Skip to Content
Author's profile photo Thorsten Franz

BPM in Ten Years? My Best Shot

At SAP TechEd 2009 Vienna, I met Chris Horak, with whom I share a passion for good single malt, music (he’s into Metal and Bebop and plays Donna Lee on the e-bass), and the occasional witticism. Chris told me that he wondered what the future of Business Process Management might look like: “What will the successors of today’s BPM and ERP software in 10 or 15 years look like?”

Chris works at the office of the CTO at SAP and it appears that they try to look very far ahead. I don’t normally try to do that – frankly, I think it’s often difficult enough to see clearly what is happening right now in front of my nose -, but I took Chris’ question as a challenge and tried to come up with a few ideas. I let it sink in and a few thoughts emerged. Here’s my best shot at playing poor man’s Ray Kurzweil.

Goal-driven BPM

Today’s workflow processes are in a sad state (and I don’t mean Delaware). A workflow or BPM process is essentially a list of steps to be executed with a few conditions that define which steps to skip. Every workflow has a goal and exists because we want to see that goal realized. But where is this goal ever expressed? Nowhere! The step-by-step instructions we have instead are a poor substitute.

Imagine a team leader who tells her team members what she expects from them. Imagine she would do this by giving each step-by-step instructions for the next year but not telling them about the goals and purpose of these activities and what she hopes to achieve in year three or even after one year as the total result of their different efforts. The poor employees can only get hopelessly lost in the woods! Thankfully, such puppet-master style leadership should be very much the exception among humans.

But as the performance of computer is increasing logarithmically (if you haven’t seen the recording of Ray Kurzweil’s keynote speech from SAP TechEd Phoenix, please watch it at once) and we’re nearing the point where they develop human-like smarts, we have to learn to guide them differently and more like humans.

BPM as a Strategy Engine

The BPM of the future will not be based on step-by-step instructions, which are for idiots, but on the description of goals and purposes. Pursuing these goals, the BPM software will find a way as unstoppably as water flows downhill and Bill Gates came to the top.

It will simulate multiple possible execution paths, re-evaluate them iteratively, and incorporate any outside information that becomes available (which will be an awful lot) to update the strategy and tactic. It will respond flexibly to unexpected problems and obstacles.

Meta User Interface: A Living, Thinking Wave

I imagine it as a living, thinking Google Wave that dynamically adds the right people and bots as they are required to solve problems or provide expertise, and facilitates their work by blending in just the right information sheets, application interfaces, and tools. In most cases the engine will choose them autonomously so that navigation become as a thing we rarely have to do. The notion is old: context-sensitivity, except that the notion of context will be infinitely more subtle than today.

This means that the user interface is actually a meta user interface whose main purpose is to act as a substrate for all the other, situatively embedded user interfaces.

Exception Handling in a Living Wave: Resolving Goal Conflicts

In today’s workflow processes, exception handling by human intervention plays an important role. Many businesses have put all the business rules of their high-volume mass processes into their software systems, thus allowing for millions of documents to be processed “in the dark”, i.e. without human intervention. Only when a case cannot be decided by the machine – usually because it is a rare exception – is a human agent determined and given the task to correct the document or resolve the case otherwise.

Smarter BPM systems will require this type of human intervention much less frequently because they will be able to work with their input more flexibly. So what remains?

Even much smarter machines will require human exception handling when goal conflicts occur or the definition of the goals and purposes is not precise enough. In a complex strategic environment where a BPM engine pursues several goal at once but needs advice on their priorities or confirmation for sacrificing one goal for the benefit of another, it will query a human for clarification.

Obviously, this is similar to the way we consult our boss when a goal conflict occurs and the priorization of the conflicting goals is not clear enough.

Cross-Company Part 1: Business Process Representation

A challenge that was recently posed to SAP Mentors was to think about the company of the future. I came up with a slogan: “The company of the future is the cross-company.” By this I mean that as outsourcing, body-leasing, strategic cooperations, and many other forms of cross-company collaboration become commonplace, business processes transcend company borders. Business processes that span several organizations, each with their own IT infrastructure, require seamless infrastructural coverage and representation as exactly one business process (not many processes that are connected through interfaces).

This has clear implications on security: Nobody wants to be entirely transparent to their collaborators, which may also be (future) competitors. Future BPM systems must allow you to keep certain things hidden from external collaborators (but not your own company’s workers) while other information should be easily accessible to them. This will be especially challenging because due to the smarter BPM engines, the actual running processes will be much less foreseeable than today’s rigidly pre-fabricated workflow processes.

The data protection challenge calls for some kind of federation, just like the one that is part of the e-mail (SMTP) protocol or Google Wave’s architecture. (The more I think about Google Wave, the more I am convinced that it is the result of some serious visionary thinking and we will discover many more ideas in it that point at important trends.)

Cross-Company Part 2: Agent Determination

A smart agent determination system that looks at the business context, then finds the next human agent by matching it with the skills and responsibilities in the organization will have to work for external consultants and specialists working at a partner company as well as for its own company’s workers. You need to be able to find the right person for the task no matter if that person is an employee, as long as the person is available for the task. He could be a community member, a freelancer, or a job-seeker: The smart agent determination system of the future BPM system will find him and invite him to collaborate.

Rich User Context

The user context will play a much larger role in future UIs. First of all, UIs will play a bigger part in determining the user contexts via sensors and other sources of information: Who is using the system? Where is the user currently? Are we moving and if yes, where to? What time is it? When will we arrive? Is the light dim or bright, is it a noisy environment or silent, who else is there, where were we previously and where will we be next? What has happened a day, an hour, a minute ago, and what future events are we expecting? What is the rhythm, the flow state, of the situation?

Secondly, the UI will consider the situation and act accordingly. It will not bother us when we don’t want to be disturbed because we’re playing with our kids or driving our (Better Place ;)) car. It will provide information that is relevant to the current situation – e.g. an important meeting – as it becomes available.

The outside world (see the “internet of things” notion) will become part of the BPM engine’s user interface. A good example is the beer tap in this year’s SAP TechEd Demo Jam session that didn’t open when the person trying to operate it wasn’t authorized to do so. Doors and keycards are another example. The future user interface of a business process will not be limited to a computer screen, but be all around us.

Closing comment

I expect that most of what I wrote above is rather trivial to most readers and will become part of our everyday experience much sooner than 2019. But as the title says, it is my best shot at guessing the future of BPM, even if I make a fool out of myself. It might even all be yesterday’s news – I have no idea of what researchers are saying about where BPM is headed. Be that as it will, looking into my blurry crystal ball has been a fun experience and I will try it again.

Assigned tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Former Member
      Former Member
      I am here because the "Help" button on the SAP operated US Minneapolis MN Public School Human Resource system catapulted me into this community of Sapheads, no offense.
      I just wanted to find out whether W-2's (yearend wage and tax statements in the US) were available
      on this system.  Do you think there is a simple
      thing that answers questions. Oh, no! May be hidden, but not obvious.
      As far as BPM goes, your theoretical musings are phantastic, but to the end-user (the one who has zero knowledge and not enough logic or foresight) software in general comes across as "incapable of anticipating what it could possibly be that the end-user wants to accomplish". 
      You should hire more "zero knowledge people" like me in order to see what it is a person might want, or might misunderstand, or might try to see, but doesn't see.  Certainly, I would be under pressure to maintain my zero knowledge state of mind, and my natural forgetfulness helps a lot in this regard, but as soon as I know too much, you'd have to fire me. Because then I won't be enough of a useful idiot.  If you think about it, there is something about what I am saying.
      It's about understanding the psychology of the end-user, the human agent.  Since software architects are too logical, generally, to understand this, you'd have to artificially put yourself into a "no-brain status", or hire someone who is naturally like that.
      Author's profile photo Thorsten Franz
      Thorsten Franz
      Blog Post Author
      Hi Wolfgang,
      Thanks for your comment and feedback. About your first question, you might want to try the search function at I usually use Google search and restrict the site to SAP Library by typing in after the search words.
      Your remarks re: working with "real", unspoilt users - I couldn't agree more. It is only by watching humans interact with the systems that we can learn to make them user-friendly. Some companies (Google, Adobe, Apple) have gotten quite good at it and they do what you describe: scrutinize and analyze the human-software interaction and improve it wherever possible. If you look at the user interfaces of Business Objects, SAP's recent acquisition, you will find that kind of user-friendliness for the first time in the SAP space. And it's good to know that the Business Objects employees who bring us this UI goodness are now responsible for SAP UIs in general.
      BTW, I'm not an SAP employee. 🙂
      Anyway, thanks for your comments and I hope you'll stop by more often.
      Author's profile photo Jan Penninkhof
      Jan Penninkhof
      These predictions could both be excellent and scary at the same time. It is excellent, simply because the future BPM that you described would instantly make a obsolete. Computers would finally be understanding us, and we would be understanding computers much better at the same time. Its user-interface would adapt to what we're doing, where we are and what we want or even expect and would anticipate to it.

      The scary side of is that if this environment detected is smart enough, it could us all obsolete. If it is true that:
      - The user-interface of computer is really able to detect and understand the so-well documented social, economic and political environment (do I hear semantic web, anyone? 😉
      - could independently create BPM processes based on goals we feed into it, and
      - a $1000 computer would become as smart as the entire human population (as mentioned in Vienna Teched's keynote),
      where does that really leave us? In the Matrix, becoming the hands and feet of "the computer" that doesn't have such a great role in mind to meet its goals?

      I can't (or maybe refuse) to believe that we can achieve all this in only 10 years, but let's just follow the white rabbit and see where it leads us...

      Author's profile photo Thorsten Franz
      Thorsten Franz
      Blog Post Author
      Hi Jan,
      Thanks for your comments. It was nice meeting you personally in Vienna. 🙂
      I'm beginning to understand that as computers get smarter, the relationships with our computers and software will be more and more like our relationships with human colleagues and team members.
      If you're a manager, having highly able people working under you is a two-edged sword. You want the best people in the world on your team, but if they're not loyal, their smarts and ability can become dangerous and they'll take over your job.
      Some managers go for hiring only stupid employees who will never rival or overtake them. Some managers go for hiring the best and try to create win-win situation in which success can be shared and becoming obsolete in one position allows you to move on to be better position.
      I believe that's the strategy we should go for. But it's still entirely possible that "colleague computer" will turn out to be nasty, play foul, mob and discriminate against its human colleagues. I don't know what to do about that.
      Author's profile photo Former Member
      Former Member
      Hi Thorsten

      I believe that finding the right user within an organization, or even in a crowd sourced scenario for a given context should be possible even now.

      All the data is there - org hierarchies, text analysis etc...If an engine has access to SDN, Google, twitter etc in addition to org hierarchy, your emails and IMs etc - I would think that we can figre out what your expertise is, and whether you can be matched to a given context. OK ..maybe not 100%, but even if you can do it more than 50% of the time, wouldn't that be great?


      Author's profile photo Thorsten Franz
      Thorsten Franz
      Blog Post Author
      Hi Vijay,
      I think it's all possible today. Most of what is programmed today could have been programmed twenty years ago, but as tool support changes and platforms grow, what is now a matter of simple business content modeling would have required a major hardcore programming project with plenty of risks and open questions in the past.
      So my expectation is that for all the things I depicted in my post, tool support will be so good that e.g. rules and responsibilities for cross-company agent determination will be modeled in platform-independent and intuitive graphical notations similar to how we're describing the step-by-step of business processes in today's BPM.
      Author's profile photo Rolf Paulsen
      Rolf Paulsen
      Hi Thorsten,

      not only step-by-step instructions are for idiots, drawing BPMN-diagrams for every 4% variant of a process is for idiots, too. So I'm looking forward to a machine that derives BPMN-diagrams from well defined goals expressed in a declarative "4GL" process language (BPMN is "3GL"). But I think it will be difficult to give the machine sufficient metrics to find the best path. With this "4GL" the "P" in "BPM" looses impact - replace it by "G" for "Goal" instead - "BGM" is born (maybe to close to a film studio...)

      Resolving goal conflicts becomes fascinating if even two of the companies have different goals.

      Finding the right person to do a job is too expensive to be done only for one process. Why not let the machine interprete, remember, filter all these profiles and skills in advance and build up a huge database? This deserves the name "service registry" for human services - finally a practical use case for SOA. Then the BP, ehm, BG-Machines in the world can lookup the best fit ressource.

      But not too fast...getting the rich user context will be stopped by our workers' council and laws for data protection 🙁


      Author's profile photo Twan van den Broek
      Twan van den Broek
      Great that you share your crystal ball vision with us.

      "Today's workflow processes are in a sad state", I feel that this is a statement a bit too bold. It is already so much better than 10 years ago.

      * 10 years ago companies were internally department focused. Doing business within the boundaries of a department. When you are done, you do (hopefully) a handover to the next in line.
      * Nowadays there is a end-to-end process focus. Department have to work together to fulfill the process goal. But this is still an internal focus.
      * In 10 years (or sooner, my crystal ball is smaller 😉 companies focus on collaboration BEYOND its boundaries (Thorstens future cross-company).

      This cross-company thing has already started within communities as SDN and BPX. We are using Twitter, wikis and Google wave to collaborate beyond boundaries. Our current focus is more on issues that we encounter during our daily work and to discuss that collaboratively. But hey in a few years we will start working together on the process goals. Not because you're next in line but because you have special expertise that really adds value to the process.

      I totally agree on your UI statements. I think that today's UI's are in a sad state. A lot of info is pushed to you but a lot of relevant info not. Balancing needs and sources and filtering are key in doing your job right. We need to have access to the right info at the right time at the right place. And the right place is not behind your desk at your office, it must be available everywhere as processes run 24 hrs a day.

      Interesting thoughts on the future, maybe input for next years hackers night/design slam. Call it a future night, however we do need that extra day of SAP TechEd in Europe to get this on the agenda.

      Hope to share more crystal ball views with you.

      Author's profile photo Thorsten Franz
      Thorsten Franz
      Blog Post Author
      Hi Twan,
      Thanks for your feedback. When I wrote that current workflows are in a sad state, I meant that in comparison to what they can be. Of course you're right in pointing out the great improvements both in thinking and technology that took place in the last 10 or even 5 years such as breaking up organizational and technological silos.
      Hacker Night and Design Slam can be excellent vehicles to experiment there. And who knows where we will stand in one year? I think the innovation speed is increasing very rapidly here as creative geeks are swarming all over the topic.
      Author's profile photo Jon Reed
      Jon Reed
      Thorsten, good stuff - such speculation is good - an imagining of the future may bring lessons on how we are to get there.

      I don't have much disagreement with your vision. It strikes me that in the long term future, business managers/line manager are able to tweak processes and press "go live" and there you are. No one in their right mind would claim NetWeaver BPM or any other tool is there yet.

      So would developers go away? No, I don't think so, but they might seem more like orchestra conductors than virtuoso violin players, if that analogy makes any sense. The rhythm and flow would be structured in techno-functional collaboration.

      I agree about cross-company collaborations being that "wave" of the future" Project teams assembled across organizations, disbanding when work is done - with all unstructured information generated during the project tied neatly into the system where all who are accessing the results of the project can see it and add to it - like a wiki on steroids. Are we there yet? No, not anywhere close!

      But: seeing Gravity at work at the Process Design Slam made me think we're a lot farther down the path than we were 10 years ago or even 5.

      To turn the question into what SAP should do, Sikka himself admitted that SOA hasn't worked out - only because it hasn't gone far enough. SAP needs to empower customers to be able to truly "compose" on top of the suite, in a way that does not require as much explanation and proprietary know-how as it seems to now. NetWeaver BPM could be a big part of this, we'll see - though the tool is less important than the fluidity of use for those project team members who want to differentiate their company with unique, flexible processes- without either disrupting the underling transactional security/flow or having to sort through protocols on how to make the integration work. Hey, on your comment log a guy can dream! Though I think SAP does vision something like this also, very much so: let's see it come to fruition, sooner rather than later.

      Author's profile photo Thorsten Franz
      Thorsten Franz
      Blog Post Author
      Hi Jon,
      Thanks for sharing your thoughts. Isn't it amazing how much we can get out of that composition/orchestration metaphor? Let's play with it. The job of a real music composer/arranger includes a lot of "engineering" skills, craftsmanship, expertise, and quiet concentration, much like software development. Many people have no idea how carefully crafted even hardrock song arrangements are and how much musical skill and training goes into those.
      If we want to extend the metaphor and describe lightweight, ad-hoc process design (if that is not an oxymoron) that doesn't require a lot of training, it should me more like whistling a tune than working out the contrapunctual movements of 6 voices in a fugue.
      Another approach would be that of the jazz improvisation: Someone wrote the harmony (defined through chord progressions), there's also a tune that is consistent with the harmonies, but everything else you make up as you go along.
      In many cases, as musicians enter the stage, they have never met before but are able to play along together without practice, preparation or even the opportunity to discuss anything in advance. I think I feel a blog coming on how this works and what we can learn from it. 🙂

      Author's profile photo Jon Reed
      Jon Reed
      Thorsten -

      good stuff. I like the notion of jazz improvisation also: "Another approach would be that of the jazz improvisation: Someone wrote the harmony (defined through chord progressions), there's also a tune that is consistent with the harmonies, but everything else you make up as you go along."

      The structure of jazz would be akin to the stability of the transactional ERP system, whereas the vital freeform improvisation would be BPM at its best and most evolved state.

      Nenshad Bardoliwalla (@Nenshad on Twitter) recently wrote in a blog talking about E 2.0 that E2.0 and the core enterprise would be perpetually intertwined, rather than one trumping the other. I think same will be true of BPM and ERP. Of course, we have a long way to go to achieve this kind of fluidity.
      Sometimes the lessons are that simpler is better. James Governor (@monkchips) had a great quote on this regarding SOA versus REST: "While enterprise architecture astronauts were delighting in complexity, REST was taking hold. "  Hopefully with BPM we don't get carried way with complexity either, but focus on empowering users to get directly involved.

      - Jon

      Author's profile photo Former Member
      Former Member
      Hi Thorsten - I really enjoyed your blog and the responses to it.

      One additional change I hope to see - greatly increased process agility. By that I mean that the process should be at least as easy to change as the paper process it replaces.

      Cheers, David

      Author's profile photo Former Member
      Former Member
      Hi Thorsten,

      you describe a nice vision. From a research perspective, most have been already investigated. Sadly enough, most research has been ignored and research that has been done ten years ago is now repeated.
      I agree also with you that step-by-step instructions are not the future to go. This originated from the development area. Subsequently, most development paradigms (changes to process models at run time, aspects, rules, model checking etc.) have been applied to this without ever thinking about if it makes sense.
      I have a provocating statement (on purpose): Most processes can NOT be supported with current flexible/adaptive/... bpm technology. Imagine engineering processes, innovation processes, creative processes, software development, service (not web service) processes, personal processes (i.e. your own life processes), crisis management processes ... They all have not adopted BPM technology for supporting their processes.
      Hope this provocation initiate another nice interesting discussion 🙂



      Author's profile photo Thorsten Franz
      Thorsten Franz
      Blog Post Author
      Hi Joern,
      Thank you for a very interesting comment. 🙂 I'm sorry if I have to disappoint you, but of course your statement doesn't provoke me at all. On the contrary, I agree completely that the current step-by-step design approach is unsuited for many essential types of business processes. In my blog Wave, Web Dynpro, and the Magic Glue: Google Wave and Web Dynpro Integration Part 1 I tried to show that the combination of Google Wave and ERP systems can lead to a kind of next-generation BPM approach that covers the middle ground between the rigidly pre-structured (classical BPM) and the completely freely improvised (pure Wave). (This thought shows up here in the term "living, thinking wave".) Here I see the potential of covering new types of processes. I'd very much appreciate your feedback on this one.
      Author's profile photo Former Member
      Former Member
      Hi all,

      first of all, just to say that, even not knowing much about the subject, I have had a very good time reading the article and all the opinions about it. Really like the "jazz improvisation" example with chords progression, linking music with programming, oincredible 🙂

      So I wanted to pose a question. There is a mention to "having the technology and the means ten years ago to do all this". As it is stated, when we humans work together, the most difficult part is usually to get to understand each other and cooperate. It takes time and a lot of effort. So development usually comes by the hand of new standards, work groups, platforms, agreements, etc. which are the result of a myriad of private interests. Just like in the stock market the price of a value is the result of the mix of different interests of many agents, the development and future of BPM (as other technologies) will be the result of private interests.

      So, this long long introduction to get to the point 🙂 Do you think that all agents are already taking the steps alltogether towards the direction of the future described in this article? I get the feeling that in such an extended interconnection state, crossing company boundaries, etc. each company could have a completely different system (ERP, etc.) without loosing interconnection capabilities. That could benefit some "in the cloud" companies (Google?), but, does it benefit rather than damage software vendors like SAP i.e.? If the standards for such an interconnection are set, where will SAP (and others) will try to set their added value? Will each company try to create their own standard? (in this case we may still wait for another 30 years! 🙂

      I know this article was an exercise of imagination from a technical perspective, but as the subject goes, I am just trying to see some practical opinions.