≡ Menu

More thoughts on RuleBurst and Haley


Having posted some initial thoughts on RuleBurst’s acquisition of Haley, I was lucky enough to get some time with Peter Still, VP Strategy. Peter and I spent an interesting hour discussing the merger and the combined companies plans so I thought I would share some of my thoughts.

The first interesting thing to note is that, although RuleBurst has bought Haley, the new company will be adopting the Haley name. Initially in the US but then worldwide. Given the size of the US market, this probably makes sense, as does breaking the implicit tie of the name to just rules. Decisioning is increasingly perceived as needing more than just rules so a name that leaves that open clearly makes sense.

As you would expect, the plan is to merge the two products over time while supporting upgrade paths for both. While both products are “natural language” based they do bring some different strengths to the party. RuleBurst is strong in using office documents for editing and supports multiple languages while Haley is based on a conceptual model and has a strong repository. As always the question is how quickly can they bring the products together without abandoning either user community – particularly important when both companies have a significant user base. Details of the roadmap were, as you would expect at this early stage, sketchy.

One of the key differences between their approach and most other rules vendors is that they do not start with a business object model. They like the way that their approach allows mapping to an execution model later. Not only does this allow development of expert systems for interactive diagnosis (with no database involved at all), it also allows mapping to multiple systems later. They also have some customers using the natural language approach to capture rules at a source rule level with implementation to come later. As a result some of their deployments involve tight integration with production systems, some do not. They have even done some work plugging the RuleBurst front end onto someone else’s backend such as their ability to generate rules for Microsoft’s emerging rules engines. This is an area that intrigues me, though I remain to be convinced. I am hoping to get a demonstration of how this will all work together sometime soon and will blog again then.

It seems that the merged company will have an increasing focus on business solutions that include rules rather than the product as a pure platform play. Examples Peter gave included social security, tax and benefits in the public sector (where RuleBurst has a strong presence, especially in Europe/Asia/Australia), fraud detection, claims assessment, underwriting and other financial services and banking solutions (more I think from the Haley side). While this is a trend amongst most rules vendors – adding vertical content – this was one of the strongest positions on this I have seen taken by a rules vendor. It will be interesting to see how this plays in the general technology market, though I think their choice of focus areas is spot on (I might have added marketing/customer service decisioning). One difference is that they see increasing value from adding vertical content to both rules and related technologies. For instance, RuleBurst’s framework for compliance has workflow, neural networks (for fraud detection) and rules. This seems like the model for the kind of solution-oriented approach they want to take going forward. Despite the inclusion in these frameworks of analytics and BPM, they don’t see this as excluding work with other analytic/BPM vendors so much as part and parcel of delivering a complete solution. Indeed they tout pre-built integration with BEA Aqualogic, IBM Websphere Process Server, Tibco, Netweaver and Biztalk.

One of my big questions was around the use of Haley in Oracle Siebel CRM. Siebel 8 includes Haley for configuring some of the behavior of the CRM component and customers can (and do) build on the OEM license to expand rules use beyond CRM. Despite Oracle’s known opposition to OEMing technology in their applications, Peter assured me that Oracle is continuing to sell the solution aggressively and to sell consulting packages around adopting this component. Given the need for Haley consulting support I am sure Oracle is pleased with the merger as the combined company can offer consulting worldwide much more effectively. That said, I am still not convinced that Oracle will continue to rely on an OEMed component. RuleBurst is confident that Oracle have no plans to drop it, despite its somewhat anomalous situation as the only OEM component in the Oracle stack. RuleBurst’s focus with Oracle is on business content around CRM rather than on integration with Oracle’s development tools such as their BPEL tool with its APIs for Decision Services.

Meanwhile they also have a good relationship with SAP from the Ruleburst side, especially in Public services. Despite the fact that they are one of many partners they feel that they have a special relationship and presented some compelling evidence that this, in fact, the case. They don’t feel that the SAP purchase of Yasu (on which I commented here) as part of the Netweaver platform is a threat to this relationship thanks again to a focus on business content (this time in public sector). They remain confident that they can continue to work with both Oracle and SAP – I wish them luck!

In my original post I asked three questions:

  • Will the combination be able to deliver the features now common in Business Rules Management Systems and Natural Language Processing?
    This remains to be seen but Peter seemed confident they can and that their unique approach offers benefits
  • Will the two companies different geographic focuses become an asset, making them a worldwide player, or a liability, by creating tensions between different parts of what should now be a single development team?
    The adoption of the Haley name and the acknowledgment that merging the products sooner rather than later is desirable along with Peter’s discussion of the relative benefits of the two technologies makes me think they have a shot at this.
  • Will the combined company focus on decisions not rules
    While they do have many customers who use the engine like this and support having the engine deployed as a service, it is not the only way to use it. Although, therefore, you can build decision services that are widely accessible as a stateless component that can receive data, return decisions and explanations, flag additional required data etc., the company seems like it will remain split between this view and a more traditional expert-system-like view of rules.

So we will all have to wait and see what happens but consolidation is a sign of market maturity I think so I wish them, and the folks at Yasu/SAP, the best of luck.


Comments on this entry are closed.

  • Paul Vincent November 29, 2007, 12:14 am

    Its curious that the first “business rules” standard to be commissioned by a bona fide standards authority (ie OMG SBVR) addressed natural language specification of rules, without any participation (AFAIK) from these natural-language-approach BRE vendors. And generally the government-based market is very sensitive to standards compliance…

    I wish Softlaw/RuleBurst/Haley good luck, too!

  • Paul Haley November 29, 2007, 8:38 am

    I don’t think that Ruleburst will have problems with the first of your bullets any longer, James. Haley has met the requirements of customers or Oracle, Countrywide, United Healthcare, Medco, etc. in high-value and highly-competitive sales. Haley has been an effective competitor within its limited presence (due to marketing and business development that have been weak relative to Ruleburst). With regard to natural language, Mr. Still has communicated a significant but distinguishing charactaristic of both companies that is a further synergy – the emphasis on a smooth flow from pre-IT capture, harvesting, and formalization of business requirements, regulations and policy through to implementation rather than publishing an IT-centric model through a more business accessible front-end (i.e., for rule authoring). These synnergies, one additive and the other emphatic and overlapping bode well for the combination.

    With regard to OMG SBVR, Paul, you know I am a big fan. Authority was pretty far ahead of where SBVR landed up, imo, but SBVR is critically correct in separating semantics and logic from implmentation. Now all we need to do for SBVR is add langauge capabilities (e.g., real dictionaries and linguistic flexibility) plus logic generation for Rete Algorithm rules engines or logic programming systems…

    One more thought on SBVR, Benjamin Grosof pinned the vendor panel at the Business Rules Forum with a question about interchange (e.g., using OMG PRR or W3C RIF). All of the vendors dismissed these standards as being inadequate for interropability given their higher level differences. I suspect that SBVR will be a more practical interroperability mechanism. Specifically, I think it is interesting to migrate between or from vendors via SBVR, including to open-source/free engines, such as Drools.

Next post:

Previous post: