Intelligent Automation: The Future of Print Efficiency is in Flexibility and Adaptive Logic

Hi, I’m Pat McGrew with McGrewGroup. This is my friend Ryan McAbee from PixelDot Consulting, and we’re here today with one of our favorite technology people to talk to, Drew Sprague, Solimar Systems. Drew, thanks so much for taking some time with us today. Oh hi, hi Ryan, so nice to see you both this morning. Thanks for making time for me.

So, we want to come to you today to talk about intelligent automation, because I think there’s not as much understanding in our market as there really should be. We hear everybody talking about how artificial intelligence and GPT technologies are so amazing. They are, but for most printers there’s a baseline you have to achieve to get to where automation is actually going to help you out. And, as we’ve been talking, you coined the term to me, intelligent automation, and I like it because I think that it helps position the role of automation in a workflow. And, certainly the pantheon of Solimar products address the issue. But, could you talk more about what the value proposition of intelligent automation is for those who really don’t have a way to frame it in their business experience?

Sure, Pat. You know, I was thinking about this a couple of hours ago and the first thought that came to me was that what’s really important for individuals trying to apply automation better in their organizations is to get their heads around the scope and fit of the problem they need to solve. Now, for most people, that’s they need to get their production done reliably on time and accurately, and they need to do that with staff that can be variable, right? You might have some turnover, you might have illnesses. You know, frankly, people exit as they age out sometimes. So, you need to be able to maintain a steady stream of business. Steady states always work better than those that are interrupted and then stop, start, right? So, and that then allows the organization to focus on what we all want to focus on, which is bringing more opportunities in the front door and being able to handle those methodically and accurately and successfully.

So, a lot of the print shops that Ryan and I have done assessments in, as we approach them, we wouldn’t call those simple workflows, would we Ryan? There seem to be a lot of loops and turns. So, as you think about intelligent automation, how does it establish a baseline for how to approach simplifying that production output environment in a way that maximizes use of available labor and minimizes using them for repetitive tasks?

Well, what we have found is that one of the critical elements in automation is having flexibility because nobody stays in a static state for an indefinite period of time. Opportunities come up in the business that require the ability to adapt the system that you have. So, flexibility is kind of a cornerstone to that, and then you layer on top of that some adaptive logic. So, for example, one of the challenges people have commonly now that they didn’t have just a few years ago is they tend to get a lot of PDF content in and the source of that PDF content can be highly variable. They can be systems that people have had in place that were built by an individual in the organization that is no longer in the organization, and it’s producing some content that has some marginal characteristics that can upend a device that’s in the production workflow. And, so what we find that one way that this intelligent automation actually deploys in the marketplace is helping people isolate content that maybe needs a little secondary or tertiary spruce up to better play in the environment. And, it’s that first step, that isolating marginal content that then allows those other steps to take place in an automated fashion. And, some of our case studies that you can see out there relate to how people have benefited from this relatively simple step of applying adaptive logic.

You know, Drew, when we go into to shops and, you know, it’s a really consistent theme, no matter what type of printer, you kind of identify yourself as, whether you’re transactional printer, you do commercial printing across direct mail across the board. It seems like the challenge for people to really automate has a lot to do with that upfront. And that up front is basically receiving the content or the data from the customer, getting that into their production system that can then do the automatic triggers for basically any kind of file conversion, any kind of optimization of that file, any kind of imposition that’s needed for the actual print output. And, it’s those first steps that seem to be really a challenge for intelligent automation for most shops to basically kicked it off because they’re on the sneakernetting it be basically they’re having a person click a button, go download something, click the next button to get it to the next place, that sort of thing. And, with Solimar, I know you have such a good toolset there for that up front part that spans from Rubika to ReadyPDF, and so on. So, how do you see that kind of playing with intelligent automation to help that so many the pieces, so many printers struggle with which is that up front onboarding.

Well, Ryan, it’s an excellent point. You know if everybody had a highly standardized set of inputs that they dealt with and produced a very rigidly structured output from that, that was highly reliable as far as the information it passed to each and every stage of production that they could possibly have, then, you know, there really be no need for us. But that’s not the way the world works, because technology deploys evolutionarily. People incrementally build components for their systems as opportunities arise and as issues need to be dealt with. And, so what we basically provide is plumbing, just like the plumbing in your house, right? You got some elbows, some check valves, things like that. And, Solimar provides digital plumbing for the transactional printing workspace. And it’s this plumbing, this digital plumbing that helps to knit together the issues that arise from having sources for content that don’t produce necessarily what you as an organization want your outputs to be optimized to for the devices and delivery mechanisms that you have to support.

Drew, one of the things that you had said was that the idea of intelligent automation is to simplify producing accurate output. And I wonder, does that mean I have to be like a JDF expert in order to get to intelligent automation or is it the software platform that helps me get there?

It’s definitely the software that helps you get there and many, many, man years of experience that we put into building it with the components that are necessary to knit things together. So, in intelligent automation, a lot of that intelligence is actually our experience that’s gone into building into the software the, for example, heuristics and feedback loops that allow us to knit these pieces together. It might not seem obvious, but when you’re going out to index content, for example, you apply some heuristics where you deal with situations where there may not be a space character, but there’s separation between words, and you have to determine if that’s an actual space intended to be a word separation or not. That’s a very simple case, but that’s where intelligence gets applied in this process.

Does an environment that has already invested in using JDF to try and gather some data or who have vendors who have invested in JDF together, some data, does that data that is being passed back and forth, Is that valuable in intelligent automation?

Well, you know, Pat, whenever you put together an automation system, one of the critical components of that is feedback. And, what JDF enables that many interfaces don’t is that feedback loop so that you can, for example, know that the job actually completed printing. A lot of printing traditionally has been send it and forget it and hope it gets done because an operator said it got printed. But with JDF and other standards, you actually get actionable feedback. And, that’s really important because when you don’t have feedback, the theory obviously with control systems is without feedback, the outputs can go to unbounded and you don’t want that. You don’t want to have a system that has unbounded outputs because that results in bad things. So JDF, for example, is one way that you can strain your system and know that you’re actually producing what your system needs to produce.

So, that brings up, I love the fact that you use the term knit before because basically that’s to get to intelligent automation, the solution sets have to be able to communicate with each other. And, JDF is obviously one component of that feedback loop. But prior to even getting there, how does Solimar approach it? Maybe a little bit more uniquely in the environments that you are working and to basically be able to integrate with not only your solutions but particularly other solutions so that they can communicate to get all the preparation done before printing and everything that happens even after that.

Yeah Ryan, good question. So, the whole workflow that goes into a system, for example, in an ADF, an automated document factory, involves obviously connectivity and content. And, that content may have actionable information in it. It may have, for example, a page piece dictionaries that tell you what the important metadata is on the page for taking subsequent actions on that document. For example, splitting out mail pieces or sorting the mail pieces to take advantage of postal discounts and the communication is obviously critical for getting the information from whatever the host platform is. Could be a mainframe. And, then ultimately to what you want to do with it. Could be an IPDS print device, could be a PDF print device, so on and so forth. So yeah, there are a lot of pieces that have to be knit together. And then, in the middle are those secondary and tertiary processes you want to perform on that content to make it fit into a workflow that is determined by the product that you’ve promised to deliver. That could be as simple as envelopes with First Class mail, or it could be kitted content, for example. So, as you look at the Solimar universe of solutions and you apply the concept of intelligent automation to it, how does that inform how you deliver solutions like piece level tracking and the other things that to me seem to require some highly technical programing to make them all work, but also really relieve a lot of manual labor in the actual workflow production.

One of the things that always scares me about piece level tracking is that it always seemed to me to rely on having really knowledgeable people writing that program code to really understand how these pieces move through the system in order to be tracked appropriately. How does intelligent automation inform that?

Yeah, that’s a great question, Pat, and in fact, piece level tracking is an area of a lot of interest to our users because it solves a really important problem in their customer service organization, which is knowing what got delivered and when it got delivered and so on and so forth. And so, for a long, long time we have had a system called SOLsearcher Enterprise that allows organizations to serve content electronically, bank statements, insurance policies, things like that. So, a long time ago, we knit that together with our product called SOLitrack that provides more operational print shop kind of information about what jobs are being processed and what state of completion they are. And, the two of these together provide piece level tracking. So, we’ve knitted together with some nice, simple out of the box interfaces that users can then customize if they want. They don’t necessarily need to, but they can customize them to their own operational situation. So, for example, an organization might just simply put stuff into the postal system with USPS directly, or they might go through an organization that does presort or they might have some content that actually goes to UPS. And, this piece level tracking system is sophisticated enough to deal with any combination of those, in fact, and also at the same time provide a platform for electronic delivery of content and help you gauge if content actually got delivered electronically. If not, do you want to go and then post that at some subsequent date? So yeah, a big part of our intelligent automation is a platform. We call it Chemistry, broadly, that has these different components in it that provide different levels of functionality, flexibility and adaptive logic so that organizations can, as Ryan echoed, knit together these sophisticated systems that ultimately not only produce the content, but allow you to manage it successfully from start to finish and the finish being all the way to the consumer.

So, you know, piece level tracking and everything that you just ran through, sounds, you know, obviously it’s complicated because the workflows themselves are complicated. Does that mean that I have to have like a roster of developers to basically take advantage of the intelligent automation? Or is there some other approach that you’ve taken at Solimar.

Well, that’s a great question, Ryan. You know, it is complicated in a sense, but only in the sense that if you don’t have it, what we have and the tools that I discussed and you have to do it from scratch yourself, it’s almost an impossible task. But we provide an out of the box solution for solving this problem. So, it resolves the problems of, for example, you start with AFPDS or your printer is a PDF printer. Those kinds of issues and a whole lot of issues in between, including you, for example, want to be able to take documents that are formatted in letter and you want to print them efficiently N-up on some wide platen device. So, for the user, it’s a simple task because we make it simple for them, but for us it’s a difficult task. But if it weren’t difficult, there wouldn’t be value in it, right?

So, when you, it sounds like the approach then is to build a low code solution that leverages templates and libraries of things that can be easily roped together to automate these common processes that you know, your customers have to achieve. Am I getting that right?

Well, that’s right. And Pat, I would even say I’m no code, you know, basically out of the box and you just configure it. But if you want it to have a look and feel, that’s you, your brand, then yes, you customize it in a low code environment. But yeah, it’s out of the box and no code.

So, we’ve heard a lot of talk in the market right now about, you know, ChatGPT and GPT solutions and artificial intelligence, and you and I have talked in the past about how in fact an awful lot of the solutions that get branded that way are really machine learning solutions and more maybe really closer to intelligent automation than they are to artificial intelligence. In the print industry, do we expect that intelligent automation is going to be the best practice in the coming years as those artificial intelligence things sort of go off in other directions?

Well, the challenge with a lot of the ChatGPT and all that kind of fun stuff is it’s really misrepresented, I think, in people’s heads that it is like a problem solver. In my head, it’s really just an input for further decision making. When you go out and you use Bing, for example, you do a search, it gives you a nice collection of information, but it generally says, Well, I can’t tell you the answer, but here’s what I know. So, yes, in the space that we operate, the print production space, it is really more about intelligent automation than it is about AI. And, the reason for that is it needs to be not only evolutionary incremental, but it also has to have two really important components on the back end, which is verification and feedback. And, what you typically get with ChatGPT is not verification and feedback, you just get an input and then good luck. So, now that will evolve over time and there’ll be some really nice applications of it. You know, having something that can go out and collect a lot of information for you in a hurry is wonderful. It can provide quicker diagnoses and more accurate diagnoses for people that are dealing with a health condition, for example. But in the space that we work, what’s needed now is essentially systems that can deliver verified accurate output and that includes components that don’t really fall in the domain of AI at this time.

I’m curious, is there in the role of intelligent automation, do you see that extending to, because we focused on it from a print production kind of aspect, do you see that intelligent automation also extending out to the printer’s client and maybe the end consumer, as well? And, does AI have a kind of a role in that aspect with that kind of, just flipping the perspective around a little bit?

Well, it certainly does have a role there. And, you know, one of the areas that gets a lot of discussion, right now, is how do we make content more accessible for individuals. Not just making it easier to read in a physical kind of sensory kind of context, but in a contextual context. Sometimes people get documents that they just can’t understand because they’re written in for a user that is able to read it, for example, a college level and the words used and the organization of the content is just overwhelming. And so, they choose not to read content that they should. So, applying AI to be more of a contextual solution for people can really unlock some consumer benefits.

I can imagine that if you took a poll and you asked people if they’ve actually read any of their service legal agreements on the software that they’ve purchased, many of them have no idea what’s in those software contracts, whether it’s for a piece of software that you buy off the Internet or even an enterprise class solution.

So, wouldn’t it be wonderful, Pat, if every time you got an announcement that, hey, we’ve updated your agreement, you had a little assistant, they could say, Well, here’s what’s changed and here’s what I think you need to be worried about.

Absolutely. I think I’m waiting for that day because that would help me immensely. So, as long as we bring it around to how Solimar moves forward in this world of intelligent automation, what sort of directions do you think you’ll go? What are the things that you see customers struggling with that automation is going to be able to help them with in coming years?

Oh, that’s a great question, Pat. You know, we’re a funny blend of legacy and forward looking technology. So, for some people, and it’s not a trivial set of people, they still need to maintain legacy systems going indefinitely into the future. So, we’re going to continue to expand out our solutions for dealing with legacy content. And, what constitutes legacy content evolves over time, because even some PDF content now is truly legacy content. So, we’ll continue with on that end, and also expanding the feedback that we can get from any number of systems that might be used to produce content is going to be critical for the use cases that our market space needs and expanding our ability to integrate with verification systems of different sorts. Not necessarily just print, but it could be print quality things to, for example, on the fly, select a new ICC profile because something changed on the back end.

00:23:45:06 – 00:24:12:15
Unknown
Those are all great use cases too, and I can start to see how the evolving world of automation is going to make everyone’s life continue to get better. I don’t know that it’s ever perfect in print. I think everybody is always going to have challenges, but your point about legacy is well taken. We know from case studies that we’ve done with Solimar clients, that some of the data that they get is coming off of systems written in the seventies and eighties. So, it’s not going to change. There’s no intent to change any of those systems because it’s considered too risky. So, I think you have a long, healthy life in intelligent automation ahead.

Yeah, and we have to keep our expertise up because, you know, there’s not a lot of it around. So, as we onboard people, we teach them the old ways and the new ways.

So Drew, one of the things that Solimar has that is kind of unique is of a very, very long lived customer advisory council, like 30 years of history with your customer advisory council. And, could you talk to the value proposition of having customers who are willing to talk to you when and how you use their feedback and what it might have led to in terms of features that you have actually implemented?

Thanks, Pat. Yes. Solimar is fundamentally a referral business, right? We get most of our business from existing clients who say good things about us and share that with others. And, we really appreciate that, and we worked very hard to earn that trust. And, it manifests in ways that can actually be a little annoying for some users because when they open a case with us, we literally hound them to make sure that we solve their problem. It’s important to us to make sure that we’re doing the work that they’re expecting us to do. So, we make a big effort to make sure that we’re listening not only to the complaints and issues that they have, but also to make sure that we’re solving the problems that they have and looking forward to the things that they might need.

Your customer counsel, when we meet with them, is very giving group of people. Ryan and I sit with these people almost monthly and they’re not shy about saying what they need, but they’re also not shy about sharing how you’ve helped them be successful.

Well, we’re certainly grateful for the nice thoughts and things that they say about us. We appreciate that everyone is very busy and that they take the time to spend a little time with us, to share with us what’s worked and to share with us the things that may be a struggle in the work that they’re looking to expand into. So yeah, we’re super grateful for our user base and in fact, our partners too have done a great job in helping us to identify what their needs are and they kind of go hand in hand. The users and the partners.

One thing that comes out of talking to the members of the advisory council is that they really seem to appreciate, and it relates to the intelligent automation aspect, is that because there’s venues for them, whether it’s virtual or in-person events, they really get to talk and communicate with each other and find out, Hey, what are you doing to, you know, do this little bit for your automation or how are you using this tool from Solimar, and it seems like such a great benefit that they get real world experience from previous people so that they can kind of fast track their automation efforts internally in their own shop. Do you kind of see that happen a lot where it’s this community that’s kind of supporting each other and not from the vendor to vendor, or vendor to client, but it’s actually the advisory people in and of themselves that are sharing the expertise?

Well, yes, Ryan, I do see that even though I’m in the background and Mary Ann does a fabulous job of staying out in front of us and has for a long time now, we do get a lot of nice feedback from the community that we have built. Mary Ann has done a great job. Our support organization has done a fabulous job of just staying close to the user base and over many, many years, a very significant percentage of our employees have been with us for more than 15 years. I daresay a lot of them more than 20, 25 years. Of course, we’re only, what, 33 going on 34 here now. So, we’re super grateful for our employees. We love them. They love us and they love our users. It’s been a very, it’s been a wonderful relationship over a long, long period of time. And, we’ve all accomplished a lot of good things together.

Your CAC members regularly name check your support people on the CAC calls and talk about how helpful they are, how they ask the right questions and help lead to the best conclusions. Because, very often, the customer will put a support problem in thinking they know what the answer is and your team is so good at helping them get to the right answer, not necessarily the answer that the customer thought they were going to get. So, it is a really great group of people. So, with 35 companies on your CAC, which is an amazing number of people, 45 active users, we know, we talk to them all the time. Congratulations on having one of the most vibrant customer councils, I think, anywhere in the industry.

Well, I appreciate your saying so, we really do try hard to make sure that we’re solving the problems that they have.

Drew, thanks so much for taking the time to walk us through the concepts of intelligent automation. I know it’s going to be helpful to a lot of people who are worried about what they should be looking at next and worried about whether they have to talk about AI in their shop or if there are kinder, gentler ways to approach their automation requirements as they start to look at their challenges in handling their employee levels and just their workflow requirements as their workflow capacities change. Ryan, thanks so much for your time this morning as well, and we hope to see you on another one of these great SolmarSecrets videos.

Yeah, thanks so much, Ryan. Thanks, Pat. Nice to see you.