Showing posts with label Consulting. Show all posts
Showing posts with label Consulting. Show all posts

Thursday, August 8, 2013

Custom Applications vs. Automating a Process with IPA

Considering IPA? Or maybe you're an ININ Reseller looking to automate your first internal process to get your feet wet with the product. If either of these are the case, please read on! I had lunch with one of the IPA folks from ININ a while back and one of the topics we discussed was using Interaction Process Automation to automate a business process vs. building a custom application with it.

What's the difference?

Automate a Business Process
Employee Onboarding
IT Change Requests
Loan Origination
Insurance Application

Build a Custom Application
Case Management / Incident Management, such as a trouble ticketing system
Customer Relationship Management (CRM)
Just about anything where there are multiple off the shelf or cloud-based solutions available

While these two types of solution are similar, it's the areas where they differ that are worth considering before you implement. After Interactions 2013, where I co-presented at 4 IPA sessions and talked to lots of partners, customers and prospects about IPA, it became clear the initial customers deploying fall into one of these two types of solutions and the overall success and customer satisfaction varies depending on what type of solution was deployed.

When it comes to the Automate a Business Process, the best chances for overall success, customer satisfaction, maintainability and ROI present themselves when the customer has an existing data store.

Take AccuQuote, an early customer in the insurance industry, who had an existing brokerage management system. They wrapped IPA around it to great success. No secrets here - they reduced paper, increased efficiency, with all the value you'd expect from IPA. Watch the AccuQuote video and see for yourself. Yes, that's the same Geoff Calhoun who now works for ININ as the product manager for IPA.

Automating a Business Process and Integrating IPA with an Existing Solution
So you have an existing CRM solution, which after a couple of years in production does about 90% of what you need, but with some enhancement you could get to 95% plus for your most critical business functions. Might not be time to rip it out and replace it something new. Maybe there could be measurable gains from automating parts of it with IPA.

What do you gain from automating a business process around another solution?

  1. The Data Store - this is pretty huge. Less need to create a custom database, although you still might want to consider a Database Driven Design for the events when you republish and process and want to restart your running processes.
  2. Web Services - in the relatively modern world of tech solutions, these magical things called Web Services are available or can be implemented relatively cheaply. IPA happens to have an extremely slick Web Services Call Operation tool that does the work in one step for many SOAP-based Web Services. I'm told RESTful is planned, but for now let's just talk what's available.
  3. Reports - if you have a CRM, or as in AccuQuote's case, a brokerage management system, it already has reports that come with it. Remember IPA's reports are more along the lines of call-type stats: handle time, volume, etc., but not necessarily what's going on in the process (who clicked or entered what) outside of the Dynamic Details.
  4. Enhance the existin gsolution - push work, add or simplify business rules as needed and translate them into the other solution. Large implementations of stuff like SAP, Salesforce, etc., could greatly benefit the organization with some automation. Yes, many have workflow engines built in, but can they push and pop work to individuals or workgroups like we can to the Interaction Client with IPA? Why expect one solution to do everything under the sun? I've seen it time and again with customers over the years. AccuQuote used the word synergize. That's fancy-schmancy for integrate. Seamlessly. No solution is perfect. IPA helps make it more perfect, or suck less depending on how dire the situation is...!
  5. Give very limited access to an existing system to other people in the organization who don't need direct access. I built a process for ININ that allows some folks who spend lots of time in Salesforce to summarize and share data with others in the organization who rarely or never use it.
Drawbacks? Pretty much none. Take your existing process, automate it, and make it even more awesome. Probably save some money too. AccuQuote and DefenderDirect are great examples of successful customers who have taken this approach. Double-bonus if you're hooking into an IPA Template you bought from the ININ MarketPlace. Talk about Happy Processing!

Now what's a custom application? Well, as suggested above, pretty much anything that you could buy. I mentioned about a year ago that ININ had commissioned me to implement a Timecard process for their hourly team members, which, while fun, certainly can be sourced. Thankfully that process was relatively small and quick to implement, but it definitely wasn't simple. Remember, there's no such thing as a simple process.

There are a couple of customers who have used IPA to build relatively large custom applications, which sounds a little scary. Because it could be, but don't necessarily let that stop you. One in particular that did essentially case management said they suggested customers start smaller when they spoke at Interactions 2012 last year.

Creating a Custom Application
The biggest benefit of creating a custom application is you have complete control. Make it as small or grand as needed. Don't like the order of the radio buttons? Change'em. Need another value in the drop down list? Add it. Test. Republish. Voila!

What are the potential drawbacks to creating a custom application?
  1. Potentially reinventing the wheel. While you or some of your colleagues might be definitive subject matter experts on trouble ticketing systems, for example, building one from scratch is probably going to take more time and resources than buying one. Why not buy the one that is the best fit and potentially use IPA to enhance a few aspects of it? 
  2. You built it - you and your team (or maybe it's just you?) get to maintain it for the lifecycle of the solution. Now of course you're in this mode with any IPA deployment, just consider the bigger it is, the more complex the process, the more there is to maintain, (hopefully) including your process flows and other documentation. Customers have strategic reasons for Do-It-Yourself solutions and they happen all the time. Does that make them wrong? Depends. It's kinda like the trade-offs of using Attendant vs. custom Handlers for call flow. We'll get to some points to ponder in a minute.
  3. Biting off too much at the start. This could apply to wrapping around an existing application as well, but your risk is even bigger here. You'll have to think the whole data model through, your custom database, and unless you're very controlled about starting small, you'll likely spend weeks to months implementing something that may or may not meet the exact requirements. Use Agile, Scrum or another similar methodology suited to your organization for rapidly deploying smaller-ish, multiple iterations/phases/sprints.
Here's some potential warning signs you might have gone too far down the path with a custom application:
  • Lots of buttons on your Work Item pages that go off and do many complex things, like launch lots of different Handlers or make heavy use of Run Process.
  • Lots of Parallel Splits. Use this feature sparingly and try to resist including it in your first phase.
  • There are many pages of To-Be process flows. How many is too many? Hard to say, but if you have more than a handful or two to start in your first phase, iteration or sprint of your process, you might be biting off more than you could try to chew to start with IPA. 
  • First process is a custom application that's "bigger than a breadbox." Timecard could be a potential exception. It was pretty small. Unless you've already deployed multiple processes in your business, which at this date most have not, you're going to be in learning mode. It's the same learning we go through with Handlers, our call flows, and everything else with the ININ solutions, or most any technology implementation for that matter. Start small. Start with a small iteration (or sprint if you prefer) in your first process. You'l be glad you did. But don't take my word for it - show up at Interactions or ask for a customer reference. There's a learning curve, and I mean that in a good way!
Things careening all over the place, kinda like my old '74 Chevy Caprice when I went a little too fast down a dirt road one time in high school and was over-compensating to get'er going straight again. So let's try to summarize things a bit.

Best Chances for Success with IPA
  • No matter what, smaller processes or a small piece of your process is the best place to start or use an IPA Template from the MarketPlace.
  • Custom Applications - the smaller the better. Acquire an existing solution and enhance it with IPA for best results.
  • Existing Applications - make them more accessible to others in the business, push work and use existing reports when you integrate with IPA.
Got any initial IPA experience to share? Feel free to contact me or drop a comment.

And Happy Processing!







Tuesday, October 9, 2012

Benefits of a Work Items Sprint with IPA

While working on my latest process with Interaction Process Automation, I had a bit of an ah-ha moment I'd like to share: A Work Items Sprint. It worked out great the the first try this past week. We recently finished the front end consulting with sign-off for As-Is, To-Be and the Work Item mock ups. Keep in mind I didn't mention Detailed Requirements doc.

In early discussions with the customer, I had suggested the idea of trying to get something in front of them "sooner than later," but I hadn't gotten any further than the idea stage. Once we got sign off I had briefly started digging into the Detailed Requirements document, filling in the State names and some other obvious details from the flows.

Then last week I had an eureka! moment - why not do a Work Items Sprint? I'm no Agile expert but I am familiar with the basic concepts. I have been trying to consider how this can be accomplished with IPA since John Heiberger (www.lucilium.com) suggested the idea at Interactions2012 back in June during an IPA session. He might not have been the first to suggest it but it's the first I'd heard of it, so great idea, John!

My customer liked the idea of getting to see the screens (which are called Work Items in IPA) and I mentioned it would help drive the Detailed Requirements as well. I also talked it over with the PM, who also has IPA development experience and we were all in agreement that there were no obvious drawbacks and it was \worth a try...

So Monday I dove into IPA and started created a new Process. Keep in mind you can't jump straight into the Work Items, unless of course you don't want to actually be able to Test or Publish the process, but what fun is that? The greatest benefit is seeing the Work Items rendered on the screen because they look slightly different in the Process Automation Designer (PAD).

Turns out you have to build just enough of a shell of your process that it really helps lay the groundwork. Since we'll also be using the Database Driven Design with this process, the Work Items Sprint also required a bunch of variable definitions, which in turn will become columns in the database.

Turns out I didn't need quite the entire week and was able to get the Work Items implemented and a process published in production, which I used to demo for the customer on Friday. They liked it and I learned a bunch as well. Here's the highlights:
  1. Customer gets to see something live almost immediately after sign-off, as opposed to having to wait weeks or months.
  2. The inevitable tweaks are made early in implementation. I noticed a few oversights where I needed to add a simple text label and read-only edit box or a Text Label bound to a variable for display purposes, and adding those drove a few format changes, all of which made for cleaner looking and more intuitive Work Items.
  3. Rapid Implementation - there's no getting bogged down in details with the logic. Just build enough to route the Work Items.
The feedback was very positive from the customer, plus it's always gratifying to see something running in production!

So consider doing a Work Items Sprint on your next IPA implementation. In another post I'll go into technical detail to break down the anatomy of a Work Items Sprint.

If you have other Agile-esque suggestions for Interaction Process Automation, feel free to contact me or drop them in the comments and I'll consider taking them for a spin too!

Happy Processing!