Can artificial intelligence truly be used in a positive manner to enhance the efficiency and effectiveness of our business processes, thus getting us closer to the Grail first considered 100 years ago? Can making our business tools synthetically smarter, help us make informed decisions, make our businesses better and augment the achievements of our business objectives?

Artificial Intelligence: Back to the basics

Believe it or not, the concept of automating business processes to increase efficiency and effectiveness has been around for centuries. Author Gil Press for Forbes traces the origins back to 1308! The concepts and coined phrases of “machine learning” and “artificial intelligence” trace back to the 1950’s. But it has only been in the last decade or so when the generation, access, storage and analysis of enormous amounts of data have made the rational application of automation to business process feasible. Even more recently, the ability to hash through this data creating information and then applying this information in an almost real-time fashion, and then, further modifying this information as it becomes actionable, provides a simulacrum of intelligence.

So much has been said about the rules-based computing that forms the basis of all AI, it is important to recognize that the output is ultimately a decision which is then used to either affect another action or process or to inform another decision (or decision-making process). It is equally important to recognize the distinctions between data, information and knowledge which are as fundamental to decisions and AI as protons, neutrons and electrons are to atoms.

Understanding Data, Information and Knowledge 

Data is comprised of facts, figures and numbers without context, such as can be found in the Internet of Things output, revenue, headcount, names of locations, clients, business partners and anonymized clinical data. Data is not necessarily correct or accurate, but there is almost always a lot of it.

Information is the result of applying context and relevance to data. Examples include clinical patient profiles, revenue by geographic region, sales per business developer, grade progression per student. The take-away here is something has been done to the raw data which has enhanced its value because it now has meaning and context.

Still, information is two-dimensional in a sense. Yes, analytics has made it more intelligible and meaningful by “endowing with context, relevance and purpose,” as pronounced by the brilliant Peter Drucker, but it remains static. That is, until it is made actionable. When something is done with that information (via insight), when the information yields a decision or is used cyclically to refine (modification through fine-tuning), the raw data that was the original basis of the decision, knowledge, is brought forth. Some would contend there is a further dimension euphemistically referred to as “wisdom”. Just as “knowledge” is spawned from “actionable information,” wisdom comes from “experienced-informed decision-making”.



At which point in this hierarchy does so-called “Artificial Intelligence” loom?

I do not contend we get too bogged-down in semantics, but it is important to recognize the underlying “Currency” of knowledge and hence, AI. How and where does the data come from? Can we trust it? It takes (a lot of) time and (a lot of) money to analyze and mold this data into something comprehensive enough to make a decision…so how do we decide which data to use? What are the algorithms dictating the decision process itself for the above?

In recent years, the ability to “crunch” these vast amounts of information to better enable better business decisions has grown vastly. This gives us the cabability to finally utilize automation as not only the originally envisioned time-saver, but also as a source of informed recommendations. Structuring additional rules atop these recommendations allows for the automation of just about any business process. Furthermore, automated feedback can be generated as part of these business processes which can refine and enhance the origin data making for “smarter” automation of said processes (perhaps fashioning “artificial knowledge”?).

Applying Artificial Intelligence to Healthcare 

In the universe of bio-pharmaceuticals, medical devices and healthcare in general technology and software advances, have brought us to the point where not only is all this feasible, but it is commercially viable and readily available.

In early 2017, Harpreet Singh Buttar from Frost & Sullivan, predicted that artificial intelligence in healthcare will see a “dramatic market expansion” in the next couple of years, with the potential to reduce the cost of medical treatments by half across the board. “By 2025, AI systems could be involved in everything from population, health management, to digital avatars capable of answering specific patient queries.”

Currently, a vast number of life sciences firms are applying machine learning algorithms and predictive analytics to reduce drug discovery times, provide virtual assistance to patients, and diagnose ailments by processing medical images, among many other things.

Companies that provide AI-based tools to life sciences and healthcare firms appear to be coalescing in specific areas:

  • Imaging and diagnostics (which some investment firms believe is getting increasingly “crowded”)
  • Remote patient monitoring
  • Core AI platforms
  • Drug discovery
  • AI in oncology therapeutics

This is by no means exhaustive. In fact, the companies providing “Core AI Platforms” are finding their use in life sciences firms in areas as diverse as research, to regulatory monitoring and compliance, to manufacturing to sales and marketing. An interesting study entitled “AI in Pharma and Biomedicine – Analysis of the Top 5 Global Drug Companies” surveyed how the top pharma firms were currently employing AI solutions.  What was found was a trend for:

  • Mobile coaching solutionsprovide advice to patients, product/medicine users, using real-time data collection. This can enhance patient compliance, knowledge dissemination and outcomes (applications noted in many firms including AstraZeneca, Johnson & Johnson, Pfizer and others);
  • Personalized medicine– The ability to analyze large amounts of patient and clinical data to identify treatment options. Typically utilizes a secure cloud-based system and base software that can process natural language (such as IBM Watson and Infosys’ NIA)
  • Acquisition frenzy– The global talent pool for AI solutions is limited though certainly not limited only to the large life sciences firms!  Just as we have seen in the drug development arena in the last 10 – 15 years, a sudden increase in acquisitions and exclusive partnerships, we are now seeing a similar burst of activity occurring with smaller companies that can fill a niche in the life science firms’ AI needs.
  • Drug discovery – As I alluded to in a May 2017 white paper Automating Healthcare: Balancing Efficiency and EthicsAI has found a home augmenting the analytical and decision-making capabilities of pharma research scientists. It’s well known and fully documented that the process of drug discovery is hugely expensive, lengthy and risky. The ability for AI to rapidly analyze large volumes of data based on rules, requirements and algorithms specific to the research is essential in this sector and well supported by these AI tools.

In early 2017, Forrester published an interesting “TechRadar” report entitled Artificial Intelligence Technologies, Q1 ’17While the report discusses 13 technologies that “enterprises should consider adopting to support human decision-making”, what I found most interesting was the analysis of obstacles to the implementation of AI (as expressed by “companies with no plans to [invest] in AI”). I was amazed that 42% of these firms said “there was no defined business case.” In fact, 39% noted that they were “not clear what AI can be used for.”

What does the Future hold?

The benefits of adopting AI in life sciences are becoming clearer and clearer. And these benefits can clearly be tied to enterprise-level business objectives. So, those 42% of the companies without a “defined business case” should consider taking the time and doing the analysis, developing a business case and deploying a strategy. These strategies can be at a departmental or functional level or at a macro enterprise level. 

Regardless, it is a current reality that the torrent of data and access to this data is ever increasing. We have reached a point where we are encountering “data walls.” The influx of enormous amounts of data have obfuscated the answers we seek. And in a quantum physics manner, the mere act of looking for specific bits of information, affects and adds to the “data wall” moving us further away from our target. AI can help. AI analytics can chip away at the proverbial wall and help us develop the insights needed to achieve knowledge and ultimately, wisdom.

Michael Breggar

Michael Breggar

Michael Breggar is a managing partner at Infosys Consulting, and runs the firm’s life sciences practice in North America. He has 25+ years of industry experience, having run advisory work for some of the biggest firms in the world in this space. Mike can be contacted via LinkedIn or at

Pin It on Pinterest

Share This