Genpact

Amaresh Tripathy

There’s a lot of hype around artificial intelligence, but we are light years away from businesses that are outright replacing humans with machines. Human-level artificial intelligence may never be a reality, and at best, these technologies can only complement the qualitative aspects of human intelligence to make decision making faster and more real-time.

While it’s a highly anticipated digital technology, a better way to think of artificial intelligence is to consider how it can be leveraged to amplify human intelligence. Augmented Intelligence is more than just unaided AI – it’s a confluence of data, digital technologies, and human intelligence: where data processing, data science algorithms, and digital technologies extend, but do not replace, human ability to make better and faster business decisions. Augmented intelligence enables organizations to build trust between humans and AI technology while addressing regulatory concerns as well.

In the Pharmaceutical and Life Sciences Industry, data is fast becoming the most valuable resource and mobilizing this data effectively is the key to surviving in a dynamic business: An increasingly complex stakeholder environment is creating a highly complex selling environment; specialized therapies are demanding more personalized sales relationships; and demands to demonstrate value require nuanced, substantiated messaging.

Availability of data from a wide array of sources—including internal, third-party syndicated, and social media data— makes it possible to harness targeted insights and drive tailored conversations with healthcare providers, payer networks and patients. However, making these conversations productive requires solving for relevance and “data overload.” To address this challenge, insights must be filtered and contextualized with the personalized information so that sales and marketing professionals receive timely insights that are germane to their jobs— the right insights at the right time in the right way.

Augmented intelligence can greatly enhance the quality of the customer experience and improve top-line sales when applied to field execution—pharma’s dominant sales channel. Sophisticated machine learning algorithms that take into account individual prescriber affinities, recent prescribing behavior, affiliation changes, and digital footprints present the sales force with a deterministic output for the “next best action” for individual prescribers. When integrated into existing sales management systems, these algorithms can help the sales rep by tactical guidance in predicting the best message, timing and content for each provider through real-time alerts and notifications upon which the sales rep can then take action while overlaying their own judgement of individual prescribers. Each action taken is fed back into the system as an optional result, and the reinforcement learning algorithms can then learn from the feedback and optimize the next set of actions.

Another powerful example of augmented intelligence in Pharma is the use of natural-language processing focused on detecting language nuances from unstructured data. Almost all pharmaceutical companies receive customer feedback, complaints & issues from multiple sources including Phone, Emails, Web Chats, Veeva, social media, and so on with a large portion of this data being verbatim text. Currently, existing data and analytics teams spend a significant amount of time manually collating, tagging, and classifying some of these customer feedback. Medical and therapy area experts can now leverage their deep domain knowledge to augment contextual linguistics algorithms to establish the right contextual dictionary – a detailed library of keywords, language nuances, and drug behavior taxonomy broken by key stakeholders (HCP, Patients, Caregivers) that overlays any stock NLP engine. 

This can accurately discover insights from current conversation themes, topics, and sentiment, and present data and analytics teams the ML muscle to add domain-specific knowledge and tease out key insights. This bilingual talent (analytics and domain) thus significantly reduces lead times in addressing customer concerns while optimizing communication decisions.

Transactional and problem-solving activities are shifting from humans to machines, allowing humans more time to think, build relationships and innovate, in addition to overseeing the machines. As machine intelligence matures and gains adoption, enterprises need an augmented intelligence framework to bring together machine predictions & human judgment. Getting to this stage needs three pillars to be established:

• Domain-focused: Business operators know the inner workings of businesses better than anyone else. With bilingual (analytics + domain) talent, they find the best transformation opportunities.

• Data foundation: Machine intelligence finds, sorts, and learns from data that was previously inaccessible.

• Intelligence: Human intuition, backed by technology and business know-how, unlocks predictive and prescriptive insights and ultimately leads to smarter processes.

Although there are varying degrees of adoption, most Pharma companies have embarked on the journey of embedding data science and machine-based learning into their commercial operations. However, the journey to augmented intelligence needs a few other structural components:

1. The democratization of data: via a next-generation data ecosystem that is the single source of truth for all structured, unstructured, internal, or syndicated data available. This ecosystem delivers strong data integration and analytics capabilities with clear ownership and accountability for data procurement, usage, technology licensing and consumption

2. An agile operating model, with executive sponsorship that breaks functional silos and focuses on consistent communication and collaboration around objectives, timelines, and outcomes. Executive sponsorship is pivotal in ensuring adoption and in limiting fears around technology-led job losses.

3. Global scalability: As the next-generation ecosystem matures, there will need to transition from a distributed country- or region-specific model to a federated global network model driven by business collaboration and data consumerization.

4. Culture and people: When it comes to data science, skill shortages are and will continue to be a defining factor of enterprise competitiveness. Organizations will have to pursue a multi-pronged strategy encompassing internal programs and training charters, academic partnerships, and start-up ecosystems.

One of the biggest hurdles in the pervasive adoption of augmented intelligence will be in demonstrating viable results: its ability to deliver outcomes that go beyond human judgment consistently. Organizations that succeed are the ones that put into place an agile operational model that can scan processes to identify the right prediction problems, identify and nurture the right bilingual talent while maintaining a laser focus on continuous improvement and re-skilling.