Key takeaways
Yesterday's AI was about data structures, algorithms, and machine learning. Today, the focus is on generative AI. And the AI of tomorrow is the world of agents.
——Anna Barber, M13 Partner
We need to get to a world where you can say to an AI agent ‘I’m hungry’ instead of ‘order me a pizza.’
——Sami Shalabi, Maven AGI Co-founder and CTO
An agent needs to have a component of action and execution power and a feedback loop. The action is different across every context, but the common elements are reasoning plus action and some amount of learning.
——Nami Baral, Niural Co-founder and CEO
Acceptance of AI agents is about trust, and it’s about stakes. But if you have a way to build a lot of trust, you can handle higher-stakes processes.
——John Nay, Norm Ai Founder and CEO
What is an AI agent?
While generative AI has dominated conversations in the past year, the real change to come is autonomous AI agents that not only generate information but orchestrate action. At our recent Future Perfect conference, M13 partner Anna Barber and three M13 founders—Norm Ai’s John Nay, Niural’s Nami Baral, and Maven AGI’s Sami Shalabi—discussed AI agents and the impacts on their respective industries.
Compliance: Norm Ai, which built the first regulatory AI agent platform to help companies stay in compliance, thinks of AI agents across three levels.
- Level one is a system that can autonomously assess if a document or action is out of compliance, flag it, and then pass a full legal review to a human.
- Level two agents take that assessment, then autonomously suggest what actions to take to become compliant. (This is where Norm Ai is at today.)
- Level three, the future state, will be systems that don’t need to hand documents off to humans at all in order to implement changes needed to become compliant.
Human resources management: Today, an AI agent may be able to respond to a request like, “Help me hire this person in the Netherlands.” The agent can create an employment contract, figure out whether this person needs a visa and the requirements to get that visa implemented, get all of the documents in order, make sure that all of the payment methods are aligned, handle taxes, and so on. “That kind of AI agent has a limited scope, where it’s trained on certain kinds of high-precision actions,” says Niural co-founder and CEO Nami Baral.
Tomorrow, an AI agent could respond to a request like, “Here’s our headcount plan, now execute that.” “Then whether you hire in the Netherlands, China, or Kentucky doesn’t really matter,” says Nami. “That’s the level of autonomy that agents should be able to take. Obviously, there are different layers of autonomous action that can happen with HR because HR is so sensitive—but this could create a whole other level of efficiency for the company.”
Customer service (and pizza): Maven AGI co-founder and CTO Sami Shalabi explained his take on agents as he innovates in customer service.
“It's about ultimately giving AI outcomes,” says Sami. “For a non-business use case: Today you can tell an agent, ‘order me a pizza,’ and a pizza arrives. But at the next level, you tell an agent, ‘I’m hungry,’ and it knows from your overall history that you like pizza, and so it orders you one. We’re looking to create that in the business setting. So if you’re a marketing tech company, you can say, ‘I want 100 leads.’ There's a lot of potential value in that, because it's fundamentally 100x-ing human productivity.”
In short, the AI agent technology of tomorrow is about being able to take actions and putting those together in a way where humans aren't in the loop.
What are the benefits? And how does it impact us humans?
John highlights how AI agents can transform regulatory compliance from a burdensome task to a more efficient process. By fully automating compliance tasks that are neglected due to complexity and time-consuming manual compliance, AI agents can help organizations become more compliant than ever before, ensuring higher adherence to regulations.
Sami describes customer onboarding that brings human experience into deciding where to use AI agents in a way that results in ROI:
“It’s a journey to get there on the human side—s opposed to a capability. We typically start with the human in the loop, and when we present that as part of all of our experience, people are able to see the value. Then this unlocks the next phase, and the next phase. When it's not performing, we tell you why. We're hitting about 93% efficiency where our AI agents are able to answer questions without human intervention.”
Where is the line?
As Anna points out, the future of AI agents won’t only be about technical innovation but also a willingness to accept agents that are truly autonomous on an emotional level. “On a societal and cultural level, where are we ready—or not ready—to accept true agents?” she asks. “For example, right now we are more accepting of human error when it comes to driving than of self-driving cars making errors. But people are getting comfortable faster than I thought.”
High precision and rigorous testing are essential for AI agents, especially in sensitive areas like payroll and compliance. Nami points out that Niural does not deploy AI agents until they achieve over 90% accuracy, and for financial transactions, Niural requires 99% accuracy. This high standard ensures that the AI agents can be trusted to handle critical tasks without errors, which is crucial for gaining client confidence and ensuring compliance with various regulations.
“Acceptance of AI agents is about trust, and it’s about stakes,” says John. “I don’t necessarily think stakes are the blocker. If you have a way to build a lot of trust, you can handle higher-stakes processes.”
Get in touch
If you are building in AI and want to talk, please reach out to our investing team: Anna Barber at anna@m13.co and Morgan Blumberg at morgan@m13.co.
Key takeaways
Yesterday's AI was about data structures, algorithms, and machine learning. Today, the focus is on generative AI. And the AI of tomorrow is the world of agents.
——Anna Barber, M13 Partner
We need to get to a world where you can say to an AI agent ‘I’m hungry’ instead of ‘order me a pizza.’
——Sami Shalabi, Maven AGI Co-founder and CTO
An agent needs to have a component of action and execution power and a feedback loop. The action is different across every context, but the common elements are reasoning plus action and some amount of learning.
——Nami Baral, Niural Co-founder and CEO
Acceptance of AI agents is about trust, and it’s about stakes. But if you have a way to build a lot of trust, you can handle higher-stakes processes.
——John Nay, Norm Ai Founder and CEO
What is an AI agent?
While generative AI has dominated conversations in the past year, the real change to come is autonomous AI agents that not only generate information but orchestrate action. At our recent Future Perfect conference, M13 partner Anna Barber and three M13 founders—Norm Ai’s John Nay, Niural’s Nami Baral, and Maven AGI’s Sami Shalabi—discussed AI agents and the impacts on their respective industries.
Compliance: Norm Ai, which built the first regulatory AI agent platform to help companies stay in compliance, thinks of AI agents across three levels.
- Level one is a system that can autonomously assess if a document or action is out of compliance, flag it, and then pass a full legal review to a human.
- Level two agents take that assessment, then autonomously suggest what actions to take to become compliant. (This is where Norm Ai is at today.)
- Level three, the future state, will be systems that don’t need to hand documents off to humans at all in order to implement changes needed to become compliant.
Human resources management: Today, an AI agent may be able to respond to a request like, “Help me hire this person in the Netherlands.” The agent can create an employment contract, figure out whether this person needs a visa and the requirements to get that visa implemented, get all of the documents in order, make sure that all of the payment methods are aligned, handle taxes, and so on. “That kind of AI agent has a limited scope, where it’s trained on certain kinds of high-precision actions,” says Niural co-founder and CEO Nami Baral.
Tomorrow, an AI agent could respond to a request like, “Here’s our headcount plan, now execute that.” “Then whether you hire in the Netherlands, China, or Kentucky doesn’t really matter,” says Nami. “That’s the level of autonomy that agents should be able to take. Obviously, there are different layers of autonomous action that can happen with HR because HR is so sensitive—but this could create a whole other level of efficiency for the company.”
Customer service (and pizza): Maven AGI co-founder and CTO Sami Shalabi explained his take on agents as he innovates in customer service.
“It's about ultimately giving AI outcomes,” says Sami. “For a non-business use case: Today you can tell an agent, ‘order me a pizza,’ and a pizza arrives. But at the next level, you tell an agent, ‘I’m hungry,’ and it knows from your overall history that you like pizza, and so it orders you one. We’re looking to create that in the business setting. So if you’re a marketing tech company, you can say, ‘I want 100 leads.’ There's a lot of potential value in that, because it's fundamentally 100x-ing human productivity.”
In short, the AI agent technology of tomorrow is about being able to take actions and putting those together in a way where humans aren't in the loop.
What are the benefits? And how does it impact us humans?
John highlights how AI agents can transform regulatory compliance from a burdensome task to a more efficient process. By fully automating compliance tasks that are neglected due to complexity and time-consuming manual compliance, AI agents can help organizations become more compliant than ever before, ensuring higher adherence to regulations.
Sami describes customer onboarding that brings human experience into deciding where to use AI agents in a way that results in ROI:
“It’s a journey to get there on the human side—s opposed to a capability. We typically start with the human in the loop, and when we present that as part of all of our experience, people are able to see the value. Then this unlocks the next phase, and the next phase. When it's not performing, we tell you why. We're hitting about 93% efficiency where our AI agents are able to answer questions without human intervention.”
Where is the line?
As Anna points out, the future of AI agents won’t only be about technical innovation but also a willingness to accept agents that are truly autonomous on an emotional level. “On a societal and cultural level, where are we ready—or not ready—to accept true agents?” she asks. “For example, right now we are more accepting of human error when it comes to driving than of self-driving cars making errors. But people are getting comfortable faster than I thought.”
High precision and rigorous testing are essential for AI agents, especially in sensitive areas like payroll and compliance. Nami points out that Niural does not deploy AI agents until they achieve over 90% accuracy, and for financial transactions, Niural requires 99% accuracy. This high standard ensures that the AI agents can be trusted to handle critical tasks without errors, which is crucial for gaining client confidence and ensuring compliance with various regulations.
“Acceptance of AI agents is about trust, and it’s about stakes,” says John. “I don’t necessarily think stakes are the blocker. If you have a way to build a lot of trust, you can handle higher-stakes processes.”
Get in touch
If you are building in AI and want to talk, please reach out to our investing team: Anna Barber at anna@m13.co and Morgan Blumberg at morgan@m13.co.
Read more
%201.webp)
.webp)
.png)
The views expressed here are those of the individual M13 personnel quoted and are not the views of M13 Holdings Company, LLC (“M13”) or its affiliates. This content is for general informational purposes only and does not and is not intended to constitute legal, business, investment, tax or other advice. You should consult your own advisers as to those matters and should not act or refrain from acting on the basis of this content. This content is not directed to any investors or potential investors, is not an offer or solicitation and may not be used or relied upon in connection with any offer or solicitation with respect to any current or future M13 investment partnership. Past performance is not indicative of future results. Unless otherwise noted, this content is intended to be current only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in funds managed by M13, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by M13 is available at m13.co/portfolio.