9 Lessons learned from failed AI PoCs

Forums General News (General) 9 Lessons learned from failed AI PoCs

  • This topic is empty.
Viewing 0 reply threads
  • Author
    Posts
    • #40244
      Telegram SmartBoT
      Moderator
      • Topic 5959
      • Replies 0
      • posts 5959
        @tgsmartbot

        #News(General) [ via IoTGroup ]


        Headings…
        9 Lessons learned from failed AI PoCs
        Data
        Compliance
        Realistic Expectations
        Scalability
        Size and nature of your PoC
        Implementation process
        AI Accuracy / Available Data
        PoC Evaluation
        Time Window

        Auto extracted Text……

        When working with organizations, I noticed that only a fraction of decision-makers fully understand the importance of having a good dataset by that I mean a well-curated collection of data to train an AI system.
        I believe that if your company is struggling with too much data quality issues then it is safe to think that the company isn’t ready for AI yet.
        Based on my experience, companies pursuing ML for the first time generally lack an expert understanding of what data is needed for machine-learning models and struggle with preparing data in a way that’s beneficial to those systems.
        Data ownership is heavily underestimated in AI projects.
        A successful project takes time, resources, and plenty of relevant data.
        The truth is that if your business doesn’t have a clean dataset then I strongly recommend you to start with some basic data cleansing and organizing before launching an AI PoC.
        Your PoC can fail for so many reasons (lack of data, AI accuracy, etc.) and that’s why you need to focus on concrete projects with a real added-value.
        Some companies can afford the cost of working on the research side of AI, but for the majority of them what’s really needed are applied-side efforts, or projects focused on how to make use of AI.
        In an AI project, the implementation process can be impacted by your models.
        Let’s assume that your objective through this AI PoC is to provide real-world insight then it is more important that your model provides a strong explanation of what it is doing than generating highly accurate predictions.
        AI Accuracy / Available Data
        In my latest project (AI image recognition), we reached a 95% accuracy but it wasn’t enough for the company… Companies have to evaluate their confidence versus risk, and the cost of being wrong


        Read More..
        AutoTextExtraction by Working BoT using SmartNews 1.02976805238 Build 26 Aug 2019

    Viewing 0 reply threads
    • You must be logged in to reply to this topic.