Olivier Blais is Co-Founder & Head of Decision Science at Moov AI. He will be speaking at Dublin Tech Summit on 15 & 16 June, which returns to the RDS in a physical capacity. See all confirmed speakers here.
What are some of the new trends in tech?
Despite all the new concepts such as NFTs, metaverse, blockchain and so on, I believe that the most game changer trend is that companies are starting to adopt AI in their organisations. Despite all the negative sides of COVID-19, this pandemic period was an excellent enabler for organisations to think about automation and digital transformation. AI is just a tool, but it can be transformative when you integrate it well into a process. For example, if you become fantastic at predicting your customer’s demand, you can streamline the full logistics, supply chain and purchasing functions accordingly. That is huge for companies.
How can smaller companies get a foothold in the tech industry?
The secret is to demonstrate that small companies can generate value fast. I believe that the advantage of smaller companies is delivery speed and innovative thinking. For more prominent organisations, to partner up with smaller companies is difficult otherwise. Another secret for smaller companies to grow is to multiply success stories. If a company solves a massive problem for a client, it becomes a relevant supplier no matter the size of that company.
What are some of the biggest concerns facing the tech community today?
I believe that every tech company should address the following problems so that we can ensure AI adoption:
- About 50% of organisations are blocked in their AI adoption because of data issues (AI Forum, 2022). What can we do about it? How can we deliver solutions less dependent on a high quantity of data?
- About 60% of organisations do not have enough knowledge or talent to generate value with their data (AI Forum, 2022). How can we make this very simple while ensuring safe data usage?
- In 2022, 85% of AI projects will yield erroneous results due to biases in the data, models, or teams responsible (Gartner, 2020). How can we develop approaches and tools to validate the quality of an AI system before it gets released? Most data science teams still rely on ad-hoc basic performance assessment, but it must become more sophisticated. Software development evolved its testing approaches 10-15 years ago. AI should do the same soon to avoid the fallbacks of a system of poor quality being deployed in production.
What key messaging do you want the DTS community to hear?
Simply put, artificial intelligence is a great tool to support the transformation of critical functions of top organisations. However, shifts do not happen overnight, and the usage of an instrument is usually not sufficient. Therefore, these tips optimise the success of a transformative project using AI:
- Start small and build incrementally on success stories; otherwise, you will lack traction.
- Use simple solutions to resolve complex problems; many existing and robust machine learning frameworks and algorithms exist, so make sure you reuse as many as possible.
- Think about delivering a system and not just a model; machine learning code is typically barely 20% of the code base of an AI system.
- While developing your AI system, taking change management and training into account is paramount, as an unused AI system is useless.
Now in its 5th year, DTS 2022 will see over 8,000 registered attendees, 200 speakers, and 65 partners representing 50 countries, the event, which has taken place virtually over the past two years due to the Covid-19 pandemic, promises to be bigger and better than ever before. It will bring together the brightest tech innovators, influencers and entrepreneurs globally, highlighting the capabilities and creativity which form part of a wider tech ecosystem. For Dublin Tech Summit themes, see here.