ABBYY mobile real-time recognition

Dimitry Chubanov and Derek Gerber presented at the ABBYY Technology Summit on ABBYY’s mobile real-time recognition (RTR), which allows for recognition directly on a mobile device, rather than just capturing content to pass on to a back-end recognition server. Mobile data capture comes in two basic flavors: first, the mobile user is just entering data, such as an account number or password; and second, the mobile user is entering both data and image, such as personal data and a copy of their ID.

ABBYY RTR isn’t based on taking a photo and then running recognition on that image; instead, it uses several frames of image from the camera preview stream and runs recognition algorithms on the stream without having to capture an image. This provides a better user experience since the recognition results are immediate and they don’t have to type the data manually, and better privacy since no image is captured to the phone or passed to any other device or server. They demonstrated this using a sample app on an iPhone; it’s interesting to see the results changing slightly as the phone moves around, since the recognition is happening using the previous several frames of video data, and it gradually gains recognition confidence after a few seconds of video. We saw recognition of unstructured paragraphs of text, drivers licenses, passports and bank cards. The SDK ships with a lot of predefined document types, or you can create your own by training for specific fields using location and regular expressions. They are also offering the ability to capture meter data, such as electricity meters, although some of this requirement is being by smart meters and other IoT advances.

They also have a mobile imaging SDK that can capture an image when it’s needed — for proof of ID, for example — with scene stabilization, document edge detection, deskewing and various types of image enhancement to capture the optimal photo for downstream storage and processing.

I can imagine, for example, a mobile airline app that needs to capture your passport information using mobile RTR to grab the data directly rather than having you type it in. I’ve also seen something very similar used to capture the unique number from an iTunes gift card directly into the App Store on an iPhone. Just like QR code reading is now built right into the search bar on the mobile versions of Google Chrome, and Google Translate on mobile allows real-time capture of text using the same camera preview mode (plus simultaneous translation), being able to capture text from a printed source instead of requiring a mobile user to type it in is likely to become ubiquitous in mobile apps.

bpmNEXT 2015 Day 2 Demos: Kofax, IBM, Process Analytica

Our first afternoon demo session included two mobile presentations and one on analytics, hitting a couple of the hot buttons of today’s BPM.

Kofax: Integrating Mobile Capture and Mobile Signature for Better Multichannel Customer Engagement Processes

John Reynolds highlighted the difficulty in automating processes that involve customers if you can’t link the real world — in the form of paper documents and signatures — with your digital processes. Kofax started in document scanning, and they’ve expanded their repertoire to include all manner of capture that can make processes more automated and faster to complete. Smartphones become intelligent scanners and signature capture devices, reducing latency in capture information from customers. John demonstrated the Kofax Mobile Capture app, both natively and embedded within a custom application, using physical documents and his iPhone; it captures images of a financial statement, a utility bill and a driver’s license, then pre-processes them on the device to remove irregularities that might impact automated character recognition and threshold them to binary images to reduce the data transmission size. These can then be directly injected into a customer onboarding process, with both the scanned image and the extracted data included, for automated or manual validation of the documents to continue the process. He showed the back-end tool used to train the recognition engine by manually identifying the data fields on sample images, which can accept a variety of formats for the same type of document, e.g., driver’s licenses from different states. This is done by a business person who understands the documents, not developers. Similarly, you can also use their Kapow Design Studio to train their system on how to extract information from a website (John was having the demo from hell, and his Kapow license had expired) by marking the information on the screen and walking through the required steps to extract the required data fields. They take on a small part of the process automation, mostly around the capture of information for front-end processes such as customer onboarding, but are seeing many implementations moving toward an “app” model of several smaller applications and processes being used for an end-to-end process, rather than a single monolithic process application.

IBM: Mobile Case Management and Capture in Insurance

Mike Marin and Jonathan Lee continued on the mobile theme, stressing that mobile is no longer an option for customer-facing and remote worker functionality. They demonstrated IBM Case Manager for an insurance example, showing how mobile functionality could be used to enhance the claims process by mobile capture, content management and case handling. Unlike the Kofax scenario where the customer uses the mobile app, this is a mobile app for a knowledge worker, the claims adjuster, who may need a richer informational context and more functionality such as document type classification than a customer would use. They captured the (printed and filled) claims form and a photo of the vehicle involved in the claim using a smartphone, then the more complete case view on a tablet that showed more case data and related tasks. The supervisor view shows related cases plus a case visualizer that shows a timeline view of the case. They finished with a look at the new IBM mobile UI design concepts, which presented a more modern mobile interface style including a high-level card view and a smoother transition between information and functions.

Process Analytica: Process Discovery and Analytics in Healthcare Systems

Robert Shapiro shifted the topic to process mining/discovery and analytics, specifically in healthcare applications. He started with a view of process mining, simulation and other analytical techniques, and how to integrate with different types of healthcare systems via their history logs. Looking at their existing processes based on the history data, missed KPIs and root causes can be identified, and potential solutions derived and compared in a systematic and analytic manner. Using their Optima process analytics workbench, he demonstrated importing and analyzing an event log to create a BPMN model based on the history of events: this is a complete model that includes interrupting and non-interrupting boundary events, and split and merge gateways based on the patterns of events, with probabilistic weights and/or decision logic calculated for the splitting gateways. Keeping in mind that the log events come from systems that have no explicit process model, the automatic derivation of the boundary events and gateways and their characteristics provides a significant step in process improvement efforts, and can be further analyzed using their simulation capabilities. Most of the advanced analysis and model derivation (e.g., for gateway and boundary conditions) is dependent on capturing data value changes in the event logs, not just activity transitions; this is an important distinction since many event logs don’t capture that information.

bpmNEXT 2014 Tuesday Session: It’s All About Mobile

I’ll blog this year the same as last year’s bpmNEXT demos, with each session of multiple demos in a single post. The posts are a bit long, but they are usually grouped into themes so it works better that way.

First up was Brian Reale of Colosa (makers of ProcessMaker open source BPM and ProcessMapper) on self-organizing groups, ad hoc work and expectations of simplicity. This is a topic that I’m really interested in, since I’ve been presenting on worker incentives with collaborative work, which includes some of the same issues as self-organization. One of his keys points is about the effort required to start using a typicial BPMS, and how that differs from design time (where there is typically a large degree of effort required and very little organic adoption) to runtime (where there is much less effort and is the main target of ROI). What they are trying to do is increase adoption by reducing the effort required at design time by providing more ad hoc capabilities, with a resultant lower ROI but also lower cost.  The result is FormSlider, an app environment for ad hoc workflow of structured data with minimal setup, which is what Brian demonstrated (still in alpha). He demoed the tablet interface for a loan application that allows for mobile capture of a client requesting a loan, including pictures and signatures, which then interfaces with ProcessMaker or other back-ends. More interestingly, he showed how an easily-setup app can be used for mobile data capture that hte user can then route to whomever they want (possibly limited to a selection list) with a few other fields such as due date and priority. There’s some informational context, such as seeing how long it is taking each of the possible participants to process cases, and also allows for routing to be round-trip or one-way. The standard user interface is pretty simple: My Cases for things that I’m working on, an Inbox for new things, and a simple forms interface for working on items. There’s an historical view of cases, showing the participants and their responses. He demoed a simple flow going through a round-trip from the initiator through two people and back to the initiator; this can be used for adding a collaborative workflow on top of existing pre-defined processes and systems, taking the place of emailing around for approvals and other simple collaboration. He finished up the demo in ProcessMaker showing us how an app and forms are created and deployed in a few minutes, including how potential users and groups are associated with the forms as they are designed. They have email and forum connectors for ProcessMaker and will be using the same methods with FormSlider for providing people with ways to be notified about work but also to interact with it directly.

Next up was Romeo Elias of Interneer on extending enterprise software using mobile apps by using BPM, addressing the issue that many companies have of not having skilled mobile app developers, but there being no commercial apps available for their needs. Their Intellect BPMS has mobile app capabilities, and allows custom mobile apps to be built quickly that can connect directly to the back-end processes. Since BPMS’ are often being used as full application development platforms, this is not that much of a stretch: the BPM platform already has a lot of the integration and other capabilities, and Interneer’s platform is intended to be used mostly in a drag-and-drop model-driven development environment. Romeo demonstrated creating a new application template that consisted of laying out a UI form for the mobile app using the full web interface (there could also have been a process attached, but the point of his demo was to show the mobile UI), then using it as an app on a tablet interface. The design interface on the web provides the ability to specify sidebar content as well as multiple pages (shown as tabs in the designer). The resultant app – immediately available as soon as it is created in the designer – is a native mobile app, not viewed through a mobile browser, so can take advantage of device-specific features as well as cache data offline. The app was a mobile data capture/reporting application that connected to a database; he demonstrated adding records to the table that include text (free text and restricted using a selection list) and a photo field, with any new records stored locally if connectivity is lost.

Scott Francis and Greg Harley of BP3 presented on bringing process to the people using their  Brazos mobile BPM responsive UI toolkit; at the time of last year’s bpmNEXT, they were focused on hybrid mobile apps, but now are directed towards responsible UI, that is, applications that run in a browser but behave appropriately regardless of the form factor of the device. Native apps can cause a lot of problems because of lack of mobile development and deployment skills within enterprises, but also the hurdles that many companies have to go through to deploy a mobile app that connects to their enterprise apps. Conversely, many enterprise applications already have web interfaces, so adding a new web UI that happens to be responsive and hence appropriate for mobile devices may have a much shorter adoption path, and less effort required since there’s a single application to design and deploy for any platform: no specialized mobile browser apps versus desktop browser apps. Plus, they’re giving it away for free, with plans to open source it in the future. Greg demoed a UI for an IBM BPM process in the full desktop browser version, then the same form on a phone (simulator). The same features in the full form are available in the mobile version, just resized and reformatted for the smaller screen in either orientation. He showed a bit of the form designer, although I had the sense that this would take a bit more effort than what we saw in the previous two demos but would offer quite a bit more capability. They support IBM BPM and Activiti BPM (which are the two platforms that BP3 supports in its consulting practice) and can be made to work with pretty much any BPMS that has a REST API since those APIs turn out to be surprisingly similar between different BPMS vendors. If you want to try out the Brazos UI toolkit, they have a sandbox where you can try it out running against an Activiti instance. This is quite the opposite in technology strategy from Interneer: I can understand BP3’s motivation for going with responsive UI, as well as the rapid uptake, but can also understand the challenges of a browser-based app when you have spotty connectivity (as I often do when I’m travelling), and they admittedly give up some of the device-specific capabilities.

We’re heading off to dinner, then back with a last demo (which was aborted from this session due to projector difficulties) and a keynote by Jim Sinur before we get down to the serious business of the evening drinks reception.

TIBCO TUCON2012 Day 2 Keynotes: Consumerization of IT

We’re on the last of the four themes for TUCON, with Matt Quinn kicking off the session on the consumerization of enterprise IT. It’s a telling sign that many vendors now refer to their products as being like “Facebook/Twitter/iTunes/<insert popular consumer software here> for the enterprise” – enterprise app vendors definitely have consumer app envy, and TIBCO is no exception. As we saw in the earlier keynote session about tibbr, TIBCO is offering a lot of functionality that mimics (and extends) successful consumer software, and their providing Silver Mobile Server as a way to put all of that enterprise functionality that you build using TIBCO products out onto mobile devices. We saw a demo of a app that was built using Silver Mobile Server for submitting and managing auto insurance claims, and it looks like the platform is pretty capable both in terms of using the native device capabilities and linking directly to back-end processes.

They showed some new enhancements to Silver Fabric for private cloud provisioning and management, and discussed their public cloud applications (tibbr, Spotfire, Loyalty Lab) and public Silver Marketplace functionality. Today, they announced Silver Integrator, running on Silver Marketplace, providing enterprise-class integration services on the public cloud. We saw a brief demo of Silver Integrator: it launches a cloud version of the TIBCO Designer with some additional palettes for cloud connectors such as Facebook, Salesforce and REST services.

Being able to extend enterprise applications onto mobile devices and into the cloud are critical capabilities for consumerization of IT, and TIBCO (as well as other vendors) are offering those capabilities. The problem of adoption, however, is usually not about product capabilities, it’s about organizational culture: there is a lot of resistance to this trend not in the user community, but within IT. I saw a graphic and blog post by Dion Hinchcliffe of Dachis Group today about the major shifts happening in IT, and one thing that he wrote was especially impactful:

Never in my two decades of experience in the IT world have I seen such a disparity between where the world is heading as a whole and the technology approach that many companies are using to run their businesses.

<p>Cloud, mobile, social, consumerization and big data: we’re all doing it in the consumer space, but IT departments are continuing to drag their feet at bringing this into the enterprise. The organizations that fail to embrace this are going to fall further behind in their ability to serve customers effectively and to innovate, and will suffer for it.</p>  <p>Quinn wrapped up with a list of their product announcements, including two new policy-driven governance products as part of the ActiveMatrix suite: AMX Service Gateway for providing enterprise services outside the firewall, and AMX Policy Director for managing security, auditing and logging rules for services. He covered their AMX BPM 2.0 release announcement briefly, with new functionality for work assignment and scheduling, case management, and page flow debugging, plus some new capabilities in Nimbus Control and FormVine.</p>  <p>That’s it for the keynotes, since the rest of today and all of tomorrow is breakout sessions where we can dig into the details of the products and how they’re being used by customers. I’ll be heading to the BPM product update this afternoon by Roger King and Justin Brunt, and probably also drop in on the tibbr product strategy to see more details of what we heard this morning.</p>  <p>I’ve been asked to step in for a last-minute cancellation and will be <a href="">presenting tomorrow morning at 10am on the process revolution</a> that is moving beyond implementing BPM just for cost savings, but looking at new business process metrics such as maximizing customer satisfaction. If you’re here at TUCON, come on out to the session tomorrow morning.

TIBCO Corporate and Technology Analyst Briefing at TUCON2012

Murray Rode, COO of TIBCO, started the analyst briefings with an overview of technology trends (as we heard this morning, mobile, cloud, social, events) and business trends (loyalty and cross-selling, cost reduction and efficiency gains, risk management and compliance, metrics and analytics) to create the four themes that they’re discussing at this conference: digital customer experience, big data, social collaboration, and consumerization of IT. TIBCO provides a platform of integrated products and functionality in five main areas:

  • Automation, including messaging, SOA, BPM, MDM, and other middleware
  • Event processing, including events/CEP, rules, in-memory data grid and log management
  • Analytics, including visual analysis, data discovery, and statistics
  • Cloud, including private/hybrid model, cloud platform apps, and deployment options
  • Social, including enterprise social media, and collaboration

A bit disappointing to see BPM relegated to being just a piece of the automation middleware, but important to remember that TIBCO is an integration technology company at heart, and that’s ultimately what BPM is to them.

Taking a look at their corporate performance, they have almost $1B in revenue for FY2011, showing growth of 44% over the past two years, with 4,000 customers and 3,500 employees. They continue to invest 14% of revenue into R&D with a 20% increase in headcount, and significant increases in investment in sales and marketing, which is pushing this growth. Their top verticals are financial services and telecom, and while they still do 50% of their business in the Americas, EMEA is at 40%, and APJ making up the other 10% and showing the largest growth. They have a broad core sales force, but have dedicated sales forces for a few specialized products, including Spotfire, tibbr and Nimbus, as well as for vertical industries.

They continue to extend their technology platform through acquisitions and organic growth across all five areas of the platform functionality. They see the automation components as being “large and stable”, meaning we can’t expect to see a lot of new investment here, while the other four areas are all “increasing”. Not too surprising considering that AMX BPM was a fairly recent and major overhaul of their BPM platform and (hopefully) won’t need major rework for a while, and the other areas all include components that would integrate as part of a BPM deployment.

Matt Quinn then reviewed the technology strategy: extending the number of components in the platform as well as deepening the functionality. We heard about some of this earlier, such as the new messaging appliances and Spotfire 5 release, some recent releases of existing platforms such as ActiveSpaces, ActiveMatrix and Business Events, plus some cloud, mobile and social enhancements that will be announced tomorrow so I can’t tell you about them yet.

We also heard a bit more on the rules modeling that I saw before the sessions this morning: it’s their new BPMN modeling for rules. This uses BPMN 1.2 notation to chain together decision tables and other rule components into decision services, which can then be called directly as tasks within a BPMN process model, or exposed as web services (SOAP only for now, but since ActiveMatrix is now supporting REST/JSON, I’m hopeful for this). Sounds a bit weird, but it actually makes sense when you think about how rules are formed into composite decision services.

There was a lot more information about a lot more products, and then my head exploded.

Like others in the audience, I started getting product fatigue, and just picking out details of products that are relevant to me. This really drove home that the TIBCO product portfolio is big and complex, and this might benefit from having a few separate analyst sessions with some sort of product grouping, although there is so much overlap and integration in product areas that I’m not sure how they would sensibly split it up. Even for my area of coverage, there was just too much information to capture, much less absorb.

We finished up with a panel of the top-level TIBCO execs, the first question of which was about how the sales force can even start to comprehend the entire breadth of the product portfolio in order to be successful selling it. This isn’t a problem unique to TIBCO: any broad-based platform vendor such as IBM and Oracle have the same issue. TIBCO’s answer: specialized sales force overlays for specific products and industry verticals, and selling solutions rather than individual products. Both of those work to a certain extent, but often solutions end up being no more than glorified templates developed as sales tools rather than actual solutions, and can lead to more rather than less legacy code.

Because of the broad portfolio, there’s also confusion in the customer base, many of whom see one TIBCO product and have no idea of everything else that TIBCO does. Since TIBCO is not quite the household name like IBM or Oracle, companies don’t necessarily know that TIBCO has other things to offer. One of my banking clients, on hearing that I am at the TIBCO conference this week, emailed “Heard of them as a player in the Cloud Computing space.  What’s different or unique about them vs others?” Yes, they play in the cloud. But that’s hardly what you would expect a bank (that uses very little cloud infrastructure, and likely does have some TIBCO products installed somewhere) to think of first when you mention TIBCO.