The holy grail of digital transformation is the seemingly conflicting goals of high levels of customer service and pressure to reduce costs. “Digital Transformation” has become an all-encompassing term – in a piece in this column about customer data platforms, I asked whether the term has lost its meaning:
The phrase “digital transformation” can mean anything and everything — tools, technology, business processes, customer experience, or artificial intelligence, and every buzzword that marketers can come up with. Definitions from analysts and vendors include IT modernization and putting services online; developing new business models; taking a “digital first” approach; and creating new business processes, and customer experiences.
The overarching objective of a digital transformation program is to improve end-to-end efficiencies, remove friction from information flows, and create new value streams that differentiate a company’s offerings and strengthen the customer relationship. Having assisted large global enterprises with building the data architecture, supporting processes, and governance for multiple digital transformations, in my experience, there are two broad classes of initiatives that seem to get funding and others that miss the boat in terms of time, attention, and resources.
E-Commerce – the “Easy” Win
Improving e-commerce and streamlining the customer journey are high-profile, well-funded programs that on the surface are meant to improve revenue, improve satisfaction, and drive out costs. Tens or hundreds of millions of dollars are invested in new generations of AI-powered technologies, digital marketing is revamped, customer service operations refined and yet still, many of these programs fail to live up to grand expectations. While they move the needle, something is still missing from the recipe that prevents the benefits from being fully realized. The problem is that an organization is complex – streamlining processes and information flows requires participation from multiple departments that typically work in data and process siloes while the customer journey traverses those siloes. People get their work done through collaboration both within and across siloes and need easy access to knowledge and information.
But different parts of the organization frequently use different tools and applications which adds friction and slows down decision making and problem-solving. At a large manufacturing equipment provider, field service reps needed to look in over a dozen systems to repair and maintain custom-engineered installations. This meant searching for information in each silo and resulted in longer service calls, greater downtime, and higher costs. Multiple departments were involved in creating service and support content and did not use the same naming conventions so accessing information meant understanding the nuances of these various systems.
The problem was solved by creating an information structure that described equipment, components, configurations, problems, error codes, troubleshooting procedures, and other aspects of the installation including as-built details from an ERP system.
Information was broken into chunks so a tech could get an answer to a problem rather than have to search through large technical documents. This same approach is also being used to power cognitive AI – the chatbots and virtual assistants that surface information through conversations. While this work was done to improve conventional search, the approach paved the way for more advanced AI tools.
The objective of the approach was to streamline knowledge flows across departments and ultimately to solve customer problems faster and at a lower cost.
Capturing, Retaining and Applying Institutional Knowledge
How do knowledge flows connect to digital transformations? A group of people working in a team is part of a knowledge community. This is where experts come together to solve complex problems. But if the team is transient and disperses after producing a solution, that knowledge needs to be captured as enduring institutional knowledge. Expertise leaves the organization through attrition, downsizing, reorganizations, and retirement. For the enterprise to run, institutional knowledge is embedded in processes, systems, tools, and documentation as well as the accumulated experience of employees working in various departments.
The challenge is that most of the time, these groups are left to their own devices. Collaboration and problem-solving are chaotic and too much structure can slow things down. But without some standards and structure, knowledge debt begins to accumulate. This is called technical debt in IT projects. Knowledge debt happens when things are not well documented or organized for repeatable, intuitive access. Enterprise search is a notoriously difficult problem to solve, repositories of knowledge get cluttered with outdated information, and inconsistent tagging of information makes retrieving the best solutions, answers, designs, deliverables, plans, and product documentation haphazard and difficult. This leads to friction which slows down the digital machinery of the enterprise.
This results in higher support costs, dissatisfied customers, compliance violations, manufacturing errors, and overall inefficiencies from the acts of heroics that people need to go through to get their jobs done.
Artificial Intelligence to the Rescue (really?)
Many organizations are pinning their hopes on AI solving these intractable, evergreen problems that seem to defy sustainable cost-effective solutions. Some organizations spend millions of dollars every few years cleaning things up and focusing on a specific part of the process using technology. Everything looks good for a while (because with the installation comes clean up – or at least a fresh start) but things soon go south and the problem arises again.
AI tools can help under certain scenarios. In the early days of AI vendor confusion and unrealistic promises, many claimed that all you needed to do was “point the AI to all of the data” and it will work its magic. Customers quickly learned that AI technology has to be “trained” and specific information sources and frequently with a foundational structure or data architecture. (Some may argue that machine learning can figure out all of your product names and attributes, but I have yet to find a customer who has experienced that level of hands-off functionality). Certain algorithms can make sense of messy data, but cognitive assistants are in fact trained on high-quality, curated, tagged and structured data, content, and knowledge assets. That is a knowledge and content management problem. Where does that high-value knowledge and expertise come from? Knowledge communities. The engineers that design and build solutions. The service techs who come across challenging field conditions and anomalies.
So-called “cognitive” artificial intelligence applications – the Intelligent virtual assistants and knowledge retrieval bots that many organizations are experimenting with – get their abilities not from magic AI pixie dust but rather from knowledge-engineering approaches to information management. These approaches can solve real problems today while preparing for a future of high-performing virtual assistants.
Digital Transformation Program Scope – Knowledge is the First to Go
The unfortunate fact is that organizations are prioritizing user experience “ and “usability” over knowledge processes when undergoing transformations. These decisions are made when budgets run hot, timelines slip, and unexpected issues come up (seasoned program managers know that they have to expect and budget for surprises). But cutting knowledge from the scope will leave organizations flat-footed. This is a serious error that will lead to lost market share and higher costs as they scramble to catch up. In some cases, the organizations will not be able to catch up and will go the way of a Blockbuster or Kodak. Why? Because eventually almost all interactions will at least partially be enabled by virtual assistants and bots. In some situations, those tools will be the primary way that the enterprise interacts with customers. That is the inevitable course we are on. In five years, those organizations that are not making the investments in maturing their knowledge initiatives will wake up to competitors who have been developing capabilities over the past decade and find themselves woefully and in some cases – unrecoverably – behind.
Many organizations consider content strategy and SEO to be a primary area of effort during their digital transformations or e-commerce redeployments. That is fine for attracting customers to your offerings but is short-sighted and misses the boat around larger knowledge issues. This cannot be compartmentalized into “we’re just doing enough for SEO and will come back to this later.“ That approach will not work. Knowledge and content must be aligned with customer journeys using high fidelity journey maps that can interpret the digital body language of the customer and respond to those signals through the digital machinery designed by – perhaps you have guessed – knowledge communities.
In one organization undergoing a digital transformation, the marketing team owned content for the e-commerce site however customer support content and knowledge were not part of the plan. That team only wanted to focus on SEO. Two years into the project, people started asking about their knowledge strategy for the detailed engineering information that customers relied upon. The project was already behind schedule and over budget and design decisions made earlier in the program limited options. The result was a lower customer satisfaction rating and higher volumes of calls to the call center. The project was supposed to reduce those calls by improving the user experience. For this organization, part of the user experience was access to knowledge and expertise which became harder to locate on the new site.
AI and Knowledge Flow
AI and machine learning technologies can support knowledge flow in several ways.
- Organizational network analysis – ONA identifies connectors, informal networks, influencers, and hidden structures that are critical to understanding knowledge communities and networks. Machine learning can process many data sources and prescribe actions to help improve collaboration, reach, and effectiveness of communities.
- Sentiment analysis – identifies the tone of communications among and between individuals and communities. Is there healthy task-focused debate? Or have politics and personal gripes disrupted flow and effectiveness?
- Expertise identification – self declared expertise is notoriously inaccurate. More accurate expert profiles can be derived by processing multiple sources of content such as written project summaries, discussion posts, and contribution to intellectual property (white papers, methodologies, analysis, and other documents).
- Improved search and retrieval – Semantic search allows for language and phrase variation (returning “proposals” using the term “SOW” in a search when no document contains that term) as well as results that are personalized based on role, preferences or other signals. “Helper bots” use advanced and/or federated search under the covers – the bot knows where to look for specific information reducing the friction/overhead in collating information from multiple sources.
- Knowledge recommendation – recommendation engines can surface high-value knowledge based on team goals and project profiles. For example, providing documents used in similar projects or containing part of the solution to a problem the team is trying to address.
Organizations Compete on Knowledge
Every differentiator comes down to knowledge – the institutional knowledge of how the business operates at every level, technical knowledge, and IP, knowledge embedded in software and designs, knowledge of customer needs, of routes to market, of channel partners, knowledge of how messaging will break through in a crowded market. Of better ways to design the customer experience or product features. As technology continues to speed up all processes, the knowledge lifecycle will become the critical differentiator because the race is on to build cognitive assistants that will speed internal processes and support the customer at lower costs.
The future is one of supersmart devices and distributed intelligence in everyday technologies. Not just the microchips that optimize the device, but in the human-computer interface that will all but eliminate the need to call a support rep. The support rep will be built into the device. They will diagnose themselves, open a support ticket and order the parts they need. When the human maintenance tech arrives, the devices will tell them how to perform the repair. The organizations that get there first are the ones that are making the investments in knowledge today. Many executives have been burned by knowledge programs and are therefore gun-shy. They need to understand that this will not be a nice-to-have. It will be critical to survival in an AI-powered cognitive future.