AI World : What it is in it for IT services Industry?

We @Aithoughts.Org have launched a series of Thought leadership talks once in a fortnight.  I was the first speaker and the video recording is available on our website for you to listen at your leisure. While we discussed the history of AI technologies including AI Winters etc., the main question in the audience’s mind was “What is in it for the IT services Industry?”.   I will summarize my answers to this question.

Let us look the top 4 business lines in any IT industry and see how they need to adapt, reskill and deliver services in the AI world.

1.       Infrastructure Management Services  :   The industry has come a long way in supporting large data centers and adapted to Private, public cloud, hybrid cloud and supporting global enterprises from remote Capability centers.  In the new AI world, lot of more new challenges to be addressed.  The so called “Enterprise Data” now will be extended to Audio, NLP documents, Images, Video etc.  These data will be used for AI Model Training and hence to be given the status of “Enterprise Data”.  For e.g. Current call recording data in call centers are archived for compliance and given a sort of “B” status and with AI Audio models training using these call recording data needs status of “A” enterprise grade.  The sheer size of data to be managed at Enterprise grade level will also add additional complexities.  Complexities of enterprise network and firewalls need a complete relook.  Overall, IMS sales, pre-sales and delivery teams need to reskill, rebuild processes, rebuild management tools and adapt to this new AI Infrastructure. 

2.       Business Process Services or BPO : in the new AI word, almost all  Business process agents need to work along with AI agents and need to reskill.  The Business process agent desktop will become very sophisticated with high definition cameras, high quality audio devices and also high powered systems with edge processing powers.  Most technical support teams will ask customers to use their phone cameras and send the image of the device needing tech support and AI assisted agents will use the image and determine the next set of actions.  The interaction between AI agent and BPO agent will be very interactive and the human agents need to be properly trained and upskilled.  These will require rebuild infrastructure, rebuild processes, rebuild tools etc.  The new AI world BPO will be very different from what we see in today’s digital world.

3.       Engineering Services:  In the new world, all products will be more software and AI-driven than hardware.  E.g. most cars will have many capabilities such as park assist, lane discipline, and reverse assist kind of modules of autonomous cars.  Most heavy equipment such as Power turbines, engines, etc. will have self-diagnostic and even healing capabilities.  Hence major upskills, and reskill are required for all the engineering services staff.

4.       IT Development & Support Services:  A lot is written about this everywhere and I will cover only a few key points.   The Enterprise to AI Model integration poses new additional challenges. It is expected that end users will see many enterprise applications such as SAP, SFDC UI but they will be now supported by product vendor-supplied AI engines or enterprise-developed AI engines, or licensed AI  engines such as ChatGPT.  The integration is very different and complex compared to digital integration and needs careful handling.  Knowledge of audio, documents, images, and video files needs new skills as most digital developers are only used to structured data till now.  Huge emphasis will be on Data quality, data labeling, data anonymization, etc.  These skills are usually not in IT departments and they need to be hired or internally transferred from end-user departments.  IT Production support will have to deal with probabilistic and sometimes degrading performance from AI models.  The feedback mechanism from AI model to the ticketing system such as SNOW needs to be built.  An AI model unable to make a decision is a ticket sent to the AI development team to improve the next version of the model!.   Overall the jump from Digital to AI is long and needs appropriate reskilling and upskilling.

5.       Strategy and Risk Management:  This area also needs much rework.  The current digital IT strategy needs to be upgraded to an enterprise AI strategy and get the buy-in from all the stakeholders.  The quality control, testing, and risk management associated with the production rollout of these AI models is a very big area where lots of work has to be done.  We @Aithoughts.Org have launched Training, and consulting services to help enterprises define, execute, and manage the risks.

We will be announcing the invite for the next Thought leadership talk soon., Please do join in.  Please contact Diwakar Menon for any help you need from @AiThoughts.Org

More later,

L Ravichandran