The need for an AI strategist
Bridging the gap
In the latest CIO/IDG research that surveyed more than 200 IT executives of large companies, majority of the companies agreed AI had the massive potential to drive disruptive innovations across most enterprises.In fact 90% of them are investing in AI.
In reality only a very few of them succeed.As per the report the top reasons for failure are,
- Data-related challenges hinder 96% of Organizations from achieving AI.
- Technology skills, leadership, and lack of a cohesive strategy are the biggest hurdles faced by data engineering and data science.
- Increasing Complexity:Organizations Invest in an Average of Seven Different ML Tools.
Source:"Conquer the AI dilemma by unifying data science and engineering" by DataBricks
This of course is a very IT technology implementation oriented point of view.The suggested solution is to create a unified analytics platform which fosters better collaboration between data scientists and engineering.
But before rushing ahead and trying to implement a unified analytics platform consider the findings of the strategy consulting firm McKinsey&Co. in it’s article regarding “How advanced industrial companies should approach AI strategy”
A sobering insight is that only 11% of the top management considers AI technology as a top priority. In spite of overwhelming support from IT executives(per the CIO/IDG research) the rest of the executive team do not seem to be buying the AI story.
The disconnect stems from the top management being unclear about the value AI brings to the table. In spite of their enthusiasm for AI implementation the CIO/CTOs are unable to communicate the economic value add of AI transformation to the rest of the management team.
Bridging this gap is exactly the job of an AI strategist.
First and foremost the AI strategist should have 3 core competencies,
Deep understanding of the business domain: It is important to have a keen understanding of the business.A fundamental grasp of the business’s customers,sources of revenue, costs,operations etc.(framed very well using the Business model canvas) will help the strategist understand and prioritize where in the value chain AI implementation will have the most impact.
At the EmTech 2018 conference by MIT technology review in SanFrancisco, Jia li,the Head of R&D at Google was one of the headline speakers.She in essence said, the technology of creating AI software and hardware was being taken care of, but what was missing was the domain knowledge to use these powerful tools.
This is where the Data scientists and Engineers fall short.Most of them do not understand the business problem to be solved.
Knowledgable grasp over the fundamentals of Machine learning : The AI strategist should have an understanding of the fundamentals of the math behind Machine learning and Deep learning.Experience building models from scratch will give an appreciation of what it takes to build models in terms of time and effort.It will also build an intuitive understanding of the limitations of machine learning.
Having an understanding of the AI language will make it easier to communicate with the data scientists and engineers and make the bridging of the gap between IT and Business that much easier.
An understanding of the Machine learning technology stack in terms of hardware and software will give the AI strategist a better picture of what part of the technology has to be built in-house versus outsourced.
Ability to separate AI myth versus reality:Most of the disappointments in AI implementations arise because of misguided expectations.The runaway imagination of Hollywood and popular media has created a near mythical performance expectation from AI systems.
An astute AI strategist will be able to articulate in solving a business problem how much will be done by an AI system and how much by the human system.
Setting the right expectations is only possible if the AI strategist has a keen understanding of both the business and the technical capabilities of AI
Having the above 3 core competencies enables an AI strategist in collaboration with the IT and Business teams to come up with a realistic roadmap for AI implementation.
In co-ordination and with the sponsorship of the Executive team the AI strategist has to ensure the right digital infrastructure(digital base and capabilities), right talent and right expectations are in place for a successful outcome.
In my opinion no such formalized role of an AI strategist exists in today’s corporate world.The closest role that could come close is that of a Product manager.However the Product managers are limited to the narrow view of their product only.They do not have a deep understanding of the working of the different parts of the organization and thus will miss out on multiple areas of economic value add. A person who understands the working of the organization at a corporate level and with some technical background in machine learning will be able to transition into the role of an AI strategist.
Just as CIO was a role created with the advent of big data, it is inevitable that in the near future we shall have the role of Chief AI Officer.But we can first begin with the AI strategist.