How to achieve high ROI on AI – Part 1


AI is the new Big Data 🙂

We are in the golden age of AI which does cause FOMO and a general sense of the AI hammer being used to smash (learn to smash?) every nail.

While the scale and nature of FAANG and a few other businesses justify investing heavily in AI for any edge ( a 0.001% improvement in any measure is hundreds of millions of dollars of impact) the rest of the business world is struggling to gain from practical applications of AI.

That is where well thought out use cases with cross-functional teams really shine out. A cohesive team comprised of business embedded SMEs, strong UI developers in tune with business processes, technologists who can roll out complex, scalable systems are key to rolling out high ROI technology projects. AI projects are no exception.
Yes your data scientist is the star of the show and yes they can produce very compelling research to fund a project but it takes a team to deliver results. Beatles won’t be a show if only John was singing.

Key takeaways from our experience delivering AI-enabled successful projects to production:

  1. Know your requirements.
  2. Analyze what is needed to deliver value: AI -> ML -> Statistical models.
  3. Create a cohesive team to iterate over the said requirements and deliver incrementally.
  4. Create a logical architecture, able to scale appropriately as your usage grows.
  5. Agile Devops, auto-scaling, easy access to frameworks -> if you cannot manage this you are bound to fail.
  6. Continuous business Validation, know when you are in your happy place 🙂
  7. If you are stuck in a local minima with your data scientist, take a step back and deliver value now.

Stay tuned for my next post in which we will cover a large, real-world use case where we continue to deliver real value to our clients using AI / ML / NLP.


Please enter your comment!
Please enter your name here