It is difficult to come across an advertisement that does not include a reference to Artificial Intelligence, or “AI.”  It is top-of-mind in every conversation that touches on technology. Does an AI mention represent a true technological breakthrough or is it a gleam in the copywriter’s eye? Like the words “vegan” and “organic” in the corner grocery store, AI is often on the packaging – even if the label doesn’t apply.

To be sure, everything is evolving, and we are trapped between what is real and what is possible. Many business executives see technology-driven upheaval around them, so they live in constant anxiety of disruption and extinction caused by new and unfamiliar technology.

So, how do you survive and prosper in this era of rapid technological change and conversational sleight of hand? You begin by adopting technology that makes the most sense for your customer and your business model. And you test your own survivability by trying to disrupt yourself before a new player establishes a foothold in your segment.

Where do you start? I propose you approach the journey using my technology pyramid, applying this stepwise approach to your operation.

The Technology Pyramid: A Bottom’s Up Approach

Think of technology strategy as a hierarchy, best represented as a pyramid.

Level one – Processing Technology

Processing technology can speed orders and requests through order entry, production and fulfillment using automated routing and tracking. Rather than idling in email inboxes and voice mail systems, orders can be steered to available resources for screening and processing. This has enormous potential as it can maximize worker efficiency and shorten turn-times. Today we refer to this as automated order management, and technology abounds to address this opportunity.

Admittedly, this is not new… automated order management is the norm in loan origination, appraisal, title insurance and loan closing. Secondary loan operations are a bit late to the game, so automated order management has room to grow.

Level two – Information Technology

When a process is automated it typically spawns a stream of transaction data, which can be aggregated and summarized (metadata) and used to create production metrics and standards, such as average turn-times, quality control measures and individual performance statistics. In more advanced environments, transaction data is used to train AI models and create rules-based algorithms.

Historically, this data fed dashboards read by humans who would intervene in production processes to re-route orders, make changes to processes, or address performance deficiencies. Today, this data feeds AI models that automatically make these tweaks.

If you are not mining, analyzing, and leveraging your production data, your data assets are under-performing. Properly managed, production data represents an enormous enterprise asset capable of advancing the business along its evolutionary path.

In my experience, the single, greatest obstacle to mining, analyzing, and leveraging production data is that it is often housed in disparate, disconnected platforms. In this event you may have to create a data lake rather than dynamically linking platforms to rapidly advance your technology strategy.

Companies unprepared for “big data” management often lack a dedicated analytics staff with the knowledge and resources to mine it. Creating the team of course requires resources, but a well-organized function will quickly return its cost of capital.

 Diligently mined, production data can be valuable for identifying opportunities to take impediments out of order management, producing turn-time benefits.

Level three – Exception Processing Technology

Exception processing technology is excellent for identifying exceptions and outliers in the production or order fulfillment process. These cases can be quickly elevated to humans for research and remediation. This technology is usually rules-based, which means it codifies and applies the same business rules employed by humans in manual processing. It can replace statistical quality control processes that only sample units. It can also trap errors in production before they reach the customer, reducing the workload of customer service organizations and elevating customer satisfaction.

Exception processing technology can screen large numbers of cases in a fraction of the time achieved by a human-driven process (i.e., a census rather than a sample), which means it can be frequently applied. It can identify errors in consistency and condition, compliance issues, and milestone or event lapses.

Level four – Decisioning Technology

This is the stage where the pucker factor kicks in. Technology, usually AI-driven, makes decisions typically reserved for humans. Consumers are leery of it. Workers disdain it and resent the technology encroachment. Business executives are mystified by it. As Arthur C. Clarke once said, “Any sufficiently advanced technology is indistinguishable from magic.”

Training AI models is data intensive, so it requires a robust, transaction-level database. There are several AI model types, so choosing the right model depends on how you plan to deploy it in your business. As you may imagine, training is not a one-time event. Models require ongoing testing to ensure they deliver acceptable results. Evolving AI technology will certainly yield greater performance.

In my experience. Domain knowledge of the business category is essential to achieving success with AI. A purely scientific approach, unencumbered by practical business knowledge, can create a longer and more painful learning curve. If you seek outside counsel in AI technology, align with a specialist with experience in your business segment.

This level is the cutting edge of technology today for most industries, particularly financial services companies.

Level five – Optimizing Technology

In this advanced stage it is all about refining your technological ecosystem. You strive for ever greater speed and efficiency, improvements in data integrity and better decisions. Ideally, your clients should be enjoying a better experience. You should have sufficient data so that AI can radically reinvent your business because, in theory, AI can fabricate scenarios humans cannot.

Today, AI is used to re-align supply chains, diagnose medical conditions, even manage how other machines “learn” tasks. What AI may do tomorrow is beyond speculation. Who knows—it may even create its own future.

Modernization Best Practices

Approaching modernization systematically and progressively, following an approach like my technology pyramid, has several benefits:

First, you are unlikely to overspend if you seek incremental breakthroughs rather than radical reinvention. This poses less financial risk while building enterprise-level comfort with new technology. Additionally, the return on capital will likely be more immediate.

Second, you will reduce anxiety among the ranks of employees if they are asked to embrace and adopt technology that empowers rather than replaces them. Attitudes towards AI are fraught with misunderstanding, trust and even fear. A deliberative approach that builds on small successes will likely engender comfort and confidence.

Third, your company’s collective understanding of technology as an enabler will grow and technological competence will become an enterprise asset. The more employees gain hands-on knowledge of your automation strategy, the more energy you can tap for perspective on how best to leverage it.

Keep Your Customer in Mind

When you talk about modernizing your enterprise using technology, your customers will inevitably wonder, “what does this mean for me?” Consequently, you must keep your customer in mind as you advance modernization. What benefits can you create or improve through modernization? Lower cost? Higher quality? Faster turn-times? More convenience?

A good modernization strategy does more than inoculate you from disruption: it strengthens your bond with customers, particularly if you bring them into the dialog to explain what you are doing and why you are doing it.

The modernization journey is never ending. There will be successes and failures. You must remain steadfast in your commitment and forward-looking in your approach.