process – Page 40

IBM Impact Day 2: Engage. Extend. Succeed.

Phil Gilbert spoke at the main tent session this morning, summarizing how they announced IBM BPM as a unified offering at last year’s Impact, and since then they’ve combined Business Events and ILOG to form IBM ODM (operational decision management). Business process and decision management provide visibility and governance, forming a conduit to provide information about transactions and data to people who need to access it. IBM claims to have the broadest, most integrated process portfolio, having taken a few dozen products and turned them into two products; Phil was quick to shoot down the idea that this is a disjointed, non-integrated collection of tools, referring to it instead as a “loosely coupled integration architecture”. Whatever.

Around those two core products (or product assemblies) are links to other enterprise tools – Tivoli, MDM, ECM and SAP – forming the heart of business processes and system orchestration. In version 8 of BPM and ODM, they’ve added collaboration, which is the third key imperative for business alongside visibility and governance.

We saw a demo of the new capabilities, most of which I talked about in yesterday’s post. For ODM, that included the new decision console (social activity stream, rules timeline) and global rules search. For BPM, there’s the new socially-aware process portal, which has been created on their publicly-available APIs so that you can roll your own portal with the same level of functionality. There’s searching in the process portal to find tasks easily. The new coach (UI form) designer allows you to create very rich task interfaces more easily, including the sidebar of task/instance details, instance-specific activity stream, and experts available for collaboration. They’ve incorporated the real-time collaboration capabilities of Blueworks Live into the BPM coaches to allow someone to request and receive help from an expert, with the user and the expert seeing each other’s inputs synchronously on the form in question. Lastly, Approve/Reject type tasks can be completed in-line directly in the task list, making it much faster to move through a long set of tasks that require only simple responses. He wrapped up with the obligatory iPad demo (have to give him credit for doing that part of the live demo himself, which most VPs wouldn’t consider).

The general session also included presentations of some innovative uses of BPM and ODM by IBM’s customers: Ottawa General Hospital, which has put patient information and processes on an iPad in the doctors’ pockets, and BodyMedia, which captures, analyzes and visualizes a flood of biometric data points gathered by an armband device to assist with a weight loss program.

IBM Vision for BPM, ODM and SOA

Opening day at IBM Impact 2012 (there were some sessions yesterday, but today is the real start), and a good keynote focused on innovation. The wifi is appalling – if IBM can’t get this right with their messages about scalability, who can? – so not sure if I’ll have the chance to post any of this throughout the day, or if you’ll get it all when I get back to my hotel room.

This post is based on a pre-conference briefing that I had a week or two ago, a regular conference breakout session this morning, and the analyst briefing this afternoon, covering IBM’s vision for BPM, ODM (decision management) and SOA. Their customers are using technology to drive process innovation, and the IBM portfolio is working to address those needs. Cross-functional business outcomes, which in turn require cross-functional processes, are enabled by collaboration and by better technical integration across silos. And, not surprisingly, their message is moving towards the Gartner upcoming iBPMS vision: support for structured and unstructured process; flexible integration; and rules and analytics for repeatable, flexible decisions. Visibility, collaboration and governance are key, not just within departmental processes, but when linking together all processes in an organization into an enterprise process architecture.

The key capabilities that they offer to help clients achieve process innovation include:

Process discovery and design (Blueworks Live)
Business process management (Process Server and Process Center)
Operational decision management (Decision Server and Decision Center)
Advanced case management (Case Manager, which is the FileNet-based offering that not part of this portfolio, but integrated)
Business monitoring (Business Monitor)

Underpinning these are master data management, integration, analytics and enterprise content management, surrounded by industry expertise and solutions. IBM is using the term intelligent business operations (which was front and center at Gartner BPM last week) to describe the platform of process, events and decision, plus appropriate user interfaces for visibility and governance.

Blueworks Live is positioned not just as a front-end design tool for process automation, but as a tool for documenting processes. Many of the 300,000 processes that have been documented in Blueworks Live are never automated in IBM BPM or any other “real” BPMS, but it acts as a repository for discovering and documenting processes in a collaborative environment, and allowing process stakeholders to track changes to processes and see how it impacts their business. There is an expanded library of templates, plus an insurance framework and other templates/frameworks coming up.

One exciting new feature (okay, exciting to me) is that Blueworks Live now allows decision tasks to be defined in process models, including the creation of decision tables: this provides an integrated process/decision discovery environment. As with process, these decisions do not need to become automated in a decision management system; this may just document the business rules and decisions as they are applied in manual processes or other systems.

Looking at IBM BPM v8, which is coming up soon, Ottosson took us through the main features:

Social collaboration to allow users to work together on tasks via real-time interactions, view activity streams, and locate experts. That manifests in the redesigned task interface, or “coach”, with a sidebar that includes task details, the activity stream for the entire process, and experts that are either recommended by the system based on past performance or by others through manual curation. Experts can be requested to collaboration on a task with another user – it includes presence, so that you can tell who is online at any given time – allowing the expert to view the work that the user is doing, and offer assistance. Effectively, multiple people are being given access to same piece of work, and updates made by anyone are shown to all participants; this can be asynchronous or synchronous.
There is also a redesigned inbox UI, with a more up-to-date look and feel with lots of AJAX-y goodness, sorting and coloring by priority, plus the ability to respond to simple tasks inline directly in the inbox rather than opening a separate task view. It provides a single task inbox for a variety of sources, including IBM BPM, Blueworks workflows and Case Manager tasks.
Situational awareness with process monitoring and analysis in a performance data warehouse.
Mobile access via an iOS application that can interface with Blueworks Live and IBM BPM; if you search for “IBM BPM” in the iTunes app store (but not, unfortunately, in the Android Market), you’ll find it. It supports viewing the task list, task completion, attach documents and add comments. They are considering releases the source code to allow developers to use it as a template, since there is likely to be a demand for a customized or branded version of this. In conjunction with this, they’ve released a REST API tester similar to the sort of sandbox offered by Google, which allows developers to create REST-based applications (mobile or otherwise) without having to own the entire back-end platform. This will certainly open up the add-on BPM application market to smaller developers, where we are likely to see more innovation.
Enhancements to Process Center for federation of different Process Centers, each of which implies a different server instance. This allows departmental instances to share assets, as well as draw from an internal center of excellence plus one hosted by IBM for industry standards and best practices.
Support for the CMIS standard to link to any standard ECM repository, as well as direct integration to FileNet ECM, to link documents directly into processes through a drag-and-drop interface in the process designer.
There are also some improvements to the mashup tool used for forms design using a variety of integration methods, which I saw in a pre-conference briefing last week. This uses some of the resources from IBM Mashup Centre development team, but the tool was built new within IBM BPM.
Cloud support through IBM SmartCloud which appears to be more of a managed server environment if you want full IBM BPM, but does offer BPM Express as a pre-installed cloud offering. At last year’s Impact, their story was that they were not doing BPM (that is, execution, not the Blueworks-type modeling and lightweight workflow) in the cloud since their customers weren’t interested in that; at that time, I said that they needed to rethink their strategy on this and and stop offering expensive custom hosted solutions. They’ve taken a small step by offering a pre-installed version of BPM Express, but I still think these needs to advance further.

WebSphere Operational Decision Management (ODM) is a integration/bundling of WebSphere Business Event Manager and ILOG, bringing together events and rules into a single decision management platform for creating policies and deploying decision services. It has a number of new features:

Social interface for business people to interact with rules design: decisions are assets that are managed and modified, and the event stream/conversation shows how those assets are being managed. This interface makes it possible to subscribe to changes on specific rules.
Full text searching across rules, rule flows, decision tables and folders within a project, with filtering by type, status and date.
Improved decision table interface, making it easier to see what a specific table is doing.
Track rule versions through a timeline (weirdly reminiscent of Facebook’s Timeline), including snapshots that provide a view of rules at a specific point in time.
Any rule can emit an event to be consumed/managed by the event execution engine; conversely, events can invoke rulesets. This close integration of the two engines within ODM (rules and events) is a natural fit for agile and rapid automated decisions.

There’s also zOS news: IBM BPM v8 will run on zOS (not sure if that includes all server components), and the ODM support for zOS is improved, including COBOL support in rules. It would be interesting to see the cost relative to other server platforms, and the compelling reasons to deploy on zOS versus those other platforms, which I assume are mostly around integrating with other zOS applications for better runtime performance.

Since last year’s big announcement about bringing the platforms together, they appear to have been working on integration and design, putting a more consistent and seamless user interface on the portfolio as well as enhancing the capabilities. One of the other analysts (who will remain nameless unless he chooses to identify himself) pointed out that a lot of this is not all that innovative relative to market leaders – he characterized the activity stream social interface as being like Appian Tempo three years ago, and some of the functionality as just repackaged Lombardi – but I don’t think that it’s necessarily IBM’s role to be at the very forefront of technology innovation in application software. By being (fairly) fast followers, they have the effect of validating the market for the new features, such as mobile and social, and introducing their more conservative customer base to what might seem like pretty scary concepts.

Operational, Transformation and Technical Roles for Successful BPM Projects

My first post from the Gartner BPM conference in Baltimore this week; to be honest, there hasn’t been all that much so far that has inspired me.

Michelle Cantara, who focuses on organizational issues related to BPM, spoke about the roles required for successful BPM projects. She had four main points:

It’s difficult to find BPM-skilled resources: there’s a shortage of skills not just within some organizations, but in the market in general. In many cases, BPM is not a full-time job for someone within an organization, but a set of skills that they need to apply in the context of other work.
Not all projects require the same BPM skills sets. If you look at the Gartner BPM “sweet spot framework”, which is a quadrant with frequency of process change (low to high rate) along the horizontal axis and responsibility for process change (IT to business) along the vertical axis, they show three main usage scenarios for BPM (the lower left quadrant is considered not necessarily a good fit for BPM). In the upper left quadrant, for example, where BPM is used for standardization and manageability, process visualization may be a key skill that is not as important in the other quadrants. In the upper right quadrant, which is the sweet spot for BPM, round-tripping capabilities and sophisticated process governance are important skills. In the lower right quadrant, where BPM is used for IT agility, model-driven development and other technical skills are most important.
Those skills sets are not necessarily what you already have in house for other projects, although if you take a look around, you might find some of the required skills in unexpected places. For example, influential and collaborative people can be leveraged for transformational skills, while technical skills may be coming from the business community. There is a wide variety of skills and skill levels across each of transformational, operational and technical skills.
Transformational and operational capabilities are critically important; too many organizations focus purely on technical skills.

She showed an ideal BPM organization:

A business process competency center (aka COE) reporting to an executive steering committee, and containing a business process champion, business process director, business process consultants, business process analysts and business process architects.
An enterprise architecture program office, also reporting to the executive steering committee (not IT), working together with the BPCC.
The process owner, who is responsible for improving business processes related to the corporate strategy.
One or more BPM project teams, informed by the BPCC, and including an executive sponsor (linked to the process owner) and a BPM project manager.

The most critical roles are the process owner (a senior businessperson responsible for the end-to-end process performance), the business process director (business or IT background, but with a blend of all skill types, responsible for leading the BPCC and ensuring its adoption throughout the organization), the business process architect (business or IT background, a transformational role that liaises between EA and BPM, and establishes BPM governance and standards) and the business process analyst (business or IT background, responsible for modeling, designing and documenting processes).

She also defines a business process consultant as a senior person who can fill competency gaps in transformational, operational and technical skills; although I often work in the business process architect role for clients, this is probably a better description of what I do, as much as I dislike the word “consultant”.

She walked through five scenarios for how BPM is used in organizations, outlining some tips and pitfalls for each. For example, in the “visualize and rethink the process” scenario that fits into the category of using BPM for standardization and manageability, she mapped out the key skills required from each of the roles, such as process modeling for the BP analyst.

She finished up with some best practices for hiring and managing BPM resources, including:

Overhire, since senior, experienced people can play multiple roles
Establish decision guidelines so that it’s clear which types of changes need to be approved by whom
Align to key business goals
Leverage executive influence to get things done expediently

Overall, some good pointers for any organization implementing BPM.

Linking External Social Presence To Core Business Processes

I gave the Monday afternoon keynote at Appian World 2012 last week on the topic of the impact of social technologies on the enterprise, with a particular focus on how social features and exposure are changing our business processes. I’ll post the entire presentation online – I’m thinking of recording a re-creation of my presentation and syncing it with the slides – but I wanted to focus on one slide that I feel is at the heart of what we should be doing with social enterprise.

I was discussing how important it is to link an organization’s external social presence with their business processes, and gave two examples from my own experience as a consumer: one good, one not so good.

The first case – the good one – was about Zipcar. Since I live in downtown Toronto, where the transit is good and the Zipcars are plentiful, I don’t own a car, but rely on rentals. Zipcar is perfect for me: book online, show up and have immediate access to the car via an RFID card, and rent only for an hour or two if that’s what I need. One of the rules of being a Zipster is that if the car has 1/4 tank of fuel or less, you fill it up before you return it; there’s a credit card included, so it costs nothing but a few minutes of your time and a bit of inconvenience. Some people, however, treat their Zipcar like their mom’s car, and bring it back with an empty tank and potato chip crumbs all over the seat. Heading out to the CASCON conference early one morning last November, I hopped in my Zipcar for an 80km round-trip drive only to find the tank uncomfortably low on fuel. At my lunch break, I started searching for a gas station close to my destination, and tweeted my exasperation:

In spite of my mistyping of “talk” for “tank”, the next day, I received an email from a Zipcar representative:

I received word that you experienced a low gas situation with yesterday’s reservation with Civic Champion. Please accept a 2hr driving credit along with my apologies for the inconvenience this caused.

That means that a number of things had to have happened:

Someone at Zipcar was monitoring Twitter for mention of their @Zipcar handle. (pretty standard)
Someone read my tweet and decided that it needed action. (good, but also fairly standard)
Someone figured out who I am based on my Twitter account, then linked that to my Zipcar account. (OMG amazing, and only possible because I tweet using my real name)
Someone decided that my complaint had merit, and gave me two hours of driving credit (total awesomesauce, not just for the two free hours, but for what it implied about their business processes)

All of this happened within 24 hours of the rental, which means that there is some sort of link baked in between their social presence (Zipcar Twitter account) and their internal business processes (complaint handling and customer refund).

When I told this story during my keynote, there was an appreciable murmur in the audience: everyone there understood the magnitude of what Zipcar had done to link their external social presence to their internal business processes, and in what must be a fairly efficient manner.

The second case was a perfect example of how to screw up that link, and likely is closer to what most organizations are still doing. The night that I was in DC giving the keynote, I had tickets for a concert at Roy Thomson Hall, a large concert venue in Toronto. A few weeks before, when I realized that I was not going to get home in time, I tweeted to @RoyThomsonHall asking if a refund or exchange was possible; a day and a half later, they tweeted back “please call the box office and speak with one of our customer service representatives directly”. Quite excited, I thanked them and told them that I would do that, then wasted several minutes on the phone only to find out that Roy Thomson Hall doesn’t give refunds or exchanges. I responded back on Twitter with:

Twitter silence ensued.

Comparing them with Zipcar, this is how RTH links their external social presence to their internal business processes:

Someone at RTH was monitoring Twitter for mention of their @RoyThomsonHall handle. (pretty standard)
Someone read my tweet and decided that it needed action. (good, but also fairly standard)
Their social media team gave me misleading instructions that are completely disconnected from their actual business processes. (WTF?)
I wasted my time figuring out that either they employ a bunch of morons to handle their Twitter account, or someone had a massive misunderstanding about how to handle customer requests for refunds and exchanges. (#fail)
They ignore my final tweet wherein I inform them that they wasted my time. (also #fail, but at this point, expected)

With a monopoly on certain concerts in Toronto, I don’t expect RTH to be as customer-focused as a more competitive industry such as rental cars, but I didn’t expect their social media team to be so clueless about their actual revenue-generating business processes and the customer service that supports it.

The point of these two stories is not that I love Zipcar and am not quite so fond of Roy Thomson Hall’s exchange policy, but that there is a right way and a wrong way to connect your social media presence to your core business processes. Complaints and requests arriving via social media need to be quickly passed along to customer support teams in the same manner as if they had arrived by email, phone, fax or letter, and handled with the same expediency and professionalism. Technology helps a lot with this – software is available to monitor social media channels for mentions of your company or products, and it’s pretty easy to feed those events into a BPMS or CRM system – but you need to figure out how to triage and filter those events so that your core business processes aren’t swamped by off-topic chatter by anonymous contributors.

Social media for business isn’t a game, and it’s not something just for PR or marketing to worry about: it’s a channel of customer engagement. If you do it wrong, it’s not only going to be noticed by your customers, but they’re going to make sure that their followers notice it, too.

Taming Big Process With @passion4process

I have to depart Appian World at the morning break today to catch a flight home, but first I caught up with Clay Richardson of Forrester and his keynote on the future of process as we move from isolated process improvement projects to what he calls “big process”: a more holistic approach to process improvement and transformation. He described four dimensions:

Customer – bringing the customer into the process and transform the customer experience. Greater customer satisfaction isn’t just about having happier customers, it drives revenue growth, but it requires more than just putting a pretty (read: iPad) interface on the front of the process; it requires transformation of the underlying processes that support the customer experience. Starting with the simpler environment of the mobile experience, rethink processes based on the customer journey and needs.
Chaos – taking advantage of change. He claims that the round-trip is dead: what gets executed is so complex that you shouldn’t even be attempting to model it, but use dynamic case management instead where you can combine some structured fragments with the ability for the knowledge worker to decide what to do next. Social BPM provides guidance throughout process discovery and execution, and analytics monitor what’s happening during execution and push into process optimization.
Context – putting processes in context, leveraging big data to drive process transformation by connecting data directly to the value stream. As I discussed yesterday in my keynote, it’s about monitoring the flood of events, pinpointing and responding to emerging issues before they become problems. Context also means handling localization and regionalization of processes and data, but exploiting the benefits of globalization to service customers anywhere and find the resources required to manage processes.
Cloud – accelerating development and reducing cost by moving some functions to the cloud; this doesn’t mean moving everything to the cloud, but leveraging it as a platform in a hybrid strategy.

He walked through a couple of case studies, then discussed how to build a plan for dealing with big process.

First, embrace “big process” methods and techniques, adding value stream analysis, value chain analysis and business capability maps to other more granular views of process such as BPMN and Lean/Six Sigma.

Next, build process architect skills that connect data and process.

Last, adopt new tools and architectures that support target operating models, linking data and process to support customer experiences.

MDM and BPM and finally coming together, and not a moment too soon.

CEO Keynote At Appian World 2012

Matt Calkins, CEO of Appian, spoke about how they are achieving their goal to be the world’s best way to organize work.

Key features that they have to support this:

Native mobile capabilities on iOS, Blackberry and Android, meaning that you can develop your applications once and have it run not just on a desktop browser, but on any mobile device.
Transparent platform portability, allowing an Appian application to be easily moved between on-premise, public cloud and private cloud.
Social interface to minimize training and be able to more easily track events, primarily through a participatory event streaming paradigm.

Their software sales increased in 2011 by over 200% with 95 new customers (not just expansions in existing customers), and they have 95% “very satisfied” customer satisfaction rating.

The typical Appian customer runs about 10 applications, but Appian’s goal is to actually reduce the number of applications that a customer has (who wants more apps in their enterprise, after all?) by linking the data, actions and users in the application silos into a common environment. In fact, their theme for this year is data, which they see treated as a second-class citizen in many systems, and he switched over to a demo of the upcoming Appian 7 to show how they are combining data from multiple applications and sources into their event stream.

The new Appian interface is organized into five tabs:

News, which is the familiar event stream, but with the much richer links and attachments from other sources. Adding a comment to the stream can be just a comment, or can be turned into a task that can be assigned and tracked. What do I need to know?
Tasks, which are the tasks sent to the active user, or that they created and assigned to someone else. These can be filtered by type, can can be sorted by deadlines and priorities. What do I need to do?
Records, aka data, which shows a list of data sources: client records, support tickets, employees, whatever is important to this user. These may be Appian applications or external data sources such as relational databases. He drilled into the Clients data source, which provided several ways to filter and search in the client application, then selected one client to show the collection of information about that client: basic contact data, sales satisfaction survey results, ticket history, sales opportunities and more. The interface is customizable both during the initial setup, but also on the fly by any user with permissions to do so. Beyond the summary page for that client, there’s a news feed for all items tagged with this client, then links for each of the applications that might have information on that client: projects, billing history and sales opportunities. A related records tab allows connections to other data sources, such as support cases, that are linked to that client, allowing you to navigate through a web of data in your enterprise, much as we navigate the internet by following links on a whim. Lastly, a related actions tab allows you to launch any of the related applications for this client, such as starting a new contract or schedule an onsite visit.
Reports, showing enhanced reporting capabilities with new abilities to sort and customize reports.
Actions, which links to all applications in the enterprise, allowing you to launch any application from a single point.

Furthermore, there are new Facebook/Twitter-like functions to subscribe to people within your organization, see their profile information that they have posted including their job skills, and add kudos (LinkedIn-like recommendations) for individuals. This is similar to what IBM has been doing with their Beehive social network internally: it’s a way to enable collaboration within the enterprise as well as tracking of employee skills and recommendations. In order to have this sort of enterprise-wise social network, however, everyone needs an Appian license, so they are coming up with new licensing model for what they’re calling Appian tempo that will allow this type of access to social, mobile and data (but not actions) at a much lower cost than a regular Appian user. In fact, it’s free, if you have any other Appian (paid) licenses, and if you install and use it within a year.

As always, pretty innovative stuff coming from Appian.

Delivering Enterprise Performance Breakthroughs With BPM

Daryl Plummer of Gartner gave the opening keynote at Appian World 2012 stressing an “extreme” approach for achieving breakthroughs rather than just the incremental improvement that we’re seeing through current best practices. The forces to achieve this approach are social, mobile, cloud and information, combining to create a key emerging pattern of extreme collaboration. He presented 7 key aspects of extreme collaboration: mobile, ad hoc relationships, dynamic communities, proactive notification, free flow of information, people-centric, and multiple media. He spoke about using extreme collaboration to improve outcomes, not just optimize processes.

Much of this is about what mobile technology enables: empowering the workers at the edges of an organization as well as customers by allowing access to core business processes anywhere, anytime, on any device. He sees BPM’s greatest contribution as changing behavior through extreme collaboration, not process improvement. Traditional BPM creates barriers, but extreme collaboration can leverage natural relationships. Engagement methods are changing to include gamification, crowdsourcing and dynamic communities.

Gartner predicts that through 2016, organizational politics will prevent at least one-third of BPM efforts from moving beyond one-off projects to enterprise-wide adoption, based on 53% of BPM survey respondents stating that organizational politics is the #1 barrier to BPM adoption. This is not a big surprise to me; my keynote this afternoon is about how social technologies are going to change the way that you run your business, which will echo some of the themes in Plummer’s presentation.

Kofax Transform Keynote: Martyn Christian and Anthony Macciola

The morning keynote came after the first round of breakout sessions, with Martyn Christian (CMO) and Anthony Macciola (CTO) providing some history of Kofax and its uses, plus a lot more about where it’s moving from a technology standpoint and how customers are using it. This is my last session of the conference since I have an early afternoon flight, so provided a good summary and wrapup to my Transform visit.

Their current message is around capture-enabled BPM: combining the Kofax content capture processes with the TotalAgility (Singularity) dynamic case management capability to improve the capture and processing of information. Capture becomes the driver and initiation of the business process, rather than something separate that happens before the process begins; in some cases, recognition can be sufficient that human intervention is removed and the whole business process is really just the capture and recognition process. When I first was introduced to Kofax, around 1989, it was all about scanning with a bit of recognition; now, it’s about all sort of content ingestion and recognition to reduce or eliminate human involvement in non-value-added tasks such as document classification and indexing. As greater integration with TotalAgility occurs, I expect that we’re going to see human intervention during content capture to be handled more as ad hoc or unstructured exception cases in that environment rather than the more traditional Kofax structured transactional environment.

Point of origination capture is also an increasing important part of their strategy, and was already evident from the breakout sessions that I’ve been attending over the past 1.5 days: control of MFPs and mobile devices through KFS, plus content capture capabilities on customer-facing portals. I’m seeing this trend in all areas of business process: customer-facing employees and customers want more functionality on the device of their choice, and IT can no longer specify (for example) that document capture is only done in centralized scanning facilities by data entry operators.

They walked through the case management capabilities coming in through the Singularity acquisition: I expect that some number of Kofax customers will be considering using that for at least the front end of their business processes rather than their existing BPM systems, since you can set up cases to do things such as define the documents required for a case and prompt for the missing documents. If TotalAgility is integrated pre-committal (that is, before documents are moved from Kofax Capture to a content repository such as SharePoint or FileNet), this has the capability to assemble the complete case of documents before committal for lower latency. Of course, there are many situations where you don’t want to do that pre-committal, for example if the documents need to be more widely available, but I can think of a number of use cases where assembling the complete set of documents is something that needs to be done interactively with the front-line capture user, and a round-trip to the ECM platform would just slow things down.

Is Kofax going to capture 35% of the BPM market, just like they have 35% of the capture market? No way. But can they make the front-end of document-centric processes much more capable and flexible, and likely eliminate some integration to heavier BPM platforms for case creation/completion? Absolutely.

They covered some of the emerging Kofax analytics tools – content extraction analytics, content analytics, traditional business intelligence and visualization, and predictive process analytics – and although they don’t see themselves as an analytics vendor, they provide capabilities for either standalone analytics into your Kofax-based processes, or driving into more comprehensive analytics solutions.

The keynote moved on to a panel at that point, but I had to duck out to catch my flight – check out the Transform Twitter stream for more updates on what they talked about.

Kofax Mobile Capture

I arrived at this session a few minutes late and ended up standing in the back of the room: it was a completely packed house, which is a good indication of the popularity of the idea of using a mobile device for point of origination capture. There are a lot of use cases for mobile capture, ranging from mobile mortgage brokers to home healthcare specialists to claims adjusters to proof of delivery forms in transportation; mobile capture eliminates the paper movement and also greatly reduces the time required to capture these documents directly into a business process.

The goal for mobile capture is not just to take a photo of a document – the regular camera app on your mobile device can do that – but to capture a process-ready document and push it directly into the front end of the business process, just as if it were being scanned. That requires some degree of smarts built into the mobile app for capture: ensuring that the entire document is being captured using visual framing and minimize movement during capture, some post-capture processing, and confirmation that the capture was successful. For the most recent generation of smartphones, including the iPhone 4S, that means running Kofax VRS right on the phone; older phones have to send the images off to a VRS server in the cloud before receiving confirmation of a “good” capture (suitable for passing on for back-end recognition) 3-5 seconds later. The correction that can be done on the device is pretty impressive: segmentation to separate the document from the background (dependent on good contrast), keystone correction for when the camera is not square to the document, dynamic thresholding to maximize readability throughout the page, rotation, cropping, color correction/removal (including removing colored highlights without losing solored markings elsewhere on the page), and shake/blur correction. In a near-future release, it will also be possible to do barcode recognition right on the phone, too.

The phone app downloads batch class definitions and other server-based parameters the first time that it connects, and during document synchronization, but otherwise can perform completely offline. Obviously, on a lower-end phone that needs to use VRS in the cloud, the offline capability won’t be as great. Documents are uploaded to the Kofax cloud server on demand, using SSL on the communication but no image encryption.

The app allows the user to track and add to open cases, or create a new case based on a batch class. This dictates the expected document types and the indexing fields; the user may be prompted to enter in index information, or this may be left as a back-end function. Everything passes from the cloud through Front Office Server (KFS) on its way to Capture, which means that the type of restrictions that can be placed on MFPs – as I reviewed yesterday – can be applied to mobile devices, too.

The iPhone 4S’ 8MP camera takes photos at 2,448 x 3,264 pixels; if you could perfectly frame an 8.5×11” page during capture (which you can’t), that would give you a resolution of about 290dpi. Count on losing at least 10% around the edges of your image during capture, so a letter-sized page would be captured at about 250dpi. Adequate for OCR of standard printing, but you’re not going to capture tiny fonts. Also, the user experience of capturing a lot of pages using a phone is not really the best, so use cases with higher resolution or higher volume requirements are probably going to be better served by a mobile scanner. The first version of the app will support document and photo modes for capture; future releases will expand this to allow video and audio to also be captured.

They will also publish their SDK in the future to allow organizations to create their own mobile app based on the Kofax technology, allowing organization-specific security and functionality to be included. I expect that this will be extremely popular with bigger organizations who want to control the user experience more tightly.

I had seen a bit of this from Kofax previously, but great to see more of the detail, and I’ll be looking forward to trying it out or seeing a live demo when they release the mobile application on iOS and Android in Q2. Windows Phone will follow later, but Blackberry currently isn’t on the roadmap unless demand increases.