Artificial Intelligence | News | Insights | AiThority
[bsfp-cryptocurrency style=”widget-18″ align=”marquee” columns=”6″ coins=”selected” coins-count=”6″ coins-selected=”BTC,ETH,XRP,LTC,EOS,ADA,XLM,NEO,LTC,EOS,XEM,DASH,USDT,BNB,QTUM,XVG,ONT,ZEC,STEEM” currency=”USD” title=”Cryptocurrency Widget” show_title=”0″ icon=”” scheme=”light” bs-show-desktop=”1″ bs-show-tablet=”1″ bs-show-phone=”1″ custom-css-class=”” custom-id=”” css=”.vc_custom_1523079266073{margin-bottom: 0px !important;padding-top: 0px !important;padding-bottom: 0px !important;}”]

AiThority Interview with Asaf Somekh, Co-Founder & CEO of Iguazio (acquired by McKinsey)

Asaf has been at the helm of the tech, data, and AI scene for almost thirty years. He is the founder of Iguazio, an AI platform that enables enterprises worldwide to develop, deploy and manage their AI and Gen AI applications in live business environments. Get more insights from Asaf from the complete Q&A:

Hi Asaf tell us about yourself and more about Iguazio. What inspired the platform? Take us through the acquisition journey and what the road ahead will look like?

Hi, my name is Asaf Somekh, I’m the co-founder and CEO of Iguazio, an AI company that was acquired by McKinsey last year. I’ve been in the tech, data, and AI scene for almost thirty years. In 2014, after the successful exit of Voltaire (which floated on NASDAQ in 2007 and was later acquired by Mellanox), I climbed Mount Kilimanjaro with one of my Voltaire co-founders. Together we decided to found Iguazio to address the challenge of operationalizing machine learning, orchestrating, scaling, and accelerating AI delivery to impact. We built and scaled the company, focusing on enterprise AI projects. In 2022 gen AI burst into our lives with the release of ChatGPT, making the problems of getting AI from pilot to production even worse.  We see a great need from the market to streamline the development, deployment, and management of GenAI applications across the organization at scale.  In 2023 Iguazio was acquired by McKinsey, and today Iguazio is a part of QuantumBlack, McKinsey’s AI arm.

We are very excited about the road ahead – We’ve been working together with McKinsey on new client projects, including leading Fortune 500 companies. We see a lot of demand and special interest within the Financial Services industry, a highly regulated industry, a fact that poses additional challenges for deploying gen AI at scale.  We’re continuing to serve our original pre-acquisition clients as well.

Also Read: How the Art and Science of Data Resiliency Protects Businesses Against AI Threats

How did this acquisition specifically enhance McKinsey’s GenAI capabilities?

McKinsey acquired Iguazio in January 2023, in the context of scaling and operationalizing AI. McKinsey research shows that 90% of AI projects never make it to production, and enterprises are missing opportunities to generate real business impact through AI initiatives.  Now, the Iguazio AI platform is part of QuantumBlack, McKinsey’s AI arm, and its AI and machine learning innovation hub QuantumBlack Labs. Iguazio, along with Labs’ 30+ other products, enables teams to codify domain knowledge into reusable assets that can be deployed into client environments and unlock new scales of impact.

What are some of the biggest pain points enterprises face when scaling and deploying ML and GenAI across the enterprise?

Enterprises face two main challenges when advancing from gen AI proofs of concept to live implementations within business environments:

  1. Gen AI Ops & Scaling – It is difficult to bring gen AI to production in an efficient and effective way.  The resources and infrastructure needed when you’re working in a live business environment, the level of complexity, the issues that arise – such as where do you deploy your application? How do you share resources across projects?  How do you build a feedback loop?  How do you centrally manage applications in production, to ensure company guidelines are being followed and that computing costs are being controlled? – All these are difficulties enterprises face every day when it comes to operationalizing and scaling gen AI.  LLM customization using sensitive data, GPU utilization, and hybrid deployments (including on-prem) are additional complexities in this category.
  2. De-risking & Governance – Gen AI brings with it a host of risks for the enterprise, such as securing the company’s private data, mitigating AI hallucinations and biases, making sure to adhere to regulations and compliance standards, etc.  Implementing guardrails, governance, and monitoring mechanisms are essential when deploying GenAI in a live business environment. Governing gen AI within the enterprise is critical also in preparation for regulations that are soon to be imposed such as the EU AI Act.

Gen AI made it easier to build pilots, but much harder to move them to production, widening the gap between potential and actual business value.  Enterprises need to take these challenges seriously in order to start closing this gap and see real ROI from their investments.

Also Read: Real World Applications Of LLM

Can you talk about some of the biggest risks that are associated with GenAI and what enterprises should be more aware of today?

  • Handling sensitive data – I think most enterprises today know better than to put their data into online gen AI tools, but most enterprises, especially in highly regulated environments, struggle with data privacy concerns every day.  It is important to be aware of PII removal techniques, options that exist to train your LLMs privately and deploy your Gen AI apps in your own environments, whether your virtual private cloud or even on-prem data center.
  • AI hallucinations – It’s no secret that gen AI apps do not always provide the right answer, but it’s sometimes difficult to tell when they’re wrong if you’re not a subject matter expert, and to monitor them at all times.  Think about the car dealership whose virtual agent sold a car for a dollar or the chatbot that invented a travel policy that the airline needed to honor.  These are serious risks that need to be taken into account, and here too there are ways to mitigate them, for example by customizing your LLM or improving data quality before you begin training.
  • Cost escalation – This may be one of the most overlooked risks.  As you scale gen AI, if your architecture is lacking, costs may skyrocket.  GPUs for example are often used inefficiently and are underutilized which result in the over-provisioning of these expensive computing resources.  It’s important to build a sound foundation and think about the right way to automate and scale your projects so that you don’t end up paying an arm and a leg for something that could have been done in a much more effective way.

What does the future of GenAI capabilities seem like for you: how will constant enhancements change the game across the tech sector?

I think we’re just scratching the surface of what gen AI will be able to do.

I believe we’ll see combinations of large foundational models with very task-specific ones – this combo will be able to improve accuracy on one hand while reducing risks and costs on the other.

I also believe that open-source models will play a key role in the ecosystem – we already see today that many applications, if built right, can achieve a high level of accuracy and impact just by using sets of open-source LLMs combined with classic machine learning and deep learning models.

Because of the great pace of change, we’ll see enterprises adopting open and flexible architectures, ones that enable them to swap components and weave in the latest technologies as they become available.

It’s an extremely exciting time to be a part of this industry, but also quite risky.  It is my hope that in the next couple of years, we’ll be able to look back at this time and be proud that we developed gen AI in a respectful, responsible way – for good.

Also Read: How Does Artificial Intelligence Drive Predictive Analytics Systems?

Before we wrap up, a few top myths you’d like to bust on GenAI?

Yes – Maybe a couple –

First of all, not everything needs to be solved with gen AI.  Sometimes there are simpler, more cost-effective ways of getting the same things done.

Sometimes, gen AI simply enables ‘developers to write bad code faster’ (and it’s not just developers, this can be applied to any practitioner using gen AI).  I would caution against using the technology blindly without establishing the right use case, guardrails, and standardized review processes.

Finally, just to reemphasize the importance of creating a sound backbone for your gen AI projects, with guardrails and central management.  This will make all the difference as you scale gen AI across the enterprise.

Thank you, Asaf, for your insights; we hope to see you back on soon.

[To share your insights with us as part of editorial or sponsored content, please write to]

Asaf Somekh is the co-founder and CEO of Iguazio, an AI company that was acquired by McKinsey in January 2023.

Asaf has been at the helm of the tech, data, and AI scene for almost thirty years.
He is the founder of Iguazio, who’s AI platform enables enterprises worldwide to develop, deploy and manage their AI and Gen AI applications in live business environments, drastically shortening the time required to create real business impact with AI.

Asaf maintains close engagement with leading customers in the enterprise, cloud, federal and the system/software vendor community.

Iguazio (acquired by McKinsey) provides an AI platform which enables enterprises to develop, deploy, and manage ML and GenAI applications in live business environments.  Through automation and orchestration capabilities, the platform drastically shortens the time required to create real impact with AI.  Using Iguazio, organizations can develop AI applications at scale and in real-time, deploy them anywhere (multi-cloud, on-prem, or edge), mitigate risks associated with gen AI, and bring to life their most ambitious AI-driven strategies. Enterprises spanning a wide range of verticals use Iguazio to solve the complexities of implementing and scaling ML and gen AI across the enterprise, in an efficient and scalable way. Iguazio is used for a multitude of use cases such as real-time call center agent copilots, chatbot automation, fraud prediction, real-time recommendation engines, and predictive maintenance.  Iguazio was acquired by McKinsey in 2023 and is now a part of QuantumBlack, McKinsey’s AI arm.  Iguazio brings AI to life.

Comments are closed.