and together.

We're built up one complete use case.

So assume you have to do a complete business, was roundabout five funded proposals per year. Receiving, let's say, one to one thousand requests for proposals, also goes into here. Your boiler question is how you will be able to speed up this car, how to handle these proposals, you have to read everything.

So what we talked about is that maybe we have a specific dictionary for Ts and Cs.

There's specific dictionary for all technical aspects.

If you talk about steam turbines then maybe thermodynamic aspects quality experts, casing experts. Maybe also rotating equipment, let us say power output, electrical, automation and this kind of thing.

And then maybe specific project related topics like milestones, you need or maybe specific resources you need or maybe also constraints.

As I said earlier, the example was by America and this kind of thing.

Taking all this together in specific dictionaries, let the machine work and mark and tech and draw all the texts you saw earlier.

Uh, in that case, all pages which are not marked because DDOT assumption is that you have complete set of dictionaries. All pages which are not marked.

Let us say irrelevance.

It does contain blah, blah because there's nobody compliant for the requirement for decency is not is not.

that not technically a not technically be requirement and, and, and, and based on that, you will be able to speed up yeo proposal phases.

Because all these blah, blah pages will not be read, read, read.

I'm sorry, on these proposals, maybe there's the risks.

Yes, of course it is, but you can handle that.

So, assuming this kind of scenario will be able to speed up and be measured for one year, the process before using the Intelligent tenure screening.

Drawing the dictionaries and then using it for more than one year, and then they came up with a 300% sodium reduction of proposal handling was going down from around about 45 days, down to below 20, up to 15.

This is the the example I say.

So, speeding up the proposal faces in the bid and proposal process, because we had long idle time, where we We wait for the additional experts. Everyone has to read everything.

Let us say hinders us to deliver the proposers may be, in the right point in time.

The we now, doing it with a large set of a complete dictionaries. I mentioned it earlier, how to, How to draw, not only one dictionary because we need a lot of, let us, say, aspects, Ts and Cs and these kinds of things.

And we will be able to speed up the proposal time from 30 down to 10 days, which will give out the proposition. To increase the amount of proposals on a worldwide scale, I have to say in a worldwide scale, from 700, to up to 1500, And so, if your hand and more proposals, At the end, you will do more projects.

So, this is the outcome of that.

So, well, what kind of thing is it useful?

So, the complete machine D approach, assessing PDF documents, we use it for proposal phases and proposal documents and these kind of request for proposal documents, especially, you can also use it for contracts later on if you have a specific contract, or maybe you would like to assess this kind of paper for specific. I'm not sure claims changes and whatever, You also can do that.

Additionally, what I have heard, but this is not my domain, you can also use it for let us say, P papers and reports in the I would call it marketing area or maybe not marketing, what is it called competitive intelligence. So if you have a let us say: the Dee, Dee, Dee.

What is it called? Sonic, the quarterly results from other companies, in order to the ..., onto pieces of paper, doing debt by their essay, or maybe also for patterns. So you can analyze, let us, say, data in PDF format, when other companies saying that and looking for specific boardings, especially if you have more than, let's say, 50,000 pages, and guess, if they ask the combination of debt.

But SSA jolyon, my domain, specifically, the one with the RFPs and just specifications.

Right now, what we have, because this type of system was running for more than five years on a secure server with and Siemens.

We have it on a worldwide basis. We started with, let us say, 100 users. And right now, we are more than 2000 users.

If they have just kind of a Batch, they can, they can, let us say access the system. Do the upload 40 PDF documents? And they will gather was either within, let's say, five minutes. If it's a huge document. And if you have more than 20 dictionaries maybe then you will need, let's say, 20 minutes. Not only NaN, but also spent a system which will be able to read 20 the document within 20 minutes.

Maybe it's also Interesting one.

So, on the right side, we receive some kind of feedback on Dad.

And, based on the technology which is used.

So, because we really assess the let us say. The I would call it the ASCII code of the let us say, let us, we assess and phrases. Yes, of course, you will be able also to assess different languages. so easy would be let's say Norwegian and Spanish and friends friends.

Because we we started with English, and we also can do it in German, but based on the situation that you will need the dictionary language in the same language ... document. It is still possible which Rioted also for Chinese lettuce and also what you've got you see here. So, based on that, there's an additional use case.

Because with audio in a worldwide company, you receive sometimes ISPs in the let us say, lingo from the company or the country, a country. I'm sorry for that.

Where that's coming from, so like Chinese and then if you find that show stopper right now within the original document, we were not translated into English.

So based on that, we first do, what do you show stoppers search In this kind of original language, just to prevent the cost for the translation, also, translations maybe could be risky if you will not translated best. So we do this as the first approach, putting it into a different language and assessing it into these kinds of things, but not only for the show stoppers.

Based on that, I'm really quickly through, these kinds of systems would be implemented for several business units. Most of them in the area of handling more than one thousand request for proposals per year. But also in any area better handle. huge abundance was more than 10000 pages for ISP.

So, and if you have further questions, so maybe, and we cannot answer, let us say, was in this session. So also, maybe it would be interesting. You can reach out to me via LinkedIn.

So it's just easy, because it's only maker at the end. And if you are still interested, or maybe you are more or less in the US based person. And it's also possible to reach out to Martin, which is our wallets, a vice president in the US, in Atlanta, or maybe for a writer. He's also keen to talk about just networking on just kind of scenarios sharing, some kind of additional use cases, or maybe things, why we use this kind of system, and on the other one, and about why be approached specific topics to speed up proposals. Because a lot of companies try to implement AI systems, but the overall question is, how to integrate it into this kind of proposal phases.

So for that, I would say, thank you for listening.

Hopefully I was not too fast on, thanks for joining.

So, happy to answer your additional questions.

This is this is terrific Oliver. Thank you so much. That's outstanding. What? An Incredible, what? An incredible project. What? An incredible application of some of this technology, that we have been talking about in the context of Business process management, and I'm a little bit biased because I come from a strong background on energy and infrastructure. And I understand the complexity of proposals, and in the, I'm just very impressed about what was done there. I?

The audience, you know, keep and send your questions and I'm going to keep an eye on your questions.

as we have the Q and A with Oliver here, so please continue to chew, provide your questions. And I'm going to keep an eye on the question box, two related questions to to Oliver. And then one of the questions is just, if we back up a little bit, Tell us a little bit about the history of this development. Because this clearly, in our complex proposals and having the collaborative work to work on this very complex proposals, has been in the industry for decades. How has And there are probably some many non technological approaches to making the collaborative work more efficient, more effective. Telling me a little bit about the History of Siemens. Where did this come from?

It was an idea that someone had, it's some, it's a concept that evolved how did this come about and the organization, yet it has a through line, so I had a department in the past where we had to assess or maybe do the risk assessment for large, complex solution projects.

So, a lot of for me, I don't want to say oil by the lot of large, complex, losing port project came across during the proposal phases and we did the risk assessment, more or less bugs are based, a lot of companies do it in that way.

And enter, interesting fact is if you see these one thousand or 10000 pages of tender documents or RFPs. We thought about the idea because in 2013, or maybe a little bit earlier, the technology was really there.

We have PDF documents to build the machine to read this kind of PDF documents and, and so, the idea was still, they're just starting reading these elements because there's slicing and dividing the document in. Shape, doesn't parts, is not easy.

And then we had the first, the first prototype, as you know, how this develops. So, you have to first prototype and test and discuss the first things.

And if it's the first prototype, we just talked about that, because I'm not sure it was in 20 13, if we talk about Cloud Base.

...

was keen to have it on his own computer and the next decision was, No. We will put it in the cloud. So, ... in was in Siemens, will be able to assess it and it will get the results and we have a secure server.

Um, and again, there's an additional idea behind if you have it in there the way that you will be able to assess more than one ISP in a specific way.

And will you will find two resides, and you will have not only one document, but what we do right now, because based on internal restrictions, let us say net weight W says one.

But if you have a sequence of one thousand documents, or maybe 10000 documents over more than one here, you will be, will be able to add some additional AI to find out.

And also, if you have to inflammation, which kind of them was one that not, Saudi history is coming from risky projects and doing workshops in the proposal phase.

And then, the first question was, wouldn't it be good to do it, again, four weeks earlier, or six weeks earlier? Someone that's coming from?

Very good. And that one of the questions that came up during our session was about the rules. I mean, so much of this, the power of this application, has to do on this, well, thought out rules, to decide what to do, Who does what. How, how did those rules come about? How did those rules evolve? What does the process like for creating those rules?

Then, first, I have to say, let us say based on a domain or the the vertical, or whatever you name it, you have to address like, oil, and gas, and energy, and transportation, and just kind of things. The processes are different. Because, they have different procedures and day therefore, and different amount of documents and just kind of thing, so, it will be different.

So, and based on that, uh, it, there's no one use case, one fits all use case. So, they are all different.

So, and if you say everyone, you mean just kind of, search terminology.

I think the question was more on the side of the intelligence itself. Why do you know, how, how do you even code or something like this into an automatic system? Did you have the groups, people, who are, like, What are you decided, What? The Rules. Of The game. Yeah. I know what you mean.

So, the first obligation that you have to say, Well, I have to say, the machine itself is Stupid, Is just following some single rules find, find it not a market, or market ticket, or ticket Look.

Interesting fact is the intelligence is an addiction Emmys and be fully dictionaries based on expert knowledge. So this is human intelligence.

So we started with Human Intelligence, what was, let us say, a non conformance cost or risk topic in the past.

And if we have that, we we at some kind of I would call it AI.

So intelligence to find let us say similar wording, so similar voiding starts with a single employer.

Also, if you are looking for Safety Manager, you should find Safety Management.

So it is a stemming to find the DVD, DVD stemming off of these kinds of phrases, and then also some kind of similar voting. So if If safety is interesting, so maybe it was reliable as Reliability is interesting.

But normally, you will have too much text at the end.

Because if you are looking for search terminology like reliability, the complete document is colorful.

So based on that, you have to combine it, like reliability and technical reliability, availability, and technical availability.

And this is also not relevant, because you need technical availability, enter percentage, and only the combination of technical and availability and percentage, whatever it is, or not, as a number is interesting.

So we use the false positive and reducing the false positive.

It's a combination of human intelligence, but you need the experts and some kind of let us say intelligence and the computer just to find the right wordings.

They're good. There's a perfect segue to a question that Mark McDonald posted online. And his question is on the engagement of this experts in the business. What was their reaction of this highly experienced? No experts on contributing their knowledge? Should this system, where they very collaborative on that, Did you feel that there was a certain resistance from them to contributing to this? How has that dynamic and working with the experts?

Uh, good question, because it was really interesting to see. There was some kind of, was it some kind of fear?

Because, if they give their experience into the system and all and not experienced proposal experts, let us say, in other countries will be blocked with their experience.

So, there was some kind of.

You have to let us say a company just kind of project versus change project.

If you would like to speed up, if you would like to take the knowledge from the experts, you have to give them some additional benefits or maybe advantages, like see, based on your experience, we will be able to handle the double the amount of proposals and based on that.

So yes, some of them were obviously, how do we say, was reluctant maybe?

Well, maybe Deborah really stressed and other ones were really keen to share their knowledge.

Because this is Dan knowledge into it within these dictionaries. So this is like if you are the real experts and you see your name that is the ... May be said there are also proud on becoming the real expert on these kinds of things. But SSH both ways.

Some fears and also some are really proud that they are the experts, so you need some kind of change project to accompany ideas kinds of things.

That's terrific. I think the message there is very useful for all of us, the, this change management aspects of that. So there is another question here from Dave Jannie. He has a bit of a commentary. He says, first of all, amazing technology. Obviously, being able to quantify 300% improvement in the session And the associated cost savings is huge.

Have you also being able to quantify savings from the program, catching things that a human has missed in the past? And and like you are likewise, do you have instances where the program miss something that was important? So it has this heavy use this, this intelligence to look at maybe things in the past And the have and then the current system. Do you find sometimes that the intelligence of the program misses some? I feel things are important?

Yeah, that's, that's, that's what normally the business unit with they, they, they let us say try or start to implement it.

So, this is that they assess the system. Let us say.

And the results of the system not just beat the results of the system against what they are experts would normally take as DVA as a requirement where they have to deviate or write a comment on that.

So, they, they take their count, all the, all the requirements whether they would like to do a comment and assess it to the detect requirements from the system.

And in Stunk and if you, if this is not sufficient, you have to work on your dictionaries, for sure.

So, it has to be a number of let us say below 2% or something like that for me, because it, you have to be other ways that the results are not very confident. You are not confident with the results because you have the feeling to miss something.

And the other thing is, ESB have some numbers of interesting, let us say phrases or situations.

Well, we missed some interesting requirements.

We don't I'm, I'm not able to shape the network how huge numbers, but the scenarios I mentioned.

I mentioned also the phrase Buy America.

We have a phrase.

For dysfunctional safety, we have a specific one specific requirement for casings, where we have to do helium pressure tests.

And normally, we do some kind of what our boss testing, especially for the density, that not in the density of the casings.

So, based on 10 day based on that, you can assume that a lot of are selected.

Let's say, I don't want to say sayings. But the phrases be missed in the past. I mentioned some of the examples.

But we never assessed all, let's say, or the complete amount of of phrases, which we missed and build and lead to, let's say, huge non conformance cost.

Ended at the end. Maybe one comment on that It is a combination. Doing project execution, you also need some kind of project manager with beards is also may be able to handle this kind of change or maybe not.

Very good. one other question that we have that's it's an interesting one, is that tremendous increases speed. You're certainly, of course, it shows up on the number of proposals that you're doing on an annual basis, as, I don't remember the exact number, but it looks like almost double or more than that. And, then a question on: on the quality side of things: Did you see a change on your on your proposal win rate?

Which means the rate of wins that you don't have it yesterday that there's also on, uh, but, but, not the huge effect that you can see that, because, therefore, you have to add additional use cases.

So, maybe also the quality is higher, but the ready to hit raise is also let us say, influenced by price.

Of course, we did not discuss price tags and something like that, though.

But, proposals that you should not go for, complex ones, where you assume you have a less price take on something like that.

You don't want to go for these kinds of things, or maybe you know that you are too expensive, Hmm, Then, then maybe you'll be, skip it out of the process.

Therefore, as I said, if you will be able to add some kind of case, where you're definitely identify the ones with showstoppers and take them out of the game.

And you will be able to increase the heat rate. And, therefore, you that this is tubular you will reduce the speed on all the other ones.

So, but there is no one number that we will be able to increase it by let's say 30%.

Because this is it's it's less than that.

Very good. ... Raj asks another question here. He's asking us that he has come across a similar system by Colfax. And he's he's curious if you know about Colfax and specifically how does the Siemens Document Management system compare with what what Colfax does? Do you have any experiences that?

I've never heard about Colfax before I saw similar systems during the last years: Most of them using some kind of AI, foss the most interesting question is always: how will be able to integrate is, indeed, the overall document flow process handling.

So I have I saw interesting things. Also interesting algorithms.

We try to integrate.

Um, but did you honestly saying I've never heard about Colfax?

I've ever looked at Upright in the next five minutes because it makes me interesting to see what they do.

Absolutely, absolutely. So ... Hendricks asks a question I think you may have mentioned around this right in your in your previous answer, but I want to confirm this. He's asking if, if you're using the system to, if the system has allow you to prioritize projects, and you alluded to that a little bit in your previous answer. But I, using the system to prioritize proposals and projects.

Media, honestly speaking 95 of the use cases of the more or less to identify interesting, or maybe risky, or maybe quality orientated requirements. So 95 to 97% of all the use cases are embedded in that direction.

We have additional use case in different departments, different business units, where they're also trying to eliminate these kind of, let's see, proposals, we don't go into golf, or, but there's an additional strategic elements because, there's also a strategic element, is a virtual customer for us there. There's a discussion about price and oil just kind of costs, and so you have to add a lot more into the discussion. Should be go for proposal or not.

So, the 95 to 97% of the use cases are still on the requirements.

Very good, So one last question. If you'll look at the developmental in the last 12 months, what part of it is that got you're most excited about the benefit a feature of, that you have developed in the last 12 months? What part of it is you get some more most excited about?

So, if you really far focused on last 12 months, that is the OCR because we didn't do the OCR Ing before Because, in that case, because we are responsible for the OCR.

Otherwise, we are still not responsible for, for the document we receive.

So does the ones wasting thing. And then what I'm really proud of is that the Foreign Service Solution, because there's not only having the solution, it's also additionally, it is key security.

And there's also additionally the topics about export control and Customs.

So if I upload a document and I go to China and download it there, I have to hand the export control and custom things.

And all these elements which are not related to software, not ready to exit modular Agile processes, more or less to surrounding laws and just kind of thing.

That's very exciting, Oliver, whatever remarkable project and application. And thank you for your leadership. Thank you for taking the time to share that with, with our global audience today. We really appreciate all of that.

Thank you, Jose.

Bye, bye.

Bye, bye.

Ladies and gentlemen, terrific practical application of technology to create tremendous value, not only for their, for their own organization, but clearly for their clients globally. So this is very exciting.

Now, you do not want to miss our next session. Our next session will have microphone, because talking to us about five steps to improve and business processes by delivering AI ops. So another dose of high-tech apply to business process management to deliver real value. So, I ask you that you exit this portion of the webinar. By closing, it, filling out a quick survey, providing any feedback on this session. And we look forward to seeing you back at the top of the hour. Thank you.