Business Rules Forum 2008 Experts Panel (Baisley, Ross, Sinur & Taylor) : Emerging Trends and Decisioning
Don Baisley, Microsoft Don.Baisley@microsoft.com Ronald G. Ross, Business Rule Solutions LLC firstname.lastname@example.org James Taylor, Decision Management Solutions email@example.com Jim Sinur, Gartner firstname.lastname@example.org
Kirsten Seer, Business Rule Solutions LLC email@example.com
- Is decisioning a rebranding of business rules, a morphing of business rules, or something different?
- If we base decisioning on statistics will we miss a sea change?
- With the emerging focus on decision management shouldn't more business people be invited to this conference?
- What is your assessment of natural language (controlled natural language) and what practitioners can expect in the future?
- Is there some secret sauce I need to add to my panoply of tools?
- Is strategy really necessary to make this work?
Good afternoon. How is everyone doing? Has conference fatigue caught up with you yet? You'll want to stay alert and pay attention, though, because we have some really good things happening this afternoon, starting with our "Emerging Trends and Decisioning" panel.
My name is Kristen Seer and I'm with Business Rule Solutions. I'll be your moderator for this panel. From the previous panel [which John Hall moderated] we have a wonderful set of rules. I asked him if I could adopt his rules — I'm big on re-use so here I am, re-using rules!
I really like this panel; it's a good opportunity for you, now that you've been at the conference for a couple days. You've seen what people are doing; you've seen the great products that are out there; you've heard what people are experiencing and lessons they are learning. Now there's an opportunity to focus on the future. Here we have assembled an illustrious panel with some of the big thinkers in the field. These are the folks who do the serious thinking — look at the big picture — and try to see where all this fits together and where it's all going. And this is your opportunity to ask them some searching questions about where things might be going.
I thought I would start off by asking each panelist to introduce himself, even though you likely know the panelists already.
And, also, I have two questions to start them off. These are related so I've put them together. There has been a lot in the news lately — things going on in several different areas that I believe are going to have a real impact on the direction we take in the future. What I'd like to know from the panel is, how do you think we need to shift direction, in light of these events? And the two events are, of course, (1) the global financial crisis, and (2) the introduction of the big players into the field. We heard the announcement yesterday that Oracle had purchased RuleBurst/Haley. We have IBM with ILOG. And Microsoft acquired Don Baisley — which, unfortunately, he can't talk about, so I'm being very careful in what I ask, but he'll tell you what he can. So, Don, why don't you start off.
I'm Don Baisley. I've been at this conference many times, but in the past I was working for Unisys, as many of you know. Now I'm a software architect at Microsoft, working in the same field, on a lot of the same things. I was very much involved with the development of SBVR, along with Ron. I love to see that move forward and to see tooling happen that supports that so I'm involved in that kind of work.
Do you want me to answer that question now — What do I think is happening? It's good that you ask me first because otherwise I would just have to agree with Jim over here, since he's probably the expert on some of these trends.
What I'm seeing, in terms of the financial crisis, is that companies are pulling back — they are tightening their budgets. That's going to slow down innovation, in some respects, because the far-reaching kinds of projects — the kinds of things out there in the future, in terms of where it's going to pay back — that's where they're cutting back first. They are focusing on immediate deliverables and immediate cash flow.
But there are some exceptions to that. One is a refocus on, How can we save money? Or, for a business selling product — How can my product help my customers save money? — because that's the immediate reaction. So expect to see some innovation happening in that area.
Also, you're going to see more buy-outs because the collapse in stock values is making for some good deals. Companies that have cash can afford to take over. So I think we're going to see more takeovers ... more consolidation. That may also slow down innovation a little bit because, during these takeovers, there's a period of time of re-situating before you get back again into that mode of creativity.
In the area of trying to help people save money you'll still see some innovation, and decision making — automating decisioning — can fit into that category. So I'm hoping to see some progress there.
I'm Ron Ross. I wear a lot of hats ... which is why I don't have as much hair as I used to — Chair of Business Rules Forum; Executive Editor of BRCommunity; participant in the SBVR standardization work; methodologist; columnist; book-writer; a follower of James Taylor's blog (which, by the way, is excellent).
I think the answer to the question of the events currently happening basically means that, all around, there is more pressure. If you didn't think there was pressure already on those of us trying to bring about effective solutions for companies, there will be even more pressure now.
In many ways this is good, because the more the pressure, the more people begin to realize that existing techniques and approaches and methodologies — the same old answers — just aren't good enough. So they're going to be looking around for some new ideas and new opportunities. And you guys here at the conference are, hopefully, in a position to take back some ideas that you can contribute to some new directions in your organizations.
So let me introduce a maturity model, for the first time ever, in the decisioning area. (Hopefully, James, you don't have a maturity model for decisioning yet?)
It consists of three phases.
(1) Learn. Certainly, here at the conference we've had the greatest learning experience we could because we've heard from real practitioners who have been solving real problems ... not a lot of hand-waving; not a lot of consultants; not a lot of analysts (except the good ones, like Jim here). And not even a lot of software engineers (except good ones, like Don here).
(2) The second phase is Apply. Go back to your organizations and apply some of what you've learned. Try to identify problems and case studies — socialize the message back in your own organizations.
(3) The third phase, which I think is really the important one, is Scream. What I mean by that is, now that you have the large vendors in this space — IBM, Oracle, Microsoft (and throw in SAP and whoever else should be mentioned) — when you see that your current tools are inadequate, and when you see how much more could be done, and when you see that you need the kind of support that's not available in the marketplace (by and large), go to the reps for each of those respective vendors and tell the reps that what they've got just isn't good enough. Put pressure on them. You'll see the vendors eventually listening and responding, because it represents opportunity.
The bottom line is that, really, we have to make it happen.
One last comment is that business rules — and now decisioning — in many respects is unlike any other area/focal center in the past ... at least in my experience. Because this really hasn't been software tool-driven. It really has arisen from practitioners who have seen the problems over many years — Business Rules Group, for example, very active for twenty years — and have attempted to communicate that what we have in front of us is, first, a business problem and then we have a platform problem. Yes, there is a platform problem and you can take that message back to the vendors.
I am seriously tempted to invoke Rule 5 for Ron ... but he's my boss so I decided I'd better not.
I'm a Vice President at Gartner, after doing a hiatus with a vendor and realized I was too much of a Boy Scout. So I'm back in the Gartner world, where I cover business process management and business rule management systems.
A couple of points. I stay connected with the real world. I spent twenty-plus years (I won't admit how many!) implementing large-scale processes and rule systems and rule decisioning systems. (Ron: I know the answer to that!)
Shhhh..... shhhh..... <laughter>
And so I like to stay in touch with customers. I am the most active analyst, with the most enquiries at Gartner, ever. I have 14,000 enquiries over thirteen-and-a-half years. I like talking with clients, and I try to keep in touch with what their pain is and what their problems are.
The global financial situation is doing something different this time than in other downturns. I have lived through a couple of downturns — we had some in the mid-nineties, and also in 2001. Clients are actually categorizing themselves into three categories, right off the bat. Either they are trying to survive, or they're trying to thrive (which means that they think they're going to survive and they want to grow their business), or they want to innovate (which means they want to use this time to really outpace their competitors).
When I get questions from the people who are trying to survive, it's all about ROI — it's about how much can I save, when can I save it, and let's make that happen.
The easiest thing for me to tell them comes from our surveys that show you get great ROI from business process. We're telling people to double-down their bet on business process. Here's the reason why. In the nineties there was an industry that literally almost died — it was called the telecom industry, in the late nineties. They used process as a tourniquet, to keep themselves from bleeding to death. So, we're saying, double-down on process.
Now you say, "Well, we're at a rules conference! Why are you talking about process?!"
Process drags rules in.
Even though rules (when we do our surveys) has an ROI of 15-20% minimum — maximum 120% — and BPM has 20%, maximum 220%, for some reason rules doesn't have the resonance.
We at Gartner are going to put it in a term that we kinda stole from James — a 'decision management' piece — we're going to call it "intelligent decision management" (only because we like to name things). We think that's a hot area. We're telling people that if they can make better, faster, quicker decisions than their competitors that they are going to survive, thrive, or innovate.
Now, I have to say, most of the companies that innovate are concentrating on processes rather than rules. But I'll add that a lot of companies have learned a lesson — that it wasn't the mortgage loan processes that screwed up our economies (worldwide). It was the stinkin' rules and policies.
So what we're telling people to do is, plan for scenarios that link to policies ... that link to rules. And that requires some pre-planning — maybe some predictive analysis, maybe some simulation — to be ready for these things.
Do you think our government was ready for what they faced? Do you think the world community was ready for what we faced?
Nay, nay — they were not.
You could see they were reacting. They tried to keep a nice, stiff upper lip but they were scared (can I say this?) ****less (has four letters, starting with 's' and ending with 't'). Yes, they were scared. Frankly, some of our clients — particularly in the financial services industry — are scared. And so we're telling them how to save money.
To summarize: I would say it's around doubling-down on processes and doubling-down on decision making. Both of those are really hot topics. That's the advice we are giving now.
I'm James Taylor. I think most of you know who I am — I'm the guy who's having his invention stolen by Gartner. <laughter> ... only the label.
In terms of the global financial situation what I think is interesting is, with all the customers I've spoken to over the years, who've done rules projects and done decisioning projects, it's amazing to me how many of them never did a formal ROI study. Because the problem they were doing the project to solve was so awful that no one needed an ROI. It was like, "As long as you don't allow that to ever happen again, we're good." That's the key thing here.
What strikes me about the current crisis ... I think we're going to see a wave of people — not perhaps next year but in subsequent years — look back and ask, "Okay, what could we have been doing differently that would have meant that we didn't get into that mess?" How would we have to have built those systems? How would we have to have managed our rules? How would we have to have run our businesses? ... so that we didn't get caught out by this sudden change.
I think that will drive people to look at decision simulation, to look at what-if analysis, and to realize that if they don't control those decisions they can't replay things ... they can't retest them; they can't see how things would have been had they acted differently. So I think that, in the medium term, the current crisis will create another whole wave of projects that are driven by the "don't let me get in this much trouble next time" perspective.
I think it's also going to shift the focus from (curiously enough) risk management to growth.
We often differentiate between risk-centric decision making — which is where analytics and data mining and, to some extent, rules have historically been very strong: How do I avoid making a decision that has a big downside risk? — and beginning to see more energy placed into what I call growth-oriented decisions: How do I retain customers? How do I up-sell and cross-sell people? How do I sell more valuable products? How to I price more effectively, so that I get more money out of a particular customer?
I think those things will be very appealing to people during tough economic times. And I think projects that are focused on "How do I get more revenue, more margins, out of the people I already have as customers — the ones who are surviving, the ones I'm keeping?" will show a bigger percentage than they have perhaps historically.
Great. Thank you.
So it sounds like even though initially we might think "Oh, my gosh!" there is opportunity. If you target things right and look at things in a positive way, there can definitely be some opportunities out there.
And now I'd like to turn it over to all of you. Are there questions?
[from the audience]: I think it's fair to say that decisioning seems to be an up-and-coming trend. Its appearance here at this conference is evidence of that.
Do you think that decisioning is a rebranding of business rules, perhaps in more business-acceptable clothing? Or is it a morphing of business rules — simply a maturing of it? Or is it something different?
I actually think that each of the guys up there is likely to have a different answer, so I'm not sure I want to direct this to anyone in particular ... maybe to each of them in turn.
I'd like to take a shot at that.
I think decisioning is a convergence of things that were separated in the past. Rules ... was separated from business intelligence ... was separated from process. I think we're now seeing a convergence of complex events, rules management — and 'rules' not just from an engine, language, execution perspective, but how to manage the rules — how to manage agility that we've never had before, how to deal with processes that have to adapt around a changing world. These are all able to come together because a lot of the integration infrastructure is already there.
Also, I think there's a real premium on making good decisions around existing applications, packaged applications, and the processes that we're building. I think that this is a convergence that somebody might have been able to anticipate ... but I've only seen it coming together for the last three years.
Yes, I would agree very strongly with a lot of what Jim has just said. But I would add, one of the things I notice when I talk about it, I'm using 'decision management' primarily to differentiate between the wide range of things you can do with a lot of the underlying technologies.
If you listen to people who use data mining tools, some of them are using data mining to inform a knowledge worker, or to come up with some great strategic insight. Others are using data mining to power decisions. Look at 'rules' — some people are using rules for data quality; some are using rules for constraints; others are using rules for decision making, optimization, and performance management. A lot of these technologies come together around a focus on making better decisions, particularly better operational decisions.
So decision management is both a subset of a number of technologies and an amalgam of those subsets. It's neither more than nor less than because it's different from.
The key observation I would make is that we've always believed that business rules have business value, but we weren't saying it — it's not self-evident in the words and the terms we're using. For people who are into semantics, into vocabulary, into words, that is a problem. So I have to congratulate James for being a leader in the move to reposition the focal point away from the artifact, which is 'business rule', and toward the business result, which is the decision that you use the rules for.
I think that is an important conceptual breakthrough in how to think about and position business rules. Those of us who are working in the field have always known that, but you have to say it right in order for the message to resonate.
That's one part of the answer; another part of the answer is, if it were rebranding I probably wouldn't admit it here. <laughter>
The third part of the answer (and I mentioned this briefly in my Keynote) is, I really think that, in addition to what Jim said about the repositioning (the convergence, another part of this is — without realizing it and taking a somewhat longer term view — we're in the transition to the knowledge economy. And in a knowledge economy you have to have a much better focus on the intangibles of your organization ... because that's where the value, in many respects, is — the knowledge that you use to do the day-to-day things.
So I think this change in language is going to help explain, longer term, what it is we bring to the table that moves companies from industrial age, from information age, to a knowledge economy ... a true knowledge economy, not just the hype about 'semantic web', but really where companies are organizing their fundamental know-how in the form of vocabulary, semantics, rules, and so on.
I think I agree with everything I've heard up here so far in answer to that question. I think of it, though, largely as rebranding in the sense that we've recognized that a place where a lot of the value is coming from the rule technology has been in the decisioning area.
When someone asks me, "Where are you going?" and I tell him I'm going to a "business rules conference" ... Huh? That's gotta be boring, right?
When you talk about 'business rules' it's hard to appreciate the immediate return. But when you talk about making good decisions, which is an important business activity (especially in certain fields), then it starts clicking in people's minds where the value comes from.
So 'rules' is really just a part of making good decisions — it's a part of doing things in a context of repeatable process and getting consistency and improvable conditions. Decisioning is where a tremendous amount of the value is coming from. There is also value from formalizing business rules, but I think that decisioning is important as far as the current value proposition.
I'd like to add one more thing. Dr. Kim Warren says that there's a major sea change that's occurring. For about the last fifteen years there has been a belief, in the free market world, that everything is "market-driven." He believes that things are "resource-driven." And if they are resource-driven then you had better make some better decisions about how you optimize your resources and leverage them. So, his feeling is that part of this emphasis on better decisions comes from moving from a market-driven set of economies to resource-driven (resource utilization).
That is probably right because some of the customers we're working with in developing business vocabularies and business rules are fundamentally rethinking and re-engineering their products, as in themselves a resource, in order to be able to accommodate choice and customization, rapid time to market, and so on. Certainly there is a competitive market factor to that, but it's also that they have to just manage their resources better.
All right. Next question?
|[from the audience]: So we started off with decisions and knowing what our decisions are. And then we automated them with that business rule management system so that we can apply them automatically and consistently. And then James and Neil came by and said that now we need to learn how to make better decisions, and then close the loop back into the decisioning process. A lot of the learning about how to make better decisions is using predictive analytics, which is statistics.
Now, I come from business where we wrote off six billion dollars this year on sub-prime, and we got caught up with our statistics because the problem is that a lot of statistics only look back so many years — they don't handle the once-in-a-decade/once-in-a-century kind of phenomenon. (Of course, we always knew we were supposed to do what we would call stress testing, or scenario modeling, which is "What-if the world imploded ... does my business still survive?")So, do you think there is a risk, as people start applying more business intelligence, that they're going to make this same mistake ... rely solely on the statistics and miss the sea change, the paradigm shift that could suddenly occur in the year? And now their whole decisioning, which is all automated, is off and this catches them on the wrong side pretty badly.
I think that there are a couple of things to consider. First of all, I've heard at least one well-informed speaker (at another event) arguing that this is just a bigger version of something that was almost exactly the same ten years ago. He argued that it wasn't a once-in-a-century event. But nonetheless, those things do happen...
There is always the danger with analytics that the past may not, in fact, be a good predictor of the future, for certain kinds of decisions or at certain moments. I think there is always a danger when people use an approach and fail to think through the consequences.
We are seeing this year, at this show, a lot more focus on what-if tools and simulation ... something you really only used to see at the top-end, at the very big customers who had built that kind of capability themselves. Because there is a recognition that you can get into trouble quickly — as you said — when you are automating this stuff. They have the high throughput; they have the high transaction volumes. So I do think there's a risk there, for sure.
However, I'm reminded of the story about the guy who was one of the first doctors in his part of the country to automate his practice. He used to get questions all the time about, What did you do about security? ... and back-up? etc. Eventually he was asked, "What if your building burns down? Do you have an off-site backup?" And he said, "Wait a second! If your building burns down, do you have a backup of the paper files?"
We wouldn't necessarily be in a better position if we didn't have the automation ... if we weren't using the statistics. It's easy to say we're in trouble even though we were using these statistics because they're not predicting the future very well. But how well would the people who've been selling the deals have predicted the future if they'd been doing the calculations by hand?
I think these things are not perfect, but — like Democracy — they're way ahead of anything that's second.
I absolutely agree with the premise of the question which is, yes, there is a huge danger in only looking in the rear view mirror ... at data and statistics of the past. One of the areas at the conference that we haven't spoken sufficiently about is strategy — business strategy — and the fact that business strategy is another thing that can be easily structured and should become a focal point of deliberate management in organizations.
It should be important to have well-structured, well-formulated strategy, at each level of significant business activity, with well-known goals established and risks identified. You get the best minds in the company thinking about all the significant risks and you put together a model of that. This could lend itself to all sorts of interesting simulations, aided perhaps by analytics, and constraint engines, and who knows what else.
The first step up the ladder (oooh, I shouldn't be talking about ladders — I broke my arm a couple of years ago, falling off a ladder) ... the first step up the stairs is just to get people started thinking in terms of a more disciplined approach to that process of governance, with strategy being on the front end of that.
There is a danger. A similar danger, before 9-11, was that there weren't scenarios — there weren't crazy scenarios for crazy things in the government when it came to homeland defense. Now there are.
The same thing's going to happen with businesses. Gartner has been talking about Strength, Weakness, Opportunities, and Threats for (I don't know) fifteen years. We wouldn't get many questions on this. Now the questions are pouring in — people are saying, "We've got to think about what could happen, and we need to have scenarios." Yes, there were some paranoid companies who were already doing that. (I worked for one of them. They had about sixty or seventy scenarios that they had pre-planned.) But not many.
I think business rule management needs to step up to the plate and not just manage rules but also manage scenarios and the rules that are connected to them. And they should be inventoried — be ready to go so that when a complex event comes in and says, "I think you've got this condition" you can pull the rules associated with answering that condition off the rule management system in rather short order, assuming that some humans have something to say about it.
I think this is a sea-change event. And I think we're not just going to look backward, we're going to look forward at the complex events and we're going to try to identify threats and opportunities, and deal with them. I'm hearing more and more companies ask that question in the enquiries I get ... and I expect more.
I want to add to my answer. If you're interested in this area of strategy the Business Rules Group put together the "Business Motivation Model" which is now an OMG standard. You can find it on the BusinessRulesGroup.org website. It's free; it's in business English; it's readable, with lots of good examples. Take a look at it — I think you'll be convinced just by looking at what that's all about.
By the way, this is another instance of structured vocabularies/fact models being central to a problem that you need to talk about and analyze in depth. You know that sooner or later it always comes back to vocabulary — essentially, what this model of 'strategy' is, it's just a vocabulary.
I want to add that I think the analytics are great and they are important and they help for making improvement as trends continue. But businesses that will stay well-positioned have got to be creative about considering potential risks. And that you're not going to find in your statistics. You have to have a creative mind, and you have to do a lots of what-if kinds of thinking and imagine a lot more possibilities than what you're going to see in the past.
That's going to help people prepare. Hopefully that helps them in creating guidance so that their organizations are situated to react quickly and, in some cases, even take advantage of what might be devastating for others. To be prepared ... there the analytics just don't help you. It's creativity and being willing to look at possibilities.
Sometimes rules can help there. Structural rules are about what's possible. In your business models you may even be able to use some sort of automation to toss out possibilities to you — based on the rules that you have, things that would seem completely unexpected in terms of past history. Then you can look at it and say, "No, that really is impossible." But maybe it will open your eyes.
Okay. Next question?
|[from the audience]: Many of the folks that I've talked with at this conference tend to be more tool-based, or rules-based (and this would include me). Many of us are analysts. So we're basically the tails wagging the dog when it comes to talking about big decisions in our companies. If this conference is going to move into more of a decision management direction, it seems like the invite list needs to expand to include more of the folks who actually have an input to decision management at their companies ... and not just the keypunch people.|
I think, from your question, you're talking about the business managers ... the business side. Is that the essence?
From the Audience
Yes, the owners of the business decisions.
The owners of the decisions in the business ... the people whose jobs should be on the line when bad decisions are being made constantly? Yes. I'll speak to that since I'm one of the organizers of the conference.
The problem we face today — and I'll be open about it — is that, in a lot of organizations, IT still has the budgets for training. So when we reach out and try to get your business colleagues here, in a lot of companies that's difficult. So, ironically, it's only the IT people who can come here, although in many cases it should be the business managers and the business side people.
The best I can tell all of you is to take back the message that the conference has not been completely technical. In fact, in my mind, many of the sessions have not been technical at all. They have been really addressing — from an analysis point of view and a strategy point of view and a pragmatic point of view — the question of, How do you solve these business problems effectively?
So, please, take the message back and socialize it in your company. And hopefully in the future we will see more of the business side. We'd love to have them here. We try everything we can to get them here. And, yes, there are people from the business-side here. One of them is waving his hand frantically — I know him personally.
Actually, I think our track record on attracting the business side has been pretty good, percentage wise, over the years. I don't know if we did any statistics this year, but I think it's probably about 50-50.
Jim and I were talking a little bit about those numbers before this session. So, Jim, do you want to say something?
At Gartner, we deliberately went after the process crowd because we thought the business people were more accessible. Since we are (in essence) a technology-advising company, we knew that we had to reach the business audience, going forward. So we use 'business process' as the vehicle to get a better audience.
When we had our first BPM conference, we had fifteen percent business users; now we have sixty percent. It's something we deliberately grew. We knew that business people would relate to process very easily ... because a lot of them were operationally oriented.
And I think in today's world they are going to be more interested in doing decisioning. If you take a look at the growth rates of our decisioning-related events, they're growing higher than the business process events now. So I think the worm has turned.
The only thing I would add is, one of the things it's going to take is a change in mindset. When I talk to people about using rules to automate decisions, in some organizations it's the IT department that doesn't want to give any control of the rules over to the business people. And in other places it's the business people who don't want to be put in the position where they can't just throw their spec over the wall and then point fingers.
So I think it will take a group of business people to whom it doesn't occur that controlling their business does not also mean controlling the way their systems behave. Until that kind of generational shift happens we won't have the business more fully engaged. It needs people who are completely used to the idea that their business could only run if their systems are running. They then will find it much less of a stretch to think, "I need to know how to make my systems do what I want them to do," both on a process side and on a decisioning side. That will gradually drive more people to say, "We want to be part of that; we want to participate."
I think what you're saying, James, is that if I'm a homeowner I'm responsible for leaks in the plumbing. It doesn't mean that I know how to fix them, but I am responsible for identifying the problem and making sure that they do get fixed ... and for staying with the problem until they are fixed. Same difference. And it is a cultural change in many organizations to get to that point.
I would add, I think that the proportion of people here that are from an IT perspective versus a business perspective is similar to the proportion in the nature of the tooling that drives a lot of this.
There are a lot of tools being demonstrated here — vendors supporting this. And most of the tooling right now is really at an IT-kind of level. We talk about business rule 'engines' — things that have to be set up by IT kinds of people. So people interested at that level are coming to this conference.
There's a little bit of stuff here that really is meant to talk to the business community. As that grows then I think you will see participation grow there as well. So we need to take decision support tooling to the business perspective, in business language, so that it talks to the business community. Then you will see bigger business participation.
[Ron]: I have a question for the moderator. Is a panelist allowed to ask a question? Can you check the rules on that?
[Kristen]: Oh, I don't know — I don't think John allowed for that in his rules. So it will need an executive decision.
[Ron]: Well, we live in a light world (as SBVR would say), which means that there are no prohibitions unless you say there are.
Okay, so I'd like to ask Don — since he's a world-class engineer and now working for Microsoft: You heard the previous panel. There was a discussion of language and, surprisingly, the discussion seemed to be more focused on the need for business people to be able to speak in natural language, as opposed to sort of the plumbing talk ... to have the deep technology to support rules on the web, and that sort of thing. What is your assessment of the future of natural language (or structured natural language, or controlled language) from what the practitioners/the consumers can expect?
I have really high expectations, as some of you who know me might imagine. I think that we're going to see the ability for business people to state rules in their own language and see automated systems react accordingly. I think we're going to see that more and more.
The ability to ask questions of our automated systems in natural language and get answers back that are precise — I think that we're going to see that.
It's definitely in the future; it's inevitable, whether Microsoft does anything about it, or not. Now, I'm glad to be working at Microsoft and working in that area. Unfortunately, I can't announce a product, or tell you when — they don't let me do that. But like Ron said earlier, go and scream at your reps. There's a squeeze everywhere and budget allocations are going to be based on where the customers are demanding product. So, if you want to accelerate this, the people who are demanding it are going to have control. But, again, I do think it's inevitable.
In terms of how 'controlled' are these natural language subsets? I think initially they are going to be more controlled. It's more to do with natural language in a focused area. That's why, when you're dealing in a narrow business domain where you have a business vocabulary, it's much easier to eliminate ambiguity than it is to, say, deal with casual natural language speech. There are a lot of problems to solve there that you don't need to solve in a narrow environment.
So, you're going to see more controlled in the beginning, and it's going to evolve to more and more casual capabilities, in terms of the natural language processing.
I have a follow-up question.
Okay, but this is your last one.
So there was a rule, you just hid it.
It's called the "cut-off rule."
So, Don, do you think — in your professional opinion — that it's possible to take a controlled language, or natural language subset, like RuleSpeak and generate running code that satisfies faithfully the intent of what the rules are expressing.
I absolutely think so, and (as I said earlier today in another talk) there are two parts to making that happen, realistically.
First, you have to help the business users write it clearly, so that everybody who reads it will understand it in the same way. You can help the business user do that. A lot of business users do not have a knack for that, let's face it, but there are other people who do and they're going to need facilitators to help move forward in this direction initially. Methodologists training them in RuleSpeak (and things like that) will help facilitate this. So, helping people to write it clearly is a first step.
Then, once you have unambiguous rules stated in natural language, are there people who know how to take that and turn it into software? Yes, absolutely. It actually is happening. Some of you have seen me demonstrate some of that last year — generating a natural language query against database; generating constraints; checking for violations; generating derivations; even generating interfaces to systems based on authorization rules. It's all very possible and very doable. It's a whole lot of work — it's hard work. So it's going to take a lot of investment. But it is in our future. There's absolutely no doubt about it.
It's hard work for vendors, not for the business.
Right. This goes back to something Benjamin [Grosof] said. To make it simple for the user means it's gotta be super complicated for the vendor; that's how that works. Take the C programming language. It's a real easy language to compile, and it's a real hard language to deal with in terms of writing very complex systems.
As the language gets more and more natural for the user, the system behind it, that supports it, has got to be very sophisticated. And that's why SBVR — when you get to the area that's not really the business-facing area but the tool-facing area of semantic formulations — gets rather complicated. It had to be because it has to accommodate all the kinds of rules and guidance that actually occur. And SBVR wasn't in the position of deciding what those should be.
For example, if you're creating a programming language, you get to decide what's possible. But if you're trying to deal with business rules you have to deal with all the rules that are out there. There are millions of rules already written, and SBVR is aimed at being able to actually capture the semantics of rules as they are. So it has to have that sophistication inside.
The adoption and the tooling behind it ... it's going to take awhile. It's very complex. But what that does is put something much, much simpler in the hands of the business user.
I'm actually going to take issue with pretty much everything that both of them have just said. Is that okay?
First of all, I'm not sure that the statement "unambiguous rules in natural language" is not an oxymoron. Like "military intelligence" — the idea of something being in natural language (such as English) and being unambiguous.
And the fact that you also immediately talk about SBVR and RuleSpeak, neither of which are natural language, demonstrates that, as an assertion I think. So I think there's a real issue with that.
But I also think that there are a number of underlying problems. Because it's not enough to be able to describe rules in natural language and generate a system from them. You also need to be able to describe all the information you have — its usage and its relationships — in natural language, in order for that to work.
So far, we've made a lot more progress on the data and information side with a more mathematically-oriented representation model. Without that ability to take a fact model and also say how that should be represented in a systems perspective, you can't do the rules piece.
I don't disagree that we will get there. I just think that we're a lot farther from it, for most high throughput production systems, than we would like to be. And I think there is a real danger, having heard Benjamin make more or less the same comments several years ago at a European conference, that we freeze people's willingness to make incremental progress, with "better being the enemy of good." Rules and decision management, the way they work today, are a good solution to the problem. Maybe a natural language decision would be a better one, but how many problems and how much money are we going to spend while we wait for that to be possible?
So I would urge people to "be hopeful" — as I think we all are, since we are here — but also be brutally realistic about what they can get done in the high-performance transaction systems that they are running today.
Okay. I see we have a question at the back.
|[from the audience]: So, if I have a BPEL engine, a human workflow, business rule engine or two, BPM modeler, a BAM dashboard, business intelligence tools, a couple of CEP engines — do I have a good EDM platform? Or is there some secret sauce. And, if so, what's in the sauce — can I make it; can I buy it?|
Were you talking about Oracle's actual product range when you went through that list? <laughter>
From the Audience
Yes ... and then again, no.
Yes, you have all the pieces, but I think in order for it to be a platform for decision management it also requires a focus on decisioning — something that requires those pieces to operate together. I think what you see when you listen to Michelle Edelman from Discover — people who have done a lot of this kind of work — is that they are using a lot of tools. But they've built them up; they've done a lot of work integrating them, making them work, together because many of these pieces have to move together in order to make this all work.
So, yes, you have all the pieces. It would sure be much easier for everybody if there was more of a framework around how to make the decisioning piece hang together.
Kind of like having lots of musical instruments without any music or a conductor.
Yes. So I think, potentially, yes, for sure ... just not necessarily actually.
I think you've got some good pieces but I think that the area for improvement is in support of agility. You've got Business Activity Monitoring, for example, but how much setup did that take and what does it take to change it? Can an executive simply say, "Notify me when ..." and state, in his own language, when there needs to be some kind of notification brought to his attention. Does that need to go to some programming group? Is it going to cost him ten thousand dollars to get that? Or is it just a matter of being able to ask the question?
Can he get a different kind of report of information that's important in a particular problem he's trying to solve right now? Does he have to go to IT, have somebody do some programming, to get that information? Or can he just ask the question?
To support agility — moving in the direction of putting it all closer to the business user and in his terms — I think is going to be a key to improvement. So I think you have the right pieces, but we need to bring it a long way forward toward the business user to create a much more agile business.
James dragged us over to the practical world and I want to come back a step because we got off into platform-land for a second. The real problem in a lot of companies is just a fundamental ability to communicate precisely. And that's a human problem. You're going to try to build process models that cross organizational silos — that's a communication problem; that's a vocabulary problem; that's a clarity and ambiguity problem.
So, yes, it's true, James, that SBVR-style solutions (or other natural language solutions) are going to be extremely difficult and it's going to take some time. But in organizations solving real problems today it means that you'll take on the immediate problems, which are vocabulary and communication and more effective coordination, just as a social activity. Fact models, business vocabulary, and business rules are all very fundamental to getting a handle on this dis-organization that we see in most businesses today with respect to communication.
There are other artifacts besides those, but those are certainly ones you can build on. I think you also need goals and you also need some algorithms that can balance conflicting goals within a common lingua franca. You need to build on top of that.
But I would also agree with James that we're not ready yet. None of the platform vendors is ready yet; neither are any of the business process management suites that I see.
We're all moving in a good direction. Nobody is there yet.
I think Don's comment — that all this needs to move closer to the business — is what it boils down to. We're trying to remove that impedance between deciding that I need to change my business in a certain way (or to do a certain thing) and having my systems be able to do that. We have to take small steps every day.
I agree that it's got to be gradual. That's why I'm saying that there will be more controlled languages in the beginning, moving toward more natural. That gives a chance for the technology to move along with the reality.
On the other hand, I'm not sure what James meant because, based on what he said, it must have had at least two meanings, right? ... because he did say it in natural language.
And we think we understood what he said. Although with James there is sometimes a question. <laughter>
And if the panel will allow, I have to throw in my two cents here.
You asked about secret sauce. I think it's the skilled people who use the tools. They are the secret sauce. They are the ones who are interpreting the results, deciding what to do with things, deciding what needs to be measured, figuring out how these things work together. They are closing the loop back around. So, to me, that is the secret sauce.
Brian Stucky used a great phrase. He talks about "purple people." We know about the red states and the blue ones — then you have the purple ones, places like Florida where they still run political ads. Brian drew the same analogy for people — you have business people and IT people (he didn't allocate either to red or blue). But it's the purple people who are those skilled ones. They are the people who sit at the intersection and can think on both sides of the fence.
There's another question down here.
[from the audience]: I want to make two comments first and then I have a question for Ron.
The secret sauce to EDM is your business because at the root is a business philosophy. Actually, you don't need any of those components — they just make it easier.
And a comment on natural language ... the danger there is that I have yet to meet a business person who thinks about boundary cases. Business people tend to just think about the most common case. So if you make it too easy for the business person he is going to miss the boundary condition, which is often where the guts of the problem is.
Now my question. When it comes to decisioning we've heard over and over that to really make it work it needs to be tied to your business strategy. Well, I've worked at a number of organizations and I've yet to work at one that I think is very good at strategy. Most business guys think strategy is some ethereal thing that really doesn't guide behavior. So my question is, is it really necessary that people do a better job at strategy to make this work? Or do you think a greater focus on decisioning will inherently push business users to think about strategy better and, as a result, that will drive strategy?
I'd like to credit Gladys Lam as being one of the pioneers in the area of bringing strategy — as a structured activity — to business analysis as a front-end system development requirements process ... engaging business people in it. We've found this to be extremely successful, time and time again. Business people love to engage in these dialogs. And it doesn't take as much time as you think if you have the right people in the room. It can't because they have to keep running the business.
So, yeah, maybe awareness from better decisioning will push people toward strategy. But I would hope that we are a little more proactive and that we here who practice analysis — who are the 'purple' people (I keep thinking of purple people eaters) — can bring that to the table as we go back.
I think that during this time people may become more operational, temporarily. While I agree that strategy is the right place to go, we see our clients headed toward 'operations' very quickly, to deal with the survive and thrive.
Who has read the book Execution? Okay, not many. It's an interesting book. It's focused on the fact that far too many executives, exactly as you say, think that strategy is some ethereal thing and never think through how they are going to make it happen, in an execution sense.
I think one of the opportunities with decisioning and with rules is to give people a challenge and an opportunity to do exactly that ... by thinking through, "Okay, well, what would the steps be in my process? What would the rules be in my decisions? How do I in fact decide which customer to keep and which one to dump?"
That's both something they have to do but it's also perhaps something that's never occurred to them that they could do. So I think that's an opportunity.
I'm afraid we've run out of time.
Thank you all very much. This was very informative.
# # #