Tamr Achieves Premier Technology Partner Status With SnowflakeLearn More

Competing in a Post-Analytics World

The pace of enterprise automation is accelerating with robotics, machine learning and AI. The robots are coming, and not just to the assembly line or mega-warehouse floor.

To stay competitive, enterprises will need to understand how to use robots to automate more of the strategic business-decision-making roles in enterprises: in sales, marketing, HR, procurement, finance, product management and so on. Instead of feeding analytics into decades-old, people-bound processes, they’ll increasingly be feeding the results to robots that augment strategic, decision-driven brain processes for executives and knowledge workers.

Think about it. AI has already completely automated stock trading, guiding traffic, and setting hotel prices. In the enterprise, it can certainly automate the purchasing of office equipment or parts–and eventually other, easy-to-explain jobs. But it can–and will–do so much more. It’s inevitable.

All of this isn’t news. What is news is that it’s happening faster than we ever anticipated.

So much so that Tom Davenport, author of The AI Advantage, recently issued a call to action for enterprises: “It’s time for companies to start stepping up and developing a better combination of humans and machines. My argument has always been (if) you can handle innovation better, you can avoid a race to the bottom that automation sometimes leads to if you think creatively about humans and machines working as colleagues.”

Toward the Autonomous Enterprise

A big reason for the urgency is that a major barrier–clean, curated, classified and computable data at scale–is today being solved by human/machine collaboration. Essentially using AI to enable more AI.

Writing in the Harvard Business Review, Megan Beck, Tom Davenport and Barry Libert observed: “Our experience working with boards and leaders is that creating a solid AI product that provides either customer, employee, operational or investor value is about 40% problem and product definition, 40% data sourcing, cleaning, filling and merging and only 20% algorithm development.” (That boldface is mine.)

When you can accelerate the continuous availability, trustworthiness and reliability of the data that drives AI, you really open up the road for the autonomous enterprise.

Well-fed AI, ML and robots will enable the enterprise to continually operate more autonomously. It achieves this by automating not just mechanical processes but also the thought processes of knowledge workers and strategic decision-makers.

I see the evolution of the autonomous enterprise as very analogous to what’s happening with self-driving cars.

Today, AI-cleaned and -integrated data enables industrial-strength predictive analytics. Think of this as basically “collision avoidance” for enterprises. Predictive analytics is fast becoming a standard feature in enterprises, just as collision avoidance is a basic feature in self-driving cars. As AI learns and gets smarter, the enterprise can steer and maneuver better on its own (think lane tracking), enabling prescriptive analytics. Next up is the automation of parallel parking and other more sophisticated, cognitively intense actions. While full self-driving (in both cars and enterprises) is a ways off, the intermediate steps are happening and much quicker than we thought.

Forward-looking CEOs will get this. They’ll start to set the rules by which all enterprises will operate in a post-analytics world.

The customer experience for virtually every enterprise is already being transformed with robotic automation, in industries from financial services and retail to manufacturing to healthcare (thanks to competitive pressures from Internet-native companies).

The operations experience needs to be similarly transformed. The days of executives, senior managers or knowledge-worker professionals having to be involved hands-on (brain-on?) in every strategic decision are over. Data can be collected, parsed, analyzed and (increasingly) evaluated and acted on independently by AI, much faster than humans can.  This is a done deal.

Data-driven to the Max

If data is fuel for the enterprise today, it will need to be rocket fuel for the enterprise of tomorrow: ever-more-precise and capable at delivering the needed thrust to get (and stay) off the ground.

Once decision-makers (CROs, CMOs, CFOs, CPOs, etc.) start making more data-driven decisions, an obvious question will be: Why can’t we create neural networks (robots) that are trained by these data/decision outcomes and largely replace the humans in the loop? Imagine Chief Revenue Robots (CRRs), Chief Marketing Robots (CRRs), Chief Financial Robots (CFRs) and so on. While this may sound like the script for a bad science fiction movie (you heard it here first), conceptually that’s where things could and should head.

As for the political hot potato of jobs lost to automation,Senen Barro and Tom Davenport summed it up nicely: “The greatest impact of intelligent technologies won’t be from eliminating jobs but from changing what people do and driving innovation deeper into the business.”

CEOs should start to plan for this kind of future today when building out their data and analytics strategies. Clean data is the obvious and proven place to start. Empowered by continuously clean, updated and classified data (itself managed by advanced models and AI), enterprises like GE and GSK are ready to make the next AI move.

Here are the top things you can do to get your organization on the road to the autonomous  enterprise:

  • Deploy DataOps/modern data engineering, including embracing the emerging open DataOps ecosystem. The former will help you build the data refinery and pipelines that feed robots, AI and ML with continuously updated, curated and computable data. The latter will help you move faster and more efficiently by using best-of-breed, interoperable tools (many of them FOSS), releasing the handbrake imposed by relying on single-vendor solutions (e.g., ERP systems).
  • Embrace your data variety and silos. Both are facts of life and will be for decades to come, particularly because of enterprises’ reliance on structured data. Human-guided machine learning tools for data integration, cleaning, curation and classification can feed your AI-destined pipelines, while reducing the time and cost of doing so. One example is Master Data Management at scale, an AI-driven twist on a previously human- and technology-bound tool. MDM at Scale lets enterprises efficiently create dynamic, AI-consumable “master” data packages of the entities that propel the business: customers, employees, products, parts, and so on. (No more relying on inflexible, bespoke rules.)
  • Learn from early adopters. Watch and learn how other enterprise companies have asked and answered the question “How can we best use AI and other cognitive technologies to become more agile and competitive?” by leading with a data-first approach. Yeah, it’s still early but there’s some inspired work going on. The Mark Ramsey-led initiative at GSK is one of my favorite examples. There are others. For example, there’s some use cases and advice in Tom’s AI Advantage book, including a more fully articulated version of the GSK story.
  • Raise your AI game, by putting in place some key AI roles that bridge the gap between business and technical management of AI.
  • Get started today – or you will never catch up. I saved the best advice for last (again from Tom Davenport).

Everything in our world is becoming more autonomous these days, from our (non-self-driving) cars and home appliances to our financial, retail, education and healthcare experiences. Look around you. It’s time for enterprises to catch up, starting from the top down (executive suite).

What do you think?  What’s your business doing to become more autonomous?


#AutonomousEnterprise #NoMoreBadBigData #Post-AnalyticsWorld #EntityThis!