Read our blog with Open AI here:
https://openai.com/index/model-ml-chaz-englander/
OpenAI's Executive Function series features perspectives from leaders on the frontier of AI adoption.
A conversation with Chaz Englander, CEO & Co-founder of Model ML.
Model ML is building AI infrastructure transforming how leading financial services firms operate. Model ML’s platform features purpose-built agents and an application that automates end-to-end workflows as well as bespoke research and analysis.
OpenAI recently spoke with CEO and co-founder Chaz Englander about how financial firms are evolving, and how recent AI advances are automating and streamlining their operations.
What was our first meaningful encounter with AI, and how did it influence the creation of Model ML?
After selling our last company, my brother and I realized we didn’t like investing but became obsessed with automating the investment process through GPT‑powered function calling.
We were a six-person family office, but with these GPT‑3.5-powered LLMs, it felt like we had the leverage of a 60-person team.
We built a prototype of Model ML for ourselves and didn’t plan to commercialize it. But once we saw the insight gains and efficiency from automating research workflows, we knew we were onto something.
What are we seeing change inside financial services firms?
There are tasks that historically used to take days, weeks, or even months, and some of those now can be done in minutes or hours. For example, preparing quarterly earnings summaries used to take hours. Now, agents handle this entire process: they pull the data, format the slides, and publish the Powerpoint to SharePoint, all without human intervention. I think that’s going to be the biggest shift we see this year… that you’re going to come in in the morning and your work will already be there.
“I think that’s going to be the biggest shift we see this year… that you’re going to come in in the morning and your work will already be there.”
That’s forcing a rethink of where humans add value, and how companies are going to need to remap where teams are going to be impactful both today and in future.
We’re seeing firms shift people into higher-value, judgment-based roles. The leaders of the companies that we work with, at least in our view, are the ones rethinking the architecture of the whole organization in an AI-native way. It’s incredibly difficult, which is why we find ourselves acting as consultants early on, helping them determine where AI is most applicable today and also futureproofing to where we think it’ll be most impactful in 12 months’ time.
“The leaders of the companies that we work with… are the ones rethinking the architecture of the whole organization in an AI-native way.”
We are seeing the people within financial firms are more impactful now, not less. With automation taking on the grunt work, people can focus on relationships and strategic thinking. The firms that win will be the ones who rethink their entire operating structure to take advantage of that shift.
How does Model ML stand out compared to general-purpose AI tools, and how are new model capabilities benefiting our customers?
In finance, accuracy, compliance, and workflow-fit aren’t optional–they’re table-stakes. That specificity is where general-purpose tools fall short. Model ML was purpose-built for financial services from day one at two critical levels.
First, at the agent layer, we’ve built and fine-tuned systems specifically to parse and interact with the kinds of data financial professionals use daily, structured and unstructured, across tools like Sharepoint, and common data sets like Capital IQ, FactSet, and Crunchbase, which can be hundreds of tables and 20 terabytes. 12 months ago, it was near impossible to build an agent on top of those datasets. These aren’t just models that answer questions; they are contextually aware, understand schemas, write code, and retrieve information from terabytes of complex data.
Second, there’s the application layer: the interface through which users interact with agents, specifically designed for finance. It provides firms with the tools to build agents that automate end-to-end workflows and enable previously unattainable analysis. In terms of use cases, we’re seeing dozens of new use cases daily and we’re at thousands of use cases, a lot of which customers get right out of the box when they sign up.
We’ve seen significant step-changes with every new model release that we’ve channeled into immediate benefits for our customers. The advances in things like reasoning and coding ability have sent areas of our product stratospheric. Most recently with the release of OpenAI o3‑pro, o3, o4-mini, and GPT‑4.1 models, these new models have brought dramatic improvements in reasoning, multimodal capabilities, instruction following and tool integration. With larger context windows and more advanced reasoning capabilities, we’re now able to unlock end-to-end workflows. Now, users can chain together data gathering, analysis, and presentation creation tasks, producing fully formatted outputs completely autonomously.
“The advances in things like reasoning and coding ability have sent areas of our product stratospheric.”
Looking ahead 12 months, what do we think will change the most?
I think the most profound shift ahead is the rise of end-to-end workflow automation, where your systems act as control towers overseeing an army of digital workers. As these agents take on more complex, multi-step tasks across your entire digital universe, even the UI and the way we interact with hardware will start to change. This is likely a step beyond the next 12 months, but it’s where we’re heading.
What’s coming next is the emergence of truly autonomous agents that you can build in our product. Our agents are able to execute sophisticated workflows that gather, analyze, and present data from your CRM, emails, files, external data vendors, meeting transcripts and more. These agents won’t just wait for instructions; they’ll anticipate what needs to be done, whether it’s cyclical (daily, weekly, monthly, quarterly, annually) or triggered by real-world events—just as you’d ask a team member after a meeting or in response to an email.
The real shift is that these workflows will run end-to-end, automatically, with deep reasoning and orchestration across all your systems. The outputs could be as substantial as a 100-page PowerPoint, built entirely by machine—faster, more consistent, and available 24/7.
This is the future: autonomous digital teams running the workflows that drive your business—better, faster, and always on.
How do we keep our team agile as AI evolves so quickly?
Our belief is that AI-native companies will look structurally different. Fewer layers, faster cycles, tighter feedback loops. We’ve embraced a flat structure. Arnie [my cofounder] and I each have double-digit direct reports. That might sound wild, but AI makes it manageable. All one-on-ones are AI-assisted. Notes, to-dos, context, it’s all streamlined. It lets us move faster and stay close to the product. We think that’s how modern companies will run: more like control towers than siloed hierarchies.
Part of being agile is betting on the ecosystem and foundational models getting better. The real key, and maybe this is part of the founder mindset and engineering organizations, is not being emotional about your code. We used to build everything ourselves: agent abstractions, service connectors, all of it. Now, if OpenAI or the open-source community ships something better—like the OpenAI’s Agent SDK or MCP connectors—we just plug it in and delete our code.
We have shifted to using OpenAI’s Agent SDK and MCP tooling to handle agent loops, tool calling, guardrails and integrations, and we’ve been able to move on with less maintenance for faster innovation.
We’re not trying to win by maintaining infrastructure; we’re trying to win by delivering value through customer outcomes.
Model ML uses OpenAI’s API platform, including GPT‑4.1, OpenAI o3, and the Agents SDK, to power its agents, automations, and internal tools.
Read our blog with Open AI here:
https://openai.com/index/model-ml-chaz-englander/
OpenAI's Executive Function series features perspectives from leaders on the frontier of AI adoption.
A conversation with Chaz Englander, CEO & Co-founder of Model ML.
Model ML is building AI infrastructure transforming how leading financial services firms operate. Model ML’s platform features purpose-built agents and an application that automates end-to-end workflows as well as bespoke research and analysis.
OpenAI recently spoke with CEO and co-founder Chaz Englander about how financial firms are evolving, and how recent AI advances are automating and streamlining their operations.
What was our first meaningful encounter with AI, and how did it influence the creation of Model ML?
After selling our last company, my brother and I realized we didn’t like investing but became obsessed with automating the investment process through GPT‑powered function calling.
We were a six-person family office, but with these GPT‑3.5-powered LLMs, it felt like we had the leverage of a 60-person team.
We built a prototype of Model ML for ourselves and didn’t plan to commercialize it. But once we saw the insight gains and efficiency from automating research workflows, we knew we were onto something.
What are we seeing change inside financial services firms?
There are tasks that historically used to take days, weeks, or even months, and some of those now can be done in minutes or hours. For example, preparing quarterly earnings summaries used to take hours. Now, agents handle this entire process: they pull the data, format the slides, and publish the Powerpoint to SharePoint, all without human intervention. I think that’s going to be the biggest shift we see this year… that you’re going to come in in the morning and your work will already be there.
“I think that’s going to be the biggest shift we see this year… that you’re going to come in in the morning and your work will already be there.”
That’s forcing a rethink of where humans add value, and how companies are going to need to remap where teams are going to be impactful both today and in future.
We’re seeing firms shift people into higher-value, judgment-based roles. The leaders of the companies that we work with, at least in our view, are the ones rethinking the architecture of the whole organization in an AI-native way. It’s incredibly difficult, which is why we find ourselves acting as consultants early on, helping them determine where AI is most applicable today and also futureproofing to where we think it’ll be most impactful in 12 months’ time.
“The leaders of the companies that we work with… are the ones rethinking the architecture of the whole organization in an AI-native way.”
We are seeing the people within financial firms are more impactful now, not less. With automation taking on the grunt work, people can focus on relationships and strategic thinking. The firms that win will be the ones who rethink their entire operating structure to take advantage of that shift.
How does Model ML stand out compared to general-purpose AI tools, and how are new model capabilities benefiting our customers?
In finance, accuracy, compliance, and workflow-fit aren’t optional–they’re table-stakes. That specificity is where general-purpose tools fall short. Model ML was purpose-built for financial services from day one at two critical levels.
First, at the agent layer, we’ve built and fine-tuned systems specifically to parse and interact with the kinds of data financial professionals use daily, structured and unstructured, across tools like Sharepoint, and common data sets like Capital IQ, FactSet, and Crunchbase, which can be hundreds of tables and 20 terabytes. 12 months ago, it was near impossible to build an agent on top of those datasets. These aren’t just models that answer questions; they are contextually aware, understand schemas, write code, and retrieve information from terabytes of complex data.
Second, there’s the application layer: the interface through which users interact with agents, specifically designed for finance. It provides firms with the tools to build agents that automate end-to-end workflows and enable previously unattainable analysis. In terms of use cases, we’re seeing dozens of new use cases daily and we’re at thousands of use cases, a lot of which customers get right out of the box when they sign up.
We’ve seen significant step-changes with every new model release that we’ve channeled into immediate benefits for our customers. The advances in things like reasoning and coding ability have sent areas of our product stratospheric. Most recently with the release of OpenAI o3‑pro, o3, o4-mini, and GPT‑4.1 models, these new models have brought dramatic improvements in reasoning, multimodal capabilities, instruction following and tool integration. With larger context windows and more advanced reasoning capabilities, we’re now able to unlock end-to-end workflows. Now, users can chain together data gathering, analysis, and presentation creation tasks, producing fully formatted outputs completely autonomously.
“The advances in things like reasoning and coding ability have sent areas of our product stratospheric.”
Looking ahead 12 months, what do we think will change the most?
I think the most profound shift ahead is the rise of end-to-end workflow automation, where your systems act as control towers overseeing an army of digital workers. As these agents take on more complex, multi-step tasks across your entire digital universe, even the UI and the way we interact with hardware will start to change. This is likely a step beyond the next 12 months, but it’s where we’re heading.
What’s coming next is the emergence of truly autonomous agents that you can build in our product. Our agents are able to execute sophisticated workflows that gather, analyze, and present data from your CRM, emails, files, external data vendors, meeting transcripts and more. These agents won’t just wait for instructions; they’ll anticipate what needs to be done, whether it’s cyclical (daily, weekly, monthly, quarterly, annually) or triggered by real-world events—just as you’d ask a team member after a meeting or in response to an email.
The real shift is that these workflows will run end-to-end, automatically, with deep reasoning and orchestration across all your systems. The outputs could be as substantial as a 100-page PowerPoint, built entirely by machine—faster, more consistent, and available 24/7.
This is the future: autonomous digital teams running the workflows that drive your business—better, faster, and always on.
How do we keep our team agile as AI evolves so quickly?
Our belief is that AI-native companies will look structurally different. Fewer layers, faster cycles, tighter feedback loops. We’ve embraced a flat structure. Arnie [my cofounder] and I each have double-digit direct reports. That might sound wild, but AI makes it manageable. All one-on-ones are AI-assisted. Notes, to-dos, context, it’s all streamlined. It lets us move faster and stay close to the product. We think that’s how modern companies will run: more like control towers than siloed hierarchies.
Part of being agile is betting on the ecosystem and foundational models getting better. The real key, and maybe this is part of the founder mindset and engineering organizations, is not being emotional about your code. We used to build everything ourselves: agent abstractions, service connectors, all of it. Now, if OpenAI or the open-source community ships something better—like the OpenAI’s Agent SDK or MCP connectors—we just plug it in and delete our code.
We have shifted to using OpenAI’s Agent SDK and MCP tooling to handle agent loops, tool calling, guardrails and integrations, and we’ve been able to move on with less maintenance for faster innovation.
We’re not trying to win by maintaining infrastructure; we’re trying to win by delivering value through customer outcomes.
Model ML uses OpenAI’s API platform, including GPT‑4.1, OpenAI o3, and the Agents SDK, to power its agents, automations, and internal tools.
Read our blog with Open AI here:
https://openai.com/index/model-ml-chaz-englander/
OpenAI's Executive Function series features perspectives from leaders on the frontier of AI adoption.
A conversation with Chaz Englander, CEO & Co-founder of Model ML.
Model ML is building AI infrastructure transforming how leading financial services firms operate. Model ML’s platform features purpose-built agents and an application that automates end-to-end workflows as well as bespoke research and analysis.
OpenAI recently spoke with CEO and co-founder Chaz Englander about how financial firms are evolving, and how recent AI advances are automating and streamlining their operations.
What was our first meaningful encounter with AI, and how did it influence the creation of Model ML?
After selling our last company, my brother and I realized we didn’t like investing but became obsessed with automating the investment process through GPT‑powered function calling.
We were a six-person family office, but with these GPT‑3.5-powered LLMs, it felt like we had the leverage of a 60-person team.
We built a prototype of Model ML for ourselves and didn’t plan to commercialize it. But once we saw the insight gains and efficiency from automating research workflows, we knew we were onto something.
What are we seeing change inside financial services firms?
There are tasks that historically used to take days, weeks, or even months, and some of those now can be done in minutes or hours. For example, preparing quarterly earnings summaries used to take hours. Now, agents handle this entire process: they pull the data, format the slides, and publish the Powerpoint to SharePoint, all without human intervention. I think that’s going to be the biggest shift we see this year… that you’re going to come in in the morning and your work will already be there.
“I think that’s going to be the biggest shift we see this year… that you’re going to come in in the morning and your work will already be there.”
That’s forcing a rethink of where humans add value, and how companies are going to need to remap where teams are going to be impactful both today and in future.
We’re seeing firms shift people into higher-value, judgment-based roles. The leaders of the companies that we work with, at least in our view, are the ones rethinking the architecture of the whole organization in an AI-native way. It’s incredibly difficult, which is why we find ourselves acting as consultants early on, helping them determine where AI is most applicable today and also futureproofing to where we think it’ll be most impactful in 12 months’ time.
“The leaders of the companies that we work with… are the ones rethinking the architecture of the whole organization in an AI-native way.”
We are seeing the people within financial firms are more impactful now, not less. With automation taking on the grunt work, people can focus on relationships and strategic thinking. The firms that win will be the ones who rethink their entire operating structure to take advantage of that shift.
How does Model ML stand out compared to general-purpose AI tools, and how are new model capabilities benefiting our customers?
In finance, accuracy, compliance, and workflow-fit aren’t optional–they’re table-stakes. That specificity is where general-purpose tools fall short. Model ML was purpose-built for financial services from day one at two critical levels.
First, at the agent layer, we’ve built and fine-tuned systems specifically to parse and interact with the kinds of data financial professionals use daily, structured and unstructured, across tools like Sharepoint, and common data sets like Capital IQ, FactSet, and Crunchbase, which can be hundreds of tables and 20 terabytes. 12 months ago, it was near impossible to build an agent on top of those datasets. These aren’t just models that answer questions; they are contextually aware, understand schemas, write code, and retrieve information from terabytes of complex data.
Second, there’s the application layer: the interface through which users interact with agents, specifically designed for finance. It provides firms with the tools to build agents that automate end-to-end workflows and enable previously unattainable analysis. In terms of use cases, we’re seeing dozens of new use cases daily and we’re at thousands of use cases, a lot of which customers get right out of the box when they sign up.
We’ve seen significant step-changes with every new model release that we’ve channeled into immediate benefits for our customers. The advances in things like reasoning and coding ability have sent areas of our product stratospheric. Most recently with the release of OpenAI o3‑pro, o3, o4-mini, and GPT‑4.1 models, these new models have brought dramatic improvements in reasoning, multimodal capabilities, instruction following and tool integration. With larger context windows and more advanced reasoning capabilities, we’re now able to unlock end-to-end workflows. Now, users can chain together data gathering, analysis, and presentation creation tasks, producing fully formatted outputs completely autonomously.
“The advances in things like reasoning and coding ability have sent areas of our product stratospheric.”
Looking ahead 12 months, what do we think will change the most?
I think the most profound shift ahead is the rise of end-to-end workflow automation, where your systems act as control towers overseeing an army of digital workers. As these agents take on more complex, multi-step tasks across your entire digital universe, even the UI and the way we interact with hardware will start to change. This is likely a step beyond the next 12 months, but it’s where we’re heading.
What’s coming next is the emergence of truly autonomous agents that you can build in our product. Our agents are able to execute sophisticated workflows that gather, analyze, and present data from your CRM, emails, files, external data vendors, meeting transcripts and more. These agents won’t just wait for instructions; they’ll anticipate what needs to be done, whether it’s cyclical (daily, weekly, monthly, quarterly, annually) or triggered by real-world events—just as you’d ask a team member after a meeting or in response to an email.
The real shift is that these workflows will run end-to-end, automatically, with deep reasoning and orchestration across all your systems. The outputs could be as substantial as a 100-page PowerPoint, built entirely by machine—faster, more consistent, and available 24/7.
This is the future: autonomous digital teams running the workflows that drive your business—better, faster, and always on.
How do we keep our team agile as AI evolves so quickly?
Our belief is that AI-native companies will look structurally different. Fewer layers, faster cycles, tighter feedback loops. We’ve embraced a flat structure. Arnie [my cofounder] and I each have double-digit direct reports. That might sound wild, but AI makes it manageable. All one-on-ones are AI-assisted. Notes, to-dos, context, it’s all streamlined. It lets us move faster and stay close to the product. We think that’s how modern companies will run: more like control towers than siloed hierarchies.
Part of being agile is betting on the ecosystem and foundational models getting better. The real key, and maybe this is part of the founder mindset and engineering organizations, is not being emotional about your code. We used to build everything ourselves: agent abstractions, service connectors, all of it. Now, if OpenAI or the open-source community ships something better—like the OpenAI’s Agent SDK or MCP connectors—we just plug it in and delete our code.
We have shifted to using OpenAI’s Agent SDK and MCP tooling to handle agent loops, tool calling, guardrails and integrations, and we’ve been able to move on with less maintenance for faster innovation.
We’re not trying to win by maintaining infrastructure; we’re trying to win by delivering value through customer outcomes.
Model ML uses OpenAI’s API platform, including GPT‑4.1, OpenAI o3, and the Agents SDK, to power its agents, automations, and internal tools.
OFFICE LOCATIONS
OFFICE LOCATIONS
OFFICE LOCATIONS