The world of artificial intelligence is hyperactive and constantly evolving. Is there a day that you don't hear about #IA ? Definitely not since ChatGPT and if you follow our news feeds :-) (blog biZNov, linkedin biZNov, twitter @biZNov_fr).
Regular/frequent technology watch and perpetual questioning is the state of mind that must be displayed to develop and implement a winning strategy adapted to your company. It is because AI has demonstrated its positive impact on business development and competitiveness in many areas, and increasingly, that it is worth investing the energy necessary for this constant monitoring.
Today, there are more and more "off-the-shelf" solutions, in SaaS mode, with "ready-to-use" features. This makes adoption much easier.
There are of course many service companies or ESNs that offer you to build a custom AI.
And then you have the option to build your solution with your team but how do you get started and what skills will you need in the short, medium and long term?
Why should you move towards one solution over another? You must first consider your business strategy, its valuation, your customers, and your ways to differentiate yourself from your competition. And what are your ambitions?
Is there an off-the-shelf solution that suits your company's unique needs? An impressive number of solutions are already listed: 590 companies at the beginning of 2023 for France only (see this article from France Digitale: startups specialized in AI). How do you identify the right solution among this forest of existing solutions?
What level of adaptation is required? How does this solution enable your employees to be part of the performance? Does it include an active improvement loop under the control of your teams? Which solution is tailored to your company's data volume? Which infrastructure to use (on-premises, cloud, hybrid), for the learning phase, for inference? Once deployed, and your company is threatened with becoming a unicorn (and this is what we wish you ;-) ), will you have to redo everything so it scale?
And then, assuming that we have found the ideal solution on paper, how much it costs to develop, adapt, deploy and also maintain. Since the data from your company's environment is not fixed, how often should you update the prediction/classification/aggregation models used as a basis for the new services you will create?
If you need a tailor-made solution, who is the right partner to process your data of yesterday and tomorrow? Does this partner use state-of-the-art algorithms or that are best suited to your constraints? Are its teams tailored to listen to you, with the appropriate AI/ML expertise? Are relevant open source tools used?
Are the solutions or service providers used able to guarantee the security and confidentiality of your company data and that of your customers?
And tomorrow, will the algorithm used to train your models still be the best answer to your need or will you have to migrate to better performance? Will the initial return on investment still make sense or justify stopping or evolving your services?
All this represents a lot of questions to ask ourselves, but it is because we ask ourselves the right questions, we find the right answers and we give ourselves the best chance of success (in 2020, 50% of AI projects fail).
It is still very complicated to adopt the #artificialintelligence, deploy it and reap all the expected benefits. It is a nascent industry with reach in all areas, which explains the abundance of information on the subject.
Comments