Tuesday, May 12, 2015

Are we preparing our Organisations to absorb smarter machines, as integral part of talent pool?


In the Second Machine Age, it’s not so clear whether humans will be complements or machines will largely substitute for humans; we see examples of both…… Erik Brynjolfsson, Schussel Family Professor of Management Science at the Sloan School.
Machines have always been integral instrument available to workforce to deliver work more efficiently and consistently. They performed routine and well-structured tasks, be it manual or cognitive, helping workers boost their outputs.  But with improvement in technologies such as AI, Cognitive computing, Natural Language processing, big data analytics, machines have become lot smarter.  Today some of the activities which were unquestionably human (being complex, non-routine), can also be performed with machines, not only efficiently but also with higher quality outputs. 

There are different projections about the extent to which Smarter Machines (SM) will be introduced in our work delivery model, working alongside or instead of human colleagues.  Carl Frey and Michael Osborne have conducted analysis of 702 occupations to rank their susceptibility to technological advancements linked risk.  According to their estimate, around 47% of total US employment is in high risk category.  
If this is to happen, are our organisations ready to embrace this challenge, which will manifest itself in terms of leadership expectations, performance measure, workforce strength, skill requirements or work related policies and processes?  Have we started thinking about the impact of these smarter machines on our organisation, and what should be our approach to managing it?  Is it time we start asking these questions?

New machines display significantly higher capabilities, when compared to traditional machines on three key dimensions:
SEARCH CAPABILITIES:  While traditional machines were able to search from within structured data sets in prescribed formats, SM are able to search from structured and unstructured data, coming from multiple sources, in different formats, almost in dynamic manner.

SOLVING CAPABILITIES: While traditional machines helped solve structured routine well defined and programmable set of problems, smarter machines are able to increasingly solve non-routine, context sensitive problems, with increasing accuracy with every usage.  SM can identify patterns and trends from large data sets and put forth advice with associated probabilities.      
SERVE CAPABILITIES: Machines can now serve responses in any format, and converse with humans as near humans with natural language as medium of conversation. 

Every other day, we see new examples of SM replacing human work, completely or partially, as enterprises experiment with them.  Some published examples are shared as case in point:

1.       USAA, a financial service company, uses smarter machine to handle its Armed Forces customers’ queries with regard to transitioning to civilian-life.  Machines uses data base of 2000+ questions and 3000+ military training documents as search base to solve customer questions and respond in natural language.  (IBV paper)      

2.       Associated Press is using smarter machine from company Automated Insights to autonomously create Quarterly Corporate Earnings stories, from data coming from Investment Research firm(ZackS). AI algorithm can process large number of financial announcements, press releases and other information and then offer personalised financial advice at large scale and lower cost. 

3.       Computer Assisted Translation technology is being employed to speed up the translation work, as it memorises earlier translations and use for preparing half cooked translated material for human translators to perfect.   

4.       Oncologists at Memorial Sloan-kettering Cancer Centre are using IBM Watson computer to provide chronic care and cancer treatment diagnostics using pattern recognition capabilities, with 600000 medical evidence reports and 1.5 million patient records and trials as referable data set. 

5.       Deep Knowledge Ventures (DKV), a Hong Kong venture capital fund appointed a computer algorithm named Vital  to its board of directors, claiming to be the first company of its kind to give a machine an "equal vote" when it comes to investment decisions. (BBC)

6.       Law firms rely on smart machines to scan thousands of legal briefs and precedents to assist in pre-trail research.

7.       Dr Mark Oleynik is working on creating automated kitchen housed by robotic chef that can create dishes like the professional chef it learns cooking from.  (Economist)

Yes, these examples may take time to go mainstream, but not as much time as earlier technologies have taken. These capabilities are growing exponentially and the costs of such machines are going downwards, fast, really fast!

While IT and Operational departments are busy identifying suitable business use-cases and associated technological solutions for incorporating these SM in the ways of delivering work, HR department needs to be proactively thinking of its implications for the organisation, leadership and workforce management.  Let us look at each of these areas:

Leadership: Leaders working alongside SM, that can throw unbiased data backed probabilistic recommendations, demands greater maturity and acceptance from the leaders. It can be a fight between domain expert taking out of experience and SM sharing advice that is based on pattern recognition from immense data from multiple sources.  Leaders need to be open to reconciling between gut feel and data based alternatives to problem solving. They need to be open to questioning their beliefs as domain experts and to re-examine counter intuitive suggestions coming out of machines. How do we empower leaders to have the strength of conviction to go their way, despite what SM advices? Are leaders going to be made accountable for going against SM advice, in case of things going wrong?  On the other hand, how will you keep domain experts motivated, if in most of the cases they have to go by SM advice.  Finally, leaders need to be smarter themselves in defining and asking right kind of questions to their smarter assistants (ie SM). 
Workforce Management: SM introduction will replace or redefine some jobs and consequent workforce requirements.  Those jobs that are high on creative intelligence, social intelligence or perception and manipulation seems to be least impacted, all others jobs are vulnerable. Tpo begin with, using the available thought-ware and general guidelines, we need to have view on the extent of staff that is working on vulnerable jobs.  It would have implications on the hiring vs contracting decisions, especially with medium term horizon. Also is there a way to reconfigure set of vulnerable jobs to ensure that certain category of revised jobs have greater creative intelligence, social intelligence and perception component? This readjustment effort takes time hence the need to take cognizance of the impending challenges and start thinking now.       
Skill Development:  The scope for skill development will have two clear objectives- enhancing skills among staff to work with SM and enhancing skills among employees to help tide over the transition as their existing jobs get make reconfigured/replaced and they have to look for alternate ways to stay employed and relevant. In the first category fall learnings with regard to natural language processing, database system and administration, and interface design, that would be spearheaded by technical department and supported by HR.  The second category include trainings with regard to critical thinking, evidence based decision-making, social intelligence and change management, where HR has to take lead.    

Use-case participation: SM introduction will most likely take the use-case based pilot approach to implementation.  While pilot may be focussed on validating the value generating potential of the new delivery model (using smarter machine), it is important for HR to also use the pilot to study the change impact analysis of the proposed solution.  For example, in the assisted decision making by manager, how will be the accountability of the wrong decision is tracked and established?  This may require some changes in the policies, as well.
Perception management: Also managing the expectations of the workforce on the efficacy of the new solution needs to be addressed as the SM effective benefits may take some usage to get to its full potential.  Here HR has to play its traditional role of managing naysayers that are quick to dismiss any new disruptive solutions and maintain the positive culture around the need to experiment and be open to new ways of working.    

SM idea shall be hitting our enterprise shore, in some form or other, and we need to be partnering with its carriers/sponsors to ensure organisation is ready to embrace it without much pain,

Do you agree?  Are there some other aspects of SM introduction that organisations should take note off? 

Share your views and comments, as always   

No comments:

Post a Comment

 
RKFEA34XP6KC