Interview
Agentic AI is, without question, emerging as a significant application of generative AI, redefining business potential by adding decision-making capabilities to traditional AI models. In an insightful conversation with Jonah Midanik, COO & general partner at Forum Ventures, we discussed why Gartner has recently spotlighted agentic AI as the top trend for 2025.
Unlike conventional AI, Midanik believes agentic AI takes autonomy a step further by not only generating responses or content but also taking proactive actions. This ability to operate autonomously represents a transformative leap for many applications and use cases.
Table of Contents
- Agentic AI Guidance for CIOs
- The Role & Technical Foundations of AI Agents
- Startup Opportunities in Agentic AI
- How AI Agents Will Redefine Software & Service
- Addressing AI Security, Privacy and Collaboration Challenges
- Gartner’s 2025 Prediction & How CIOs Can Capitalize
- Leading With Agentic AI in the Digital Era
Agentic AI Guidance for CIOs
As agentic AI evolves, there will be clear winners — initially, it will be companies that can seamlessly integrate AI into customer journeys, enhancing engagement and driving conversions. Midanik noted that agentic AI could revolutionize customer acquisition and engagement, potentially reducing the need for human intervention in routine tasks.
However, this shift brings forth critical challenges, especially in areas like security and privacy. The autonomous nature of agentic AI heightens the importance of robust privacy protocols and safeguards to prevent misuse or unintended actions.
For CIOs, the guidance is clear: prioritize human-agent collaboration, ensuring that AI complements rather than replaces human talent. By establishing clear governance frameworks and investing in training, businesses can harness agentic AI’s capabilities while maintaining control. Embracing this trend requires a balance between innovation and risk management, with CIOs playing a crucial role in steering agentic AI’s integration into organizational strategies.
Related Article: AI Agents: How CIOs Can Navigate Risks and Seize Opportunities
The Role & Technical Foundations of AI Agents
As an investor, why does agentic AI represent the killer app for GenAI?
Midanik argued that agentic AI has the potential to become the killer app for generative AI because it represents a completely new paradigm in how we interact with software. Instead of relying on a human-driven clickstream to accomplish tasks, agentic AI allows users to simply command, do X for me, and have the software carry out that task from start to finish. Even complex, multi-step processes can be handled autonomously.
He went on to suggest that “this leap in capability could significantly boost productivity by automating intricate workflows and freeing people from routine tasks. It’s this ability to simplify and supercharge productivity across countless applications makes agentic AI stand out as a breakthrough use case for generative AI.”
What are the key enablers of agentic AI from a tech stack perspective?
For my friends who are CIOs and CDOs, what do they need to assemble to deliver agentic AI? Midanik said, “the tech stack for agentic AI is rapidly solidifying, and it introduces some crucial considerations and changes in software architecture.
“First, a robust data layer is essential — whether through traditional databases or more specialized storage solutions. Next comes the foundation models, such as those from OpenAI, Anthropic and similar providers, which serve as the core engine driving agentic AI capabilities.”
With these foundations, Midanik said, “deployment is a critical element, encompassing how these applications are hosted and managed, whether through a cloud platform or on-premises infrastructure. Then there’s the application layer itself — the software being built. This raises interesting questions around user interface and experience (UI/UX) design, as agentic AI could transform how users interact with technology.”
Next, as another of these articles will cover, Midanik said, “integration is vital. Connecting agentic AI applications to other pieces of software remains essential, though it won’t differ drastically from current software integrations.”
However, one area that’s poised for significant evolution, added Midanik, “is security and governance. Beyond inference and output generation, platforms must ensure accuracy, compliance and fairness. This entails a strong governance layer to address issues like bias management, human oversight, privacy and ethical data usage. Collectively, these layers highlight the emerging toolsets required to support agentic AI, building upon established foundations like data and deployment layers while introducing new complexities.”
Startup Opportunities in Agentic AI
Beyond large companies, is there room for startups to invest in agentic AI?
Not surprisingly, Midanik argued, “there’s ample room for startups in any emerging field, and agentic AI is no different. While many incumbents are already building agentic AI applications atop or alongside their existing platforms, which gives them some inherent advantages, startups bring their own edge to the table. They can pivot and innovate at remarkable speed, unencumbered by legacy architectures and existing infrastructure that often slow down larger players.
“This nimbleness creates space for fast-moving agentic AI startups to gain traction and achieve wide adoption. Overall, the agentic AI space is rife with opportunity, and we expect startups to play a major role in shaping its future.”
What kind of use cases will best be served by startups?
To my personal surprise, Midanik said, “the use cases best suited for startups in the agentic AI space span several key areas.
“First and foremost are security, governance and monitoring — essential elements for launching agentic AI applications safely and responsibly. This is a space where incumbents don’t have an inherent edge. In fact, it might even be better for new, specialized players to lead here rather than the companies deploying the applications themselves. Similarly, privacy and the data layer present substantial opportunities for innovative startups.”
Beyond deployment and application development, Midanik argued, “there are plenty of verticals where incumbents simply aren’t well-positioned to build agentic AI companies. This opens doors for startups to make a real impact.
“On the horizontal front, there’s also major potential for startups to excel in areas like productivity and communication. These are spaces where incumbents often struggle to innovate quickly or meaningfully, giving agile startups a chance to carve out their own success stories. It’s an exciting time, with a lot of room for new players to shape this evolving landscape.”
Related Article: Can Agentic AI Revolutionize CX and EX?
How AI Agents Will Redefine Software & Service
How will AI agents change how sales, marketing and service is done?
A year ago, I got to spend time with the leadership team for Optimizely. I was curious how their market would be impacted by agents.
“Agents are set to revolutionize how sales, marketing and customer service operate”, said Midanik. “For years, we’ve been promised personalization, but what we’ve mostly seen is superficial — think systems that plug in your name or scrape basic details like your company or birthday. It’s a weak form of personalization.
“However, generative AI at the heart of agent applications will change that dynamic, bringing true one-to-one customization. Imagine agents analyzing your LinkedIn profile to craft uniquely tailored messages or accessing all historical data between a company and a customer to deliver contextually aware responses. On the service side, they can leverage entire customer service manuals to provide accurate and relevant assistance. This shift means agentic applications will not only possess full knowledge of a company’s data history but also understand individual customer interactions.”
Given this, I asked Midanik about the impact on jobs. He said, “first, humans will do much less of the routine work involved in sales, marketing and service. Second, the output will be significantly more personalized, moving us closer to genuine, context-rich engagements at scale. It’s a transformation that could redefine customer-facing roles and elevate the customer experience.”
Will agentic AI really transform how we think about software (when that promise has been made in the past)?
Midanik sees this wave of technology as fundamentally different. He claimed that “agentic AI is set to fundamentally reshape our perception of software. Today, we think of software as a place where we perform a series of steps to achieve a result. With agentic AI, that paradigm shifts: users will interact with software to achieve a result, while the software itself handles the steps. This transition marks a profound change in how we engage with computers.”
Webinar
Sep
25
CMS Briefing: A Live Look at What’s Next in AI-Driven Platforms
Learn how leading organizations are using AI‑driven tools to publish faster, personalize smarter and stay secure.
Register
Webinar
Oct
15
Agentic AI Playbook Real-World Customer Service Use Cases You Can Deploy Now
Boost self-service by 30% and slash call volume by 63% with agentic AI.
Register
Webinar
On demand
Ready or Not: How Data-First Organizations Are Unlocking Agentforce Potential
Learn how to cut through the noise, activate Agentforce and build a Salesforce AI strategy that actually delivers.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
Webinar
On demand
From Hype to High-Impact CX Strategies That Actually Scale
Turn buzzworthy AI and outsourcing trends into measurable CX wins with fresh 2025 data.
Watch Now
Webinar
On demand
Insights to Action Rethinking the Contact Center for Real Business Impact
Join our exclusive webinar to hear CX executives share their innovative strategies for transforming service delivery.
Watch Now
Midanik went on to say, “equally transformative is the expected move from keyboards to voice interactions. Agentic applications have the potential to excel with voice commands, allowing users to initiate complex sequences of actions verbally. Instead of sitting at a computer and navigating through clickstreams, we’ll simply speak our intentions, triggering the software to carry out tasks autonomously — even if we’re nowhere near a keyboard.
“These two shifts — AI-driven task execution and voice-based interactions — could represent a sea change in what software means and how we interact with it, setting the stage for a new era of intuitive and seamless digital experiences.”
Addressing AI Security, Privacy and Collaboration Challenges
When it comes to security, privacy and accuracy, which issues are most important for agentic AI?
“Security, privacy, accuracy and governance are critical pillars for agentic AI, and for good reason,” said Midanik.
“On the privacy front, any leakage of data — whether customer data being misused for model training or exposure through other vulnerabilities — poses a huge threat. Data leaks have always carried significant repercussions, from damaging credibility to imposing heavy financial penalties. But in an agentic AI context, the stakes are even higher. The potential for damage escalates with agents handling sensitive data, making leak prevention paramount.
“Security is also evolving alongside this technology. It’s no longer just about traditional safeguards; it’s about determining where these AI applications are allowed to operate, when and how humans should intervene and protecting new attack surfaces. This adds layers of complexity, demanding a rethinking of security protocols tailored for agentic AI’s unique challenges.”
Meanwhile, Midanik suggested, “accuracy becomes a mission-critical priority. Foundation models can produce inaccurate outputs, but the risks multiply when agents interact with each other or perform complex tasks. Inaccurate or harmful results are a real danger if left unchecked.
“Ensuring post-process accuracy, governance and traceability will be vital to keeping agentic AI on track. These areas are already gaining attention and will become essential fields of focus as the technology continues to evolve, offering both immense opportunity and responsibility for the industry.”
In his book, David De Cremer insists GenAI is about human-agent collaboration. Is that true?
Over the past several months, I have been experimenting with ChatGPT in my writing. I have discovered that outlining is critical and changing things in interactive fashion is also important.
For this reason, I was not surprised that Midanik said, “Human-agent collaboration is undoubtedly the future of software. The key question for all of us is figuring out where and when these collaboration points will occur. How far will we allow autonomous software to go before stepping in and taking control? To what extent will interactions be limited to agent-to-agent exchanges? These are uncharted waters, and much remains to be determined.”
Nevertheless, Midanik added, “what’s clear is that agentic software and this emerging agentic future will continue to be steered by humans as we navigate this new wave of technology. Identifying and refining these collaboration points — making them controllable, adaptable and testable — will be a critical part of the learning curve and likely a source of hesitation for some as adoption grows.
“Despite these challenges, there’s no doubt that this evolution will only deepen how humans use and rely on software and computing systems, pushing the boundaries of what’s possible.”
Related Article: Navigating the AI Landscape: A CIO's Guide to Success
Gartner’s 2025 Prediction & How CIOs Can Capitalize
Gartner has named agentic AI the trend of the year. Why have they got it right?
Midanik said that “Gartner’s prediction of 2025 as the year of the agent or the year of agentic AI is spot on. This projection aligns perfectly with the current trajectory of tech innovation and adoption rates.
“Over the past two decades, software adoption has only accelerated, a trend vividly illustrated by OpenAI’s ChatGPT, which reached 100 million users faster than any application in history — within a single year. The pace of change is quickening, and the scale of transformation is intensifying. Everywhere you look in the tech landscape, discussions about this agentic future are taking center stage.
“Gartner’s foresight comes from recognizing the convergence of fast-paced adoption, the explosive rise of generative AI and the emerging agentic AI ecosystem. This is a rare moment where prediction and momentum meet, signaling a truly transformative period ahead.”
What is your advice for CIOs who want to get started with agentic AI?
For CIOs eager to explore agentic AI, Midanik said, “a prudent approach is to start small. Begin with low-risk, well-defined use cases that allow for clear control and monitoring. This hands-on experience will help your team become comfortable with the technology within your enterprise environment. Equally important is a thorough understanding of the privacy, security and accuracy landscape.”
With this accomplished, Midanik said, “CIOs should develop a robust internal framework that addresses:
- Human Oversight: Determine when and how human intervention is necessary
- Privacy: Establish and enforce policies to protect sensitive data
- Security: Implement measures to safeguard systems against threats
- Accuracy: Ensure outputs meet your organization’s standards
By focusing on these areas and starting with manageable projects, you’ll likely uncover significant, transformative opportunities across your enterprise.”
Leading With Agentic AI in the Digital Era
The rise of agentic AI marks a defining moment in the evolution of AI technology, with far-reaching implications for industries aiming to boost efficiency and personalization through autonomous software. As Jonah Midanik insightfully highlighted, this paradigm shift will reshape how businesses approach customer engagement, productivity and operational automation.
However, realizing the full potential of agentic AI will require CIOs and technology leaders to carefully balance innovation with vigilance, prioritizing human-agent collaboration, security and ethical governance. Those who can skillfully navigate these challenges and embrace the transformative possibilities of agentic AI are likely to lead the next wave of digital advancement, reimagining what software can achieve.
