Companies are rapidly deploying emergent technologies (RPA, AI, ML) to create quick wins and augment their digital transformation efforts. Along the way they will need assistance to architect a strong foundation to ensure trust, foster employee engagement, and support continuous learning as this is a journey.

Initial Robotic Process Automation or RPA implementations create a lot of excitement (like Service Oriented Architecture or SOA did decades ago) and then things get complicated once they reach critical mass adoption and companies scramble or lack the ability to successfully “operationalize” RPA and permeate the impact across their business processes. The challenge companies are facing, is to ensure their data and information are of high quality (running a machine learning algorithm against data with quality concerns is going to generate spurious results, which can propagate across any/all analytics that rely upon the outputs.  Embarking on RPA without well documented processes can create unexpected anomalies as agile teams attempt to push out automated interactions with customers (bots).

Artificial Intelligence or AI is being introduced in a number of places, but success requires a very carefully defined set of boundaries. AI sounds good in theory, but scaling it is very hard and algorithms tend to generate spurious results if features are not carefully controlled. The AI spectrum is focused on two core areas: First, the engineering to improve speed and accuracy around tasks that are repetitive and where discovery of patterns requires enormous resources and time due to the sheer volume (healthcare, credit risk, behavioral patterns discerned from observation and/or social media come to mind). Second, is the “holy grail” of learning where machines can teach themselves (unsupervised). This is where Deep Learning and Neural Nets algorithms are used to discover and find patterns from vast amounts of data on their own. These tools require specialized skills and are still in their early stages of development.

Machine Learning or ML has many flavors, most are just trying to learn how to train the tools via supervised learning (humans augmenting the training for the algorithms). A few organizations are dabbling with supervised learning, but results are hard to achieve outside of a few narrow areas. Unsupervised learning is typically used to perform cluster analysis (for exploring hidden patterns). Once hidden patterns emerge, supervised (or partially supervised) learning can be used to focus on more conventional opportunities (demand forecasting, supply chain optimization, customer 360, social media analysis, diagnostics, risk, and preventative maintenance).

The Ascend has charted this course with other technologies over the last 20 years. We help clients establish a center of excellence (COE) around process, digital enablement and transformation, data management, and analytics to support scale and ensure clients realize the multiplier of productivity.  As clients step into these new technologies, they quickly realize the need to shore up older (less agile) infrastructure and systems. A large portion of or projects involve developing roadmaps and guiding solution design to help clients prepare their systems, technologies, data, analytics, and processes so they can properly leverage these technologies. We also guide leadership to build buy-in with their employees and work closely with IT teams since the digital workforce is an inclusive model that (when done properly) helps cross barriers, foster engagement, and drives significant productivity improvements.

Ascend is advancing RPA/AI pilots with our clients in enabling an end-to-end business architecture approach to process automation, introducing automated bots where necessary, inserting advanced intelligence (AI/ML) engines in the automated processes, and predictive analytical models to detect, infer, and make smart recommendations. We have employed such concepts in aerospace, logistics and supply chain, consumer B2B/B2C, financial and media industries’ use cases.

Ascend also is adept in advancing such innovation ideas to the C-Suite, building the relevant business case and enabling provable pilots that demonstrates success with high value, balancing costs and risks. As part of the Accelo MethodologyTM toolkit, we often perform a Total Economic Impact (TEI) forecast analysis up front so your innovation trials can be purpose-built, designed to fail-fast (if that happens), course-correct in an agile iterative fashion. Learn more at Digital Transformation Solutions (DTS) practice

Categories: Blog