AI in business: insights from Blackstone's Principal Yasser Ezbakhe
💡 muffin.ai: the media to understand the challenges of AI and leverage them in our society and jobs — by a collective of French engineers & entrepreneurs.
Hello everyone!
We are excited to publish a new deep dive into the fascinating world of artificial intelligence with visionaries. 🙌
Your voice matters a great deal to us. Share your thoughts, questions, and suggestions with us. 🙏
Today we are sharing our conversation with Yasser Ezbakhe, Principal Data Science at Blackstone.
🧁 We love Yasser’s insights on identifying AI use case in your business, how to approach it and sharing his experience on where AI projects usually fail.
Thanks Yasser for the interview!
On the menu today:
🎓 Yasser journey from Polytechnique to Blackstone
⚙️ The Real-World Magic Behind AI
💡 Use Case Mapping and Prioritization: A Top-Down Approach with Real Benefits
💎 Building an AI Team: From Technical Mastery to Business Savvy
🤝 Governing, Executing, Monitoring: The Framework for AI Success
👀 Some sources recommended by Yasser
If anyone shared this newsletter and you want to subscribe, it’s here:
🧁 MF: From academia to Blackstone leadership, which experiences were key in shaping your AI expertise?
👉 Yasser:
I work at Blackstone, the world’s largest alternative asset management firm which manages over $1 trillion of assets. I am a Principal in the Data Science group where I lead the firms Private Equity practice.
At Blackstone, we see AI-enablement as long-term investment theme and a tool for building stronger businesses. In my role in the 50-plus person data science team, I leverage the latest Analytics and ML models to support deal teams, and work with our portfolio companies on AI strategy and integration.
Before joining Blackstone, my journey began at Ecole Polytechnique (Paris) and Columbia University (New York), where I respectively earned a Master of Engineering in Fluid Mechanics and a Master of Science in Data Science.
I used this academic foundation as a launching pad for my career in the tech industry, where I gained hands-on experience across the AI spectrum—from database management to AI model deployment. This period solidified my core technical skills and gave me the confidence to tackle complex challenges.
My next stop was McKinsey, where I served as a Data Scientist. This role allowed me to bridge the gap between the technical and business worlds effectively. Finally, my current role at Blackstone has broadened my skill set, adding an investment and macro-level perspective to my toolkit.
🧁 MF: What's your recommended playbook for companies venturing into AI projects?
👉 Yasser:
According to a Gartner study, around 85% of AI projects fail to deliver any value. So, the question many are asking is how do businesses make AI a core part of their DNA? In my experience, the answer lies in a pragmatic three-step process:
1️⃣ mapping and prioritizing use cases
2️⃣ assembling the right AI team
3️⃣ effectively governing and executing the AI project.
I believe this straightforward approach helps to demystify the complexities of AI.
At Blackstone, we’ve been building the capabilities needed to capture the AI opportunity since 2015. This includes introducing a data science team which has developed AI technology that aims to position Blackstone at the forefront of our industry.
🧁 MF: How should leaders transition from a starting point to successful AI implementation?
👉 Yasser:
When it comes to integrating AI into a business, my main reflection is that it isn't just a task for the technology team. C-suite executives play a pivotal role in embedding AI and facilitating the collaboration required to identify how AI can add value.
Prioritization is a multi-faceted process. Teams should evaluate the potential use cases based on a number of criteria: how much revenue could be generated, what costs could be saved, and how well the project aligns with the company's strategic vision. Other practical considerations include feasibility, which encompasses the availability of data, technology, and talent.
For instance, let's take a Chief Operating Officer (COO) who sees the opportunity to use AI to automate some of its processes. By leveraging a GenAI model (e.g., ChatGPT), the company could more effectively ingest, analyze and respond to existing customers support ticket. If this initiative could make your call-centers more efficient by 10%, you could reallocate this time to more complex issues and more valuable customers.
🧁 MF: How do you approach building a successful AI team?
👉 Yasser:
Building an effective AI team isn't just about technical skills; it's about aligning those skills with the broader business objectives, all under the guidance of a leader who can navigate both worlds effortlessly.
Assembling the right multidisciplinary team is crucial for a successful AI transformation. The keystone is a strong AI leader who not only has a solid technical background but also possesses experience in large-scale transformation projects. This is critical because AI leaders must be adept at translating business problems into technical solutions and then, crucially, conveying the business value of those solutions back to the business.
This multidisciplinary team should also include data engineers, data analysts, data scientists, and machine learning engineers. But technical process isn't enough—each team member also needs to understand the business deeply. This close connection to the business ensures that the AI solutions developed will not only be technically sound but also highly relevant and impactful.
🧁 MF: How can teams ensure success in AI projects and steer clear of common pitfalls?
👉 Yasser:
Effective governance, agile execution, and rigorous monitoring are the triad that can transform an AI project.
Governance is incredibly important – the best tip I have is to set up a Steering Committee (SteerCo) which includes a project sponsor, C-suite executives, board members, and relevant business stakeholders.
In addition, the working team should be a balanced mix of data scientists and business operators. AI projects are usually 20% modeling and 80% communication, business processes and change management. Daily standups, weekly updates, and monthly SteerCo meetings to review progress are all good practice.
When thinking about the project execution – it should ideally be agile. An iterative process allows the team to adapt to unforeseen challenges, such as changes in data availability or customer behavior.
Finally, keep an eye on performance. Both operational and financial Key Performance Indicators (KPIs) should be monitored continuously. For example, if a consumer app introduces a new AI feature to optimize product recommendation, operational KPIs might include the number of successful discoveries, daily active users, and engagement time on the platform. Financial KPIs could cover new sign-ups, customer cross-sell, retention and churn rates, and increases in revenue.
🧁 MF: Which AI sources and references do you rely on regularly?
👉 Yasser:
📧 I love "This Week in Startups" (TWIST) Ticker. It is a valuable resource for staying updated on the latest happenings in the AI and tech world. The newsletter provides a digest of key tech investments, breakthroughs, and events, making it a convenient way for professionals and enthusiasts alike to keep their finger on the pulse of the industry.
🎓 For learning and developing new technical skills, I really like online courses. For example, Coursera is a great platform where I attended all the Andrew Ng classes (e.g., prompt engineering for developers). There is also Hugging Face where, in addition to importing the latest open-source models, you can learn from the community (e.g., reinforcement learning).
Enjoy your week !
— muffin.ai team
If you enjoyed the edition, consider clicking on the ❤️ button and leaving a comment so that more people can discover muffin.ai on Substack 🙏