Robotic process automation software is a stop-gap measure for a business’ integration problems and serves as a bridge to digital transformation but falls short of being transformational on its own.
That’s the view of Francis Carden, vice president of digital automation and robotics at Pegasystems Inc., a Cambridge, Mass., company in the digital process automation (DPA) space. Carden’s view runs counter to organizations that have made RPA the cornerstone of their digital transformation efforts. For example, CGS, a business process outsourcing provider based in New York, views RPA as a key vehicle for digital transformation. R.R. Donnelley & Sons Co., a Chicago-based communications company, is pursuing RPA as part of its digital revolution initiative. Numerous other enterprises have embarked on a similar path with robotic process automation software.
Carden, however, believes RPA is bumping into limitations as adopters attempt to scale initial deployments. Among those are cost and ROI. Some organizations are initially attracted to RPA as a simple and inexpensive way to integrate systems. But as bots proliferate, the need to manage them grows. An enterprise might create an RPA center of excellence and institute other governance mechanisms to keep initiatives on track. But adding those elements requires investment.
“The problem you then have is your costs go up,” Carden said. “They [RPA adopters] get lulled into a false sense of security that it is cheap.”
More effort than expected
A Pegasystems’ survey of decision makers using RPA, published Sept. 10, found organizations “spending more time and effort getting bots up and running than anticipated.” Half of the 509 survey respondents said software bots were harder to deploy than expected and 41% of those polled said bot management was taking more time and resources than expected.
RPA’s return on investment, meanwhile, may fail to justify the cost, Carden suggested. For some projects, a complete overhaul of a business process would make more economic sense than RPA, he said.
“There is a cross-over point that varies,” he said or RPA efforts. “If the ROI is significant, then go for it. If it’s small, then for the upfront dev cost, ongoing maintenance [and] governance … looking at a re-engineered process becomes more viable. It’s not a one size fits all, but if the cost of tactical RPA gets closer to digital transformation, who wouldn’t want to transform?”
And then there’s the issue of a software bot’s lifespan. Changes to an organization’s underlying enterprise architecture will likely result in an increase in software bot breakage, the Pegasystems survey noted. Eight-seven percent of the survey respondents reported some level of bot failure, although most of those reporting breakage described the problem as small to moderate in scope. Organizations, on average, believe bots will last about three years, according to the survey.
Carden said the breakage will accelerate. He believes RPA’s fragility stems, in part, from accretions of automation over time. Enterprises originally automated manual processes, an era he referred to as “computerization.” Subsequent efforts — from Excel macros to present-day RPA — have sought to automate those computer-based processes, he said.
“RPA is the automation of computerization,” Carden said, describing RPA as another layer of automation amid other attempts to stitch together old, computerized processes. “At some point of time, the house of cards must collapse,” he said.
Beyond robotic process automation software
Carden clearly has a horse in the automation race. His company, Pegasystems, views DPA as going beyond RPA to provide an “end-to-end strategy for digital transformation.” In this vision, RPA plays a role in transformation, but a transitional one.
Carden outlined the following scenario: An organization decides to “go digital” with a new customer onboarding system, spanning mobile and web engagement. Such a customer deploys RPA as fast way to tackle the necessary integrations with backend systems. RPA thus provides a bridge to digital transformation, but a temporary one. After the onboarding system goes live, the organization can, over time, descale the RPA installation, replacing bots with “true API integration into the backend systems.”
This approach makes robotic process automation software a short-term expediency. “RPA as a permanent solution for integration is madness,” Carden contended.
RPA, DPA and other emerging fields such as intelligent process automation, which blends RPA with AI, are all vying for enterprise and channel partner attention. The next couple of years will reveal which technology, or combination of technologies, gain the upper hand as the preferred tools of digital transformation.
Enterprise digital transformation projects are snagging on organizational issues and technology infrastructure, opening opportunities for consultative IT services companies.
A batch of reports published this week point to the perils of the digital era. The research reveals a landscape in which digital transformation projects frequently fail and highlights a list of obstacles including network shortcomings, legacy database limitations and spotty top-management support for initiatives. The pressure to transform — coupled with the range of technology options available to do so — complicates matters, as IT leaders and their business counterparts search for a way forward.
“Enterprises are saying today they have more options than ever before and are confused about which way to go,” said Peter Finter, CMO of Couchbase, a NoSQL database provider that conducted a CIO survey to gauge the state of digital transformation. “The role of the consultants is becoming more and more important as they are asked to make recommendations on this plethora of alternatives.”
Eighty-one percent of the respondents to Couchbase’s survey reported a digital transformation project failure, significant delay or scaled-back initiative in the last 12 months. The company polled 450 executives heading digital transformation projects.
The Couchbase survey, published Sept. 4, suggests the urgency of digital transformation and a lack of forethought may be hampering projects. The poll found most organizations pursuing transformation are doing so reactively. The top drivers are responding to competitors’ advances (23%), pressure from customers for new services (23%), responding to regulatory changes (23%) and pressure from the C-Suite (22%). Proactive ideas from within the business influence only 8% of digital transformation projects, according to Couchbase.
The reactive obsession with digital transformation is such that 71% of the respondents believe “IT teams risk working on projects that may not actually deliver tangible benefits,” the survey reported.
“[Enterprises] are doing it because they have no option but to do it,” Finter said. “That is worrisome.”
Also troubling: 52% of digital transformation efforts are the IT shop’s sole responsibility despite the business-wide implications of success or failure.
“IT is still holding the steering wheel of the majority of digital transformation projects,” Finter said. “They need the support of the rest of the board room and the C-Suite.”
A digital transformation strategy revolves around change management and business transformation, so business leaders need to work alongside the technologists, he added.
Digital transformation in business: Tech limitations
Technological issues also contribute to digital transformation difficulties, with network infrastructure one source of concern.
An Accenture survey found many enterprises adopting digital technologies, but far fewer that were completely confident in their networks’ ability to cope with business demands. The survey of 300 senior IT and business executives revealed high adoption rates for IoT/edge computing (77%), big data/analytics (83%) and digital customer experience (78%). But only 36% of the respondents reported being “very satisfied” that their networks have the necessary capabilities to support business needs, according to survey results published Sept. 4.
“I think the infrastructure and network modernization is as important as any of the major digital investments a company makes,” said Charles Nebolsky, managing director and global leader of Accenture’s Network Business.
The networking challenges point back to organizational conflicts. Forty-eight percent of the Accenture survey’s respondents cited “misalignment between IT and business needs” as among the main obstacles. Bandwidth and performance demands outpacing networks’ ability to deliver ranked among the technology-based problems, with 45% of respondents citing that issue.
Nebolsky said the network infrastructure problems stem, in part, from the practice of replacing network devices as they fail rather than investing in software-defined networking and an enterprise platform with built-in automation, analytics and security.
A network approach designed to “keep the lights on” won’t equip an organization to support large numbers of remote employees, IoT initiatives, a growing roster of SaaS applications and corporate systems running on various public clouds, he suggested.
Network outages can bring production to a halt in companies dependent on cloud systems. Nebolsky cited the example of a pharmaceutical company that couldn’t ship products because of a WAN outage that took down its cloud-based labeling system.
Networks don’t have to fail entirely to jeopardize business operations and transformation efforts, however. Indeed, network brownouts have surfaced as a challenge for enterprises. A Netrounds survey found 83% of large organizations have reported economic damage and employee frustration stemming from those events, which the company describes as unexpected and unintentional drops in network quality. Netrounds, which published its findings Sept. 3, polled 400 end user, network operations and C-level respondents in the U.S.
Technologies that enable digital transformation may exacerbate the brownout issue. Netrounds, based in Stockholm, said the adoption of technologies such as hybrid cloud, SD-WAN and network virtualization will likely lead to a sharp rise in the cost of brownouts.
“If the network platform is not running, the business cannot run,” said Cyril Doussau, CMO of Netrounds.
Databases are another technology roadblock in the path of digital transformation projects, Couchbase’s Finter said. He said the database is that last piece of IT yet to be fully transformed. “We are finding moving to a modern database is a transformation effort in its own right,” he said.
Finter pointed to some bright spots in digital transformation, one of which is increased levels of investment. That boost underscores the higher priority of projects.
“The Amazon effect is causing business leaders to realize that digital means disruption,” he said. “Their competition is just one click away. People are now willing to spend more money on these projects.”
As IT vendors struggle to modernize their channel partner programs, partner firms of all types should be actively weighing in.
That’s according to speakers at CompTIA Inc.’s ChannelCon conference, held in Las Vegas from Aug. 5 to 7. In a session led by Thom DeVos, vice president of client services at channel strategy consultancy JS Group, and Heather Murray, vice president of security solutions at distributor Tech Data Corp., it was made clear that traditional partner programs can’t cut it anymore. The channel ecosystem has become more complex and sophisticated than ever before, rendering conventional programs defunct. Vendors increasingly realize this, but few know how to modernize their partnering approach.
“What we are seeing is this big evolution where vendors are coming to us saying, ‘We know we need change. … What do we do?’ ” DeVos said.
He said channel partner programs must be flexible enough to accommodate modern buying and selling behaviors. That might mean pivoting from the revenue goals that programs conventionally focus on. “The old programs … are not going to work, because as everything goes to ‘as a service,’ ” conventional revenue goals stop making sense for partners.
Modern partner programs may also have to drop or deemphasize traditional partner incentives, including sales performance incentive funds (SPIFs) and rebates. Murray said partners are more interested in resources such as demand generation.
“I hear it all the time from partners: ‘SPIFs are too hard to manage. Don’t give me SPIFs. Let’s figure out a way you can invest in my company, ‘ Murray said.
She encouraged channel firms to tell their vendors what is and isn’t working for them. “The partners are the ones that know what they need, and they should … speak up. … For the small business, the medium[-sized] business, the MSPs [managed service providers], the MSSPs [managed security service providers], there are different models and there are different needs, and without your voice being heard, our vendors aren’t going to make the changes needed.”
Devos agreed, saying that partners should realize that they are in the power position in their vendor relationships. He said partners should take a hard look at their line cards and determine which vendors truly provide “the opportunity to better serve my customers.” Partners should also think long-term about their vendors’ strategies and areas they are investing in for growth. “Take this time to … strengthen or terminate the relationship. … Use that power position, because it is only going to make the ecosystem better as a whole,” he said.
Once vendors take into account the needs of different partner types, they can develop a partner program framework that serves each channel segment uniquely, Murray noted.
“If you are the vendor, you have to build [a partner program] that works for everybody. But you really don’t. You don’t have to. You can build out different types of programs that will work for pockets of partners … and [help] enable them,” she said.
Fresh off of its purchase of Aerohive Networks, Extreme Networks revealed it will double-down investments in the channel.
The Aerohive acquisition, which closed Aug. 9, expanded Extreme’s portfolio of wireless LAN technology but also brought in a raft of new partners, including managed service providers (MSPs). With that influx, Extreme’s partner roster now stands at about 11,000 firms. About 86% of Extreme’s business goes through the channel, according to the company.
“Basically, we are inheriting a channel that is focused on managed services, which is great, because that is exactly the way we see the market going and the need for customers is for managed service,” said Gordon Mackintosh, vice president of worldwide channels of Extreme, based in San Jose, Calif.
Mackintosh said Extreme and Aerohive’s partner programs will run separately for the foreseeable future. In part, this decision is to minimize disruption to partners. He said Extreme eventually will look to integrate “the best of both programs to create a new unified partner program.”
What Extreme, Aerohive partners can expect in FY20
For Extreme’s fiscal year 2020, kicked off in July, Mackintosh said the company will build on a partnering approach that it developed last year. In fiscal year 2019, Extreme selected a group of about 200 partners globally, 100 of which were in the U.S., to devote extra resources and support. The focused strategy resulted in selected partners growing 10% faster than other Extreme partners, he said. Extreme now plans to expand that group to about 200 U.S. partners and increase its dedicated resources by about 50%.
“More partners will be able to experience a much more focused and hands-on approach from Extreme,” he said.
Extreme plans to roll out new tools and support to its partner base as a whole. One of the tools will be Quick Quote, a tool that lets partners build quotes online and send them out to distribution, Mackintosh said. The vendor will also expand its online education platform, Extreme Dojo, with more training.
Another focus area for Extreme will be on new customer acquisition, he said. Extreme will continue to support partners in its key vertical markets, two of which — K-12 and retail — Aerohive partners already target. Mackintosh said the Aerohive acquisition “helped to solidify our position in those verticals.”
Aerohive also takes Extreme into the midmarket and SMB space. With Aerohive’s portfolio, partners can offer “a new turnkey technology for SMBs, branch retail or schools … Those customers will be major targets with the Aerohive acquisition as it is today but even more so” as Extreme integrates with the acquired technologies, Mackintosh said.
More partner resources in development
Mackintosh noted several other plans in the works. Partners that struggle with adopting a managed services business model can expect to see a new MSP program in late October. The program will offer business transformation and consulting services.
Extreme Networks also aims to help partners overcome staffing shortages, a consistent problem caused by a widespread dearth of IT skills. While still in concept form, Extreme is considering launching a program that would quickly train job-seekers who are early in their careers. The company’s program would then place those students in positions either at Extreme or partner businesses.
“We are going to be leveraging our global network of training partners as well as our online training capabilities within Extreme Dojo to target early career graduates,” Mackintosh said.
A number of traditionally hardware-focused IT vendors now pursue software, services and consumption-based pricing. Cisco provides an example in the networking field, making its first move to sell a switch as a subscription in 2017. Server-maker HPE in June announced plans to deliver its entire hardware lineup as a service by 2022. Storage vendors are moving in this direction as well: the NetApp subscription model provides one example.
NetApp channel executives in March discussed the company’s hardware-to-software transition and partner enablement strategy. Mathew Chacko, director, channel sales, consumptive solutions at NetApp, and Mara McMahon, global principal consultant, Fueled by NetApp, provided additional details in a statement describing the company’s subscription strategy and how partners fit into this approach.
NetApp’s strategy is built around its portfolio of managed and unmanaged consumptive offerings associated with the company’s private and public cloud solutions.
“NetApp, as a whole, believes that these offerings enable partners to embrace a land-adopt-expand-renew model in subscription-based selling,” according to Chacko and McMahon.
With land, adopt, expand, renew — or LAER — the seller seeks to create a long-term relationship with the buyer. The “land” step represents the initial sale, while “adopt” covers activities designed to ensure customers successfully use the offering. This phase is particularly important with cloud-based, subscription software and services — organizations are unlikely to renew an offering users don’t embrace. The “expand” step encompasses up-selling and cross-selling initiatives that extend use. All the steps point to the goal of customer renewal.
NetApp uses LAER in its partner program, with support coming from the company’s Fueled by NetApp (FBNA) program. FBNA equips NetApp partners with best practices for go-to-market strategy and planning. FBNA covers pricing, packaging, defining use cases, creating service-level agreements, messaging, sales training, demand generation and market awareness, the company executives noted. They added that the best practices create a foundation for selling subscription services.
NetApp subscription model for public, private clouds
NetApp’s lineup of consumption-based offerings gives partners various entry points to customer sales. NetApp’s Cloud Volumes OnTap, for instance, offers a way to land deals in the public cloud, putting partners in a position to help customers as they migrate to AWS, Microsoft Azure or Google Cloud Platform, according to NetApp. Cloud Volumes OnTap offers SnapMirror replications for disaster recovery, which NetApp said is a typical first step in a customer’s cloud journey.
In keeping with the NetApp subscription model, Cloud Volumes OnTap is available as an hourly pay-as-you-go subscription or as a one-, two-, or three-year subscription.
Meanwhile, NetApp’s hyper-converged infrastructure (HCI) offering, HCI Cloud Consumption, lets partners land customers in a private cloud setting, NetApp noted. The offering is available as an annual subscription.
“More and more partners are establishing private managed cloud and multi-cloud support services themselves based on NetApp technology,” according to Chacko and McMahon.
At the adoption step, the NetApp partner shifts to helping customers migrate production workloads to the public cloud. In private cloud environments, partners help customers incrementally deploy “right-sized compute and storage nodes” to boost adoption, according to NetApp.
NetApp partners have a few options for expanding upon the initial sale. For public cloud use cases, for example, partners can tap NetApp’s Cloud Data Services portfolio, which includes Cloud Sync for migration and Cloud Insights for hybrid multi-cloud monitoring and optimization. And integration between Cloud Data Services and NetApp’s HCI platform lets partners extend the company’s private cloud offerings to enable hybrid multi-cloud data management, according to NetApp.
Another expansion opportunity: Partners can position NetApp Kubernetes Service for application development on the company’s HCI platform or in the cloud.
“This would enable partners to attach high-value services focused on application lifecycle management in a hybrid multi-cloud environment,” the NetApp officials said.
To complete the LAER sequence, NetApp offers what the company describes as an “increasingly simplified renewal plan” that aims to help partners quickly finalize deals.
Remaining trusted advisors
The LAER approach within the NetApp subscription model intends to help partners avoid churn and encourage customers to renew subscriptions. Successful adoption and the resulting renewals cement the partners’ trusted advisor role among customers.
This model, according to NetApp, “provides NetApp partners with an opportunity to generate multiple, recurring revenue streams that have the potential to increase over time.”
Organizations adopting DevOps principles and moving toward the new world of cloud-native applications may be overlooking an important element in this modernization push: the data problem.
That’s the thinking of Jeff Bozic, principal architect in Insight Enterprises’ cloud and data center division. Data management, he said, involves a host of issues, not the least of which is the basic question of where does the data exist in a complex hybrid IT environment.
“It’s a big challenge and it’s only going to get worse as we create more data,” Bozic said.
Data problem: key factors
A number of factors contribute to the data problem. Bozic cited the DevOps movement, public cloud adoption, microservices-based applications residing in containers, new data sources such as IoT streams, and the pressure to create dynamic apps that react faster to user needs.
Those considerations put a strain on traditional data flows, databases and extract, transform, load (ETL) processes, according to Bozic. The newer development techniques, application models and performance expectations raise data-centric questions for organizations.
“How do these data structures need to start changing, and how do the databases need to start changing?” Bozic asked “How do I protect that data? Maybe my traditional data protection solution may not make sense. The disaster recovery process may not make sense. I want apps to be portable to run in different clouds; data may have to start moving. How do I know where it is, and how do I follow it?”
Blithely acquiring more up-to-date database technology — NoSQL offerings versus traditional relational databases, for example — may not be the answer, however. Similarly, organizations that immediately attack data structure as a point problem, without paying attention to the broader organizational context, may be in for some difficulty, Bozic suggested. He emphasized the importance of getting various internal teams talking together as a precursor to fixing the data problem.
Bozic pointed to Agile as a method for breaking down silos within an organizations and getting the database team, the infrastructure team, the security team and other groups working together to take on challenges. Breaking down cultural barriers is the first step toward understanding how to change an enterprise’s data structure and paving the way toward cloud-native development patterns, he noted.
Success is unlikely “if you don’t have the culture of becoming Agile and processes in place and cross-functional teams,” Bozic explained.
Smart regions, which seek to unite neighboring cities and towns around joint smart city projects, have begun taking shape across the U.S.
A pioneering force behind the emerging smart region trend is the Institute for Digital Progress (iDP), a nonprofit based in Phoenix. IDP developed its focus on smart region initiatives after working one on one with 22 cities and towns in the Greater Phoenix region.
“While we were doing a lot of great work, working with these cities individually, we realized that in order to be competitive, we were going to need to get the [cities and towns] to think together, act together, innovate together and procure technology together. … We realized that our competitive advantage was scale,” said Dominic Papa, executive director and co-founder of iDP.
IDP launched the Greater Phoenix Smart Region Initiative in 2018, resulting in a consortium of the 22 cities and towns iDP was working with in addition to Arizona State University.
Smart region benefits
One of the goals of iDP’s smart region initiative is to “disrupt government” and change the traditional procurement processes that often bog down government technology projects, Papa said.
The consortium, he said, has complete buy-in from state and local government stakeholders, who aim to develop cooperative purchasing contracts that will open up regional collaboration, innovation and procurement. By procuring technologies as one regional entity, consortium members can gain greater economies of scale and purchasing power than attainable otherwise.
“While procurement has been the messy part of this, we are actively working to solve it, and that is kind of the most exciting part,” he noted.
As for smart city projects, the consortium has set out to identify four to five urban challenges that it can collectively tackle. He added that part of the consortium’s efforts will include adopting interoperable technology platforms and systems. “It doesn’t make sense for the City of Phoenix to buy a transportation solution to speed up traffic through Phoenix for [traffic] to come to a dead halt in the City of Tempe,” he said.
Roles for channel partners
The Arizona smart region consortium will eventually seek partnerships with channel firms, Papa said. “We are just getting [the consortium] up and off the ground. We built the operational framework and the structure for how the consortium will run, but the channel partners are going to play a critical role in this.”
Partners, for example, can offer the technical expertise the consortium needs to deploy smart city projects. “We understand the problems … but then we need the channel partners to bring the deep understanding of the technologies and how those technologies can adapt and fit and address those challenges and issues,” he said.
He noted that North Texas, North Florida, Cincinnati, and Central Coast California all have smart region initiatives underway. He hopes to see various smart regions eventually forming a country-wide network that will share best practices and knowledge.
“The more of these smart regions that there are, the better we will be,” he said.
Microsoft has dropped planned policy changes that would have compelled Microsoft channel partners to pay for software they have been using under internal use rights licenses.
Amid considerable partner dissent, Microsoft on July 12 rolled back its plan to cut IUR licenses, a move that would have put partners on the hook for thousands of dollars in fees. The change in direction comes a few days after the new IUR policy surfaced and just before the upcoming Microsoft Inspire partner conference, which runs July 14 to 18 in Las Vegas. Microsoft revealed a number of investments in Microsoft channel partners ahead of the event.
Gavriella Schuster, corporate vice president, one commercial partner, Microsoft, discussed the company’s decision to rescind the planned policy changes in a blog post:
“Given your feedback, we have made the decision to roll back all planned changes related to internal use rights and competency timelines that were announced earlier this month. This means you will experience no material changes this coming fiscal year, and you will not be subject to reduced IUR licenses or increased costs related to those licenses next July as previously announced.”
The IUR change was to go into effect July 1, 2020.
In panning the software vendor’s proposed changes, Microsoft channel partners cited a combination of new licensing costs and a diminished ability to gain experience with products. Channel companies use vendor technology in-house to run their businesses and then deploy products for customers in a use-what-you-sell strategy.
In 1962, Ross Perot was an IBM mainframe salesman with a vision for providing technical expertise and services around the big machines.
He left IBM that year to launch Electronic Data Systems (EDS) in Dallas with an initial investment of $1,000. In the ensuing years, EDS helped pioneer what would become the IT services industry. EDS sold compute time on underutilized mainframes, built and managed computing centers for customers and created a network of data centers to run clients’ workloads remotely. Those services would eventually morph into IT outsourcing and, decades later, cloud computing.
Perot, who died July 9 at the age of 89, was ahead of his time at EDS and seemed to have a keen sense of repeating historical patterns. General Motors purchased EDS in 1984 and Perot transitioned from entrepreneur to GM board member. The shift proved unsuitable: Perot in short order became fed up with what he viewed as the car maker’s stifling bureaucracy and eventually departed GM-EDS through a $700 million buyout. In 1988, Perot was back in entrepreneurial mode, founding Perot Systems Corp. Perot’s investment company that launched the new venture was aptly called HWGA Partners, with the acronym standing for Here We Go Again.
EDS took Perot to court, claiming he had violated the terms of a non-compete agreement. Perot, however, asserted he was free to compete, provided he did so on a not-for-profit basis. The lawsuit played out in the Fairfax County (Va.) Circuit Court, where Perot amusingly told reporters no one could beat his prices, alluding to his non-profit status.
The judge sided with Perot on the question of competition, ruling that Perot was free to seek contracts provided those deals didn’t contemplate a profitable return. That restriction expired in late 1989 and Perot Systems was free to make money.
With the lawsuit out of the way, Perot Systems went on to capture projects with high-visibility clients and, in another example of circularity, advised Perot’s former employer, IBM, on the outsourcing business.
EDS and Perot Systems are gone now, with EDS finding its way via acquisition into DXC Technology and Perot Systems absorbed within NTT Data. But the industry Perot helped create continues to grow — Gartner forecasts global IT services spend of $1.065 trillion in 2020 — and evolve.
What do you get when you cross the CRISP DM model with Microsoft’s TDSP? Well, yes, another acronym. But Cognilytica, a market research firm in Washington, D.C., believes the pairing provides a methodology for AI project management.
CRISP DM stands for Cross-Industry Standard Process for Data Mining, a step-by-step approach for launching a data mining project. CRISP DM was created in the late 1990s, before the current surge in AI investment. TDSP, or Team Data Science Process, is a methodology for implementing data science initiatives. Microsoft introduced TDSP in 2016, with the aim of facilitating predictive analytics offerings and intelligent applications.
Cognilytica’s Cognitive Project Management for Artificial Intelligence (CPMAI) methodology combines CRISP DM, TDSP, Agile and its own thinking and research on best practices for AI project management. The company offers training and certification on CPMAI.
Building upon the CRISP DM model
Ron Schmelzer, managing partner and principal analyst at Cognilytica, said AI projects call for a different take on project management compared with traditional IT initiatives. Specifically, an approach for running an AI, machine learning (ML) or cognitive technology project must be “much more data centric,” he noted.
The CRISP DM gets an AI adopter partway there. The methodology starts with understanding a business’ data mining goals and works its way through phases such as data collection, preparation and modeling. CRISP DM, however, “is not AI-specific and is missing some AI details,” noted Kathleen Walch, managing partner and principal analyst at Cognilytica.
CPMAI takes the data centricity of the CRISP DM model and adds TDSP, which Microsoft describes as an “agile, iterative data science methodology.” The Team Data Science Process encompasses a data science lifecycle definition and a standardized project structure along with infrastructure, resources, tools and utilities.
Cognilytica, meanwhile, contributes components such as best practices from in-production AI implementations, ML model training approaches and ML model evaluation.
Broader AI discussion
Schmelzer and Walch discussed CPMAI during last months’ AI World Government conference in Washington, D.C. Other conference speakers outlined a three-step process for launching AI projects and discussed conversational AI’s potential in the public sector.
AI World will take a wider-angle view of the technology when the conference convenes Oct. 23-25 in Boston. The event will consider AI in manufacturing, healthcare, pharmaceuticals and financial services among other markets.