9 out of 10 professionals say they prefer face-to-face discourse but digital communication is more important, simply because of distance limitation. Not nearly enough productive discussions take place without it. Although AI’s even more engrossing for a business on a cloud market scale, most implementations get off on the wrong foot with too much
money invested in vain. The universal question, then, is how exactly
should your business start using it?
Digital transformation starts with integrating AI into your business. In so doing, many companies develop an enterpise system with simplified SOPs and machine learning administrative sequences that updates everyone in real-time across the cloud. According to Salesforce.com, however, a lot more are simply applying AI to features in isolated legacy systems. Unsurprisingly, they fall short of aspired ideals and stray off the efficient rail. In either case, one thing is clear: integrating AI is a big investment, particularly if you don’t see it as a major infrastructure transformation. You’re basically changing the reality of how everything is done to a new level of unimaginable efficiency. What previously took an entire division days, now doesn’t take even a click. Of course, all investments come with risk that either makes or breaks the bank.
To implement AI correctly and stay in the black, data from isolated legacy systems need to be synchronized and converged into a single system. That means data and table discrepancies need to be identified, formatted and settled. Luckily for micro venders and SMEs, multinational technology giants like Amazon and Alibaba have already invested in expensive IoT infrastrcutures to provide e-commerce focused cloud computing, artificial intelligence and data streaming services, for cheap.
For internal restructuring in medium-large organizations, on the other hand, the AI requirement usually comes as either a shock or a nightmare. Organizations amnestied from the Digital Revolution that shifted away from paperwork, mechanical and analogue to digital electronics like smart devices, computers and digital record keeping; although previously spent the minimal in the technology, might find it a shock now to start implementing a new enterprise system or to sign up with an e-commerce marketplace service provider and run the entire business from there, altogether. For large organizations that embraced digital but got off on the wrong foot with isolated databases that don’t talk to each other, implementing AI the right way is probably a bigger nightmare as database jumbles and mismatching data collections of differing versions from scattered spread sheets and tables now have to be synchronized with everyone on the same page. Work is also cut out in the post-synchronization stage to test, train and develop new SOPs, all of which take immense time, resources and will power to get right against status quo preference for face-to-face leverage.
That being so, unsuccessful and misimplemented first-gos should never come as a surprise nor unwelcomed. Nobody likes unlearnable lessons of failure in vain! Let’s take a look at the best practices to get AI off on the right foot.
1. IT Assessment
Many companies still use legacy systems put in place ages ago which don’t support AI. Thus, an updated SWOT analysis is required to see what strengths can be maintained, what the major weaknesses are, what opportunities are presented in the mist, as well as what threats remain out there, and determine what is needed to develop a new, technologically suitable enterprise system. Developing AI, of which, may be costly and time-consuming to properly analyse.
2. Leadership Team
You might even discover an unpleasant surprise as you come to learn that some AI capability is actually already present in your organization but underutilized. Neglect being the byproduct of conflicting convenience, clear and direct communication from leadership is required to avoid confusion, prioritise the order of business and jump start the learning curve to ascertain invaluable experience and lessons from implementation.
The Business Case
Vijay Raghavan, executive vice president and CTO of Risk and Business Analytics, RELX, recommends answering the following questions before jumping on the bandwagon. The answers you get will determine the driving power behind organizational AI.
- Do I want to use AI to create better products?
- Do I want to use AI to market products faster?
- Do I want to use AI to increase efficiency and profit in ways beyond product development?
- Do I want to use AI to alleviate certain risks that comes with data privacy, cybersecurity, compliance requirement, and etc.?
3. Measure, Measure, Measure
Erik Schluntz, Cobalt Robotics Co-founder and CTO, says that AI testing is adamant. Project outcome assessment is must be clear, rigorous in every step as well as objectivity and thoroughly evaluated. Even though AI has amazing capabilities, there is no such thing as a silver bullet that can solve solve all and any problem in one shot. Machine learning remains the ultimate assistant that carries out work and achieves goals with optimal SOPs, however, within predefined parameters using limited capacity. As such, it can also create new solutions that no one has come up with before so might want to keep an open mind when examining existing strategies.
4. Education and Collaboration
Although AI is progressing rapidly and steadily, it still relies on data science skills. The problem, then, is recruiting and retaining experienced professionals with the right skill sets for AI development. On the other hand, to expand limited expertise, we need to have access to and update data science skills and knowledge for progressive learning. There are currently many programs, courses and certificates available online for everyone with intent, some are even free of charge through education hubs like Udemy, Udacity and many others. Udemy, of which, currently has over 4,000 enterprise customers and 80% of Fortune 100 companies that use the education platform for employee skill updating (Udemy for Business).
Promoting a culture of teamwork and better together across the organization matrix is also extremely important. Who is responsible for which part of development and how to coordinate development must be clearly understood and improvised. A key success factor to strategic AI implementation is change management. Prioritized goals and joint roadmaps with short- and long-term milestones, e.g., predictive analytics prior to machine learning development, can be utilized to get personnel across departments on the same page.
5. Recognition and Reward
Celebrate success! And don’t forget to share good news with the boss. Achievements in each and every stage are a stepping stone, from the small goals in the beginning that takes 8-12 weeks to accomplish, to more challenging goals and up the hill tasks down the road.
AI is like a marathon, not a sprint, of which once completed, continued development is required in order to get prepared for the next longer and more difficult race. Underestimate it and you will regret the amount of patience and mind power required to get it right. Leadership must realize that using AI is neither an overnight nor a one-off endeavor. To the contrary, implementing AI is an ongoing job that requires progressive update and fine tuning. Personnel is no less important. Various dilemmas will arise that require professional expertise to assess what AI can do and what it can’t. Decisions made in the process will guide development and define the scope and scale of the AI capacity your organization has, thus, it’s viability as the infrastructure that tells business success.
Compiled by SCG Logistics