How to ensure that the AI solutions developed will be used? How to scale up and ensure the adoption of AI within the enterprise? How to prepare the next steps and enable scalability? We take stock of these questions in this article.
Artificial Intelligence (AI) has been invited to company meetings for several years. While there is no lack of proof of concept and pilot projects, for many, going to scale is like jumping into a void. However, the benefits for those who manage to successfully deploy AI at scale are real. Accenture is seeing a three-fold return on investment in strategic scale-up.
The tricky step of going to scale
Most companies now accept to integrate artificial intelligence into their operating mode. According to Capgemini's latest study on the subject, published in July 2020, 60% of French organizations have already passed the proof of concept or pilot project stage. For three quarters of them, the results obtained meet or exceed expectations.
Nevertheless, getting past the proof-of-concept stage and scaling up to enterprise scale is still difficult for many. To avoid the classic pitfalls at this stage, we recommend that you consult our article: "Making a good Proof of Concept in Big Data".
The key to scaling up lies in implementing and communicating a production-oriented strategy from the very beginning.
Adopt a production-oriented logic from the proof-of-concept stage
Many pilot projects will never reach the production phase. This is completely normal. Nevertheless, a company must consider a possible scale-up at the design stage.
Constraints in terms of resources, infrastructure, personnel, or time must be communicated and known. A multilateral upstream approach involving the various stakeholders within the organization will ensure efficiency right from the proof-of-concept stage.
Successfully scaling up: a few keys
Many consulting firms - including Capgemini and Accenture - are working to identify the characteristics of a successful scale-up. In this context, Accenture evokes the concept of "strategic scaling”.
There are generally certain constants that can be observed in companies that benefit from AI:
- A clear and focused strategy
- A clear understanding of the data and the value that can be extracted from it
- Competent and visionary AI managers who integrate ethical issues into their approach
- A collective effort at the level of the organization as a whole
We detail these elements in the following paragraphs.
A clear and targeted strategy
To go to scale in AI, knowing the objectives pursued is imperative. The deployment of artificial intelligence only makes sense if it has an added value.
If for many organizations this phase of reflection may seem superfluous, it will subsequently save considerable time, money, and resources.
Ask yourself the following questions:
- What means am I ready to allocate to AI?
- What goals would I want to achieve with AI?
- What are my imperatives and priorities (speed, budget, design, etc.)?
- Which projects do I want to focus on?
- How will I control my AI once scaled up and measure performance (or benefits)?
- What steps in the production, supply chain or marketing process am I ready to automate?
- What are my competitors doing and what competitive advantage would I want to gain?
- Which cells does AI impact within my company?
Moreover, a good AI strategy will almost always involve a good data governance strategy.
A good understanding of data
At the basis of any artificial intelligence application is data. The quality and reliability of the data are of paramount importance to allow optimal scaling. Data governance is gaining momentum within organizations.
In this context, the distinction between usable data and used data is fundamental. Many people believe that a large amount of data is required for artificial intelligence to function properly. This is a common misconception. In practice, it is the relevance and quality of the data that prevails. An AI must be able to extract value from data for a defined purpose. This implies a reflection on the nature of the data to be provided to artificial intelligence to enable it to obtain results.
A competent AI staff and an ethical vision
At the start of a winning company, there is a winning team. Profiles competent in artificial intelligence are highly courted. According to Capgemini, 70% of organizations deplore the shortage of talent in the field. People in charge of AI must also can develop a global vision but above all be pragmatic.
More surprisingly, it is also noted that the consideration of ethical issues by AI managers allows a more efficient scale-up. The mastery of ethical issues vis-à-vis customers or employees would lead to increased trust and better adoption of AI applications.
A collective effort at the enterprise level
Whether we are talking about big data, machine learning or artificial intelligence, teamwork and a collective vision at the corporate level will often guarantee a smooth and efficient transition to scale.
The entire organization must be involved in the transition. A multidisciplinary approach is required. Generational differences in perspective should not be overlooked. A company that wants to make a successful transition to scale must also allocate resources to training and communication.
The idea is that employees learn to view artificial intelligence as an asset, not a threat. In addition, it is important to provide them with easy-to-use and intuitive tools so that they can become familiar with artificial intelligence applications and understand how they work.
A data science platform: a real asset
Some tools exist to facilitate the transition to scale. Among these, the data science platform is becoming essential. This graphical modeling tool allows any organization to deploy its AI applications intuitively.
The use of a data science platform also simplifies issues related to the scalability of applications after they have been scaled up.
Ryax offers an open-source data science platform that gives your organization the agility it needs. With our unified framework, you can move from the design phase to the production phase more efficiently.
La Ryax Team.