#004 - AI from a business consultancy perspective

June 4 2020/45 min read

We are joined by both Christian Landgren and Isabelle Holm. Christian is the CEO of iTeam, and Isabelle is the Strategy Director at Reaktor. What they have in common is their extensive backgrounds in assisting organizations to embrace and implement emerging technologies. Together, we discuss questions such as: What phases of AI readiness does organizations experience? How should organizations structure its processes and teams to succeed with AI? What are the hurdles for AI deployment you must overcome?

(The discussion is in English)

Key takeaways from this episode

There are levels of an organization's AI readiness
  • Three levels organizations experience? 1) Where should we begin? 2) Are we behind everyone else in development? 3) Is there really value in AI implementation? 
  • Organizations are moving towards trying to understand the value AI will have. Three or so years ago many organization were located in the first phase

Embrace the mindset of learning. Throughout the organization
  • It requires time and effort to learn about AI, to then harness the benefits
  • Educators and courses make AI more simple, but it can't be oversimplified. You can not force complex problems to be simple
  • The understanding must go through all levels in organizations to create real value

The key for an organization is to learn to ask the right AI questions
  • Starting out with small corner case AI cases is okay
  • The valuable aspect is for the organization to begin asking the right questions AI can help with, size of the problem will increase over time
Don’t throw the problem at your technology department
  • Websites, apps, blockchains, and AI are all general technologies. It takes time to get value from them, dont ust throw the problem to the IT department
  • Focus on experimenting to find value for your organization
  • It is doomed to fail if you simply do it to be on the bandwagon

Experiment as a R&D organization, but involve all departments
  • R&D heavy organizations are more equipped to buy AI, as they are used to spend resources on iterative experiments
  • This is not the case for more traditional service companies
Maturity in tools for building AI is improving
  • Open source tools makes sense to use in all organizations
  • Custom made tools in AI might be less expensive than those in earlier IT projects
80% of service time around AI is cleaning the data.
  • Focusing on questions such as where is the data stored, and what is the quality of it makes the time spent on getting value of AI much lower

Changes in the external environment might imply changes needed for AI models
  • External events, such as the corona pandemic, alters humans behaviour. This makes it difficult for recommendation engines to keep up