Latest news with #CoFounder
Fox News
16 hours ago
- Business
- Fox News
Michele McManamon on Helping Military Veterans and Spouses
Michele McManamon Chief Executive Officer & Co-Founder Operation New Uniform spoke to Brian Kilmeade about creating Operation New Uniform a nonprofit organization dedicated to empowering transitioning Servicemembers from all branches of the military to find their 'new uniform'–a fulfilling career in the business world. Click here to find out more
Forbes
19 hours ago
- Forbes
The Rise Of Autonomous Cyber Agents
Ronen Cojocaru, Co-CEO and Co-founder, Imperative Inc. getty Artificial intelligence is rapidly evolving from passive tools into autonomous "agentic" systems capable of making decisions and taking actions without direct human input. These AI agents are already proving valuable as co-pilots to human analysts, enhancing threat detection and speeding up incident response. Yet their growing autonomy is a double-edged sword. As these agents gain more power, ensuring they remain secure, transparent and reliable becomes paramount. Early examples of agentic AI in cybersecurity, from automated threat-hunting bots to self-driving network monitors, demonstrate huge potential. However, they also highlight new vulnerabilities. AI agents can, unfortunately, be easily tricked or influenced by bad data, sometimes resorting to biased or incorrect assumptions, and users may place misplaced confidence in their outputs. In short, agentic AI is a force multiplier for cyber defense, but without proper safeguards, it can just as easily multiply cyber risk. Despite the promise, security leaders must grapple with several emerging risks from agentic AI systems. Notably, model drift, malicious manipulation and operational reliability issues are front and center: Model Drift Over time, AI models can become misaligned with reality as their input data changes—a phenomenon known as 'data drift.' This natural degradation in data characteristics means an AI that once performed well might start making errors as its environment evolves. For example, an intrusion detection model trained on last year's network traffic may gradually falter as new apps, devices and attacker techniques appear. Such drift opens up new attack surfaces if not caught and corrected, undermining the model's effectiveness. Recognizing this, recent joint security guidance from the U.S. and allies urges companies to monitor AI performance closely and treat drift as an expected challenge. Agentic AIs are vulnerable to adversarial exploits. Hackers can attempt to manipulate an AI's inputs or training data to distort its behavior. Tactics like data poisoning and feeding incorrect or malicious data into an AI's training pipeline can wreak havoc on its decision making. Imagine an attacker subtly corrupting the data that trains a spam filter or fraud detector—the AI might then start letting threats slip through or flagging the wrong items. Officials worldwide are increasingly fearful of hackers manipulating AI systems, especially those deployed in critical infrastructure. A poisoned or manipulated model not only makes bad choices; it erodes confidence that AI outputs can be trusted at all. Operational Reliability And Trust Like all AI, autonomous agents suffer from issues of hallucination, bias and erratic behavior, which can be amplified by their autonomy. Without proper governance, an AI agent might confidently produce incorrect analyses or take unauthorized steps. These problems aren't just theoretical—early deployments have shown that AI assistants can 'go rogue' or output toxic content if misused. Businesses have learned that an unsupervised agent's mistake can lead to serious harm, reputational damage or compliance violations. Moreover, when AI agents act unpredictably, humans tend to either over-trust them or distrust them entirely—both scenarios are risky. As one expert noted, current AI agents are still 'easily tricked' and prone to biased assumptions, yet people often trust their answers when they shouldn't. Ensuring reliability means building in rigorous testing, guardrails and oversight for AI decisions. In practice, companies are putting 'human in the loop' controls on critical uses and instituting AI red-team exercises to probe for failure modes. The goal is an AI that operates responsibly and transparently, earning trust through consistent and correct performance. Future Outlook: Roadmap For AI-Powered Cybersecurity While today's agentic AI is still maturing, the coming years promise a dramatic expansion of AI's role in cybersecurity. In this phase, organizations move from experimentation to real deployments of agentic AI for security. AI co-pilots become common in security operations centers, handling routine tasks and assisting human analysts. For instance, autonomous AI agents might triage alerts, scour logs for threats or automate responses to basic incidents. These early agentic systems are generally narrow in scope and operate under human supervision, reflecting lessons learned about governance. Shadow AI agents (unsanctioned bots running without oversight) emerge as a concern, prompting companies to institute AI governance programs. Industry experts emphasize the need for visibility into all AI agents in use and strict alignment with security policies to avoid 'rogue' deployments. Notably, businesses begin to treat AI agents much like employees: vetting their 'credentials,' monitoring their activities and granting only least-privilege access. As one analysis put it, AI agents can indeed augment overworked cyber teams, but only if we ensure these agents are deployed in a secure, explainable and reliable manner. Looking a bit further out, 2026 is expected to usher in swarm intelligence and collective defense enabled by networks of AI agents. Rather than working in isolation, multiple AI systems will increasingly communicate, collaborate and even negotiate with each other across networks. Cyber defenses could be handled by fleets of specialized AI agents, with one set watching network traffic, another analyzing user behaviors and others managing endpoint security—all sharing intelligence in real time. This coordinated 'swarm' of AI agents can respond to threats faster than any single system, mimicking a colony of ants or bees that collectively defend their nest. A new challenge will be understanding the emergent behavior of interacting AIs. When dozens of semi-autonomous agents interconnect, unexpected dynamics may arise not unlike complex financial markets or ecosystems. By the late 2020s, the industry anticipates a transition from narrow AI tools to cognitive cybersecurity ecosystems. In practice, this means AI systems with advanced reasoning capabilities are deeply integrated into every facet of cyber defense. For example, cyber defense systems will leverage AI that emulates human-like thinking and learning processes. These cognitive SOCs can ingest vast, diverse data streams, network logs, threat intel feeds, user activity and more to make connections that human analysts might miss. Cybersecurity ecosystems will become adaptive and self-optimizing. AI will not just react to attacks but continuously learn from them, evolving its defenses. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Forbes
4 days ago
- Business
- Forbes
AI-Driven Documentation In Construction: Today And Tomorrow
Shuangling Yin, CTO & Co-Founder, InspectMind AI. Construction crews pour concrete and erect steel faster every year, yet one stubborn line item keeps projects over budget: paperwork. A 2024 McKinsey analysis noted that 'the industry is lacking sufficient capable workers, and economic labor productivity (the economic value added per hour worked) has stagnated for decades globally despite technological advancements and improvements by individual firms.' Most field teams still write notes on paper and retype everything after dark so owners, lenders and regulators can audit progress—a routine that has barely changed in decades. Today, however, the latest AI makes automated reporting feasible; McKinsey noted that 'technologies such as generative AI could fundamentally transform how capital projects are delivered.' Speech recognition, large language models and mature mobile technology can allow workers to press a mic button, describe what they see, snap photos and let software handle the reporting—like a human assistant. When checking a report is quicker than writing it, skilled workers can stay focused on building, not paperwork. Visit a site trailer at 6 p.m., and you'll find a superintendent who has just handled 10 hours of rebar checks, change-order calls and a midday cloudburst—but the day isn't over. The daily log still awaits: head counts, tasks, weather, photos. Twenty minutes of typing often pushes this chore to Friday, when memories blur—or the log never gets done. While accurate documentation is crucial for timely payments and project tracking, field workers excel at building, not writing reports. One superintendent started using our mobile app, which generates reports automatically from photos and voice recordings. According to Zippia, 75.2% of U.S. construction professionals who speak a foreign language speak Spanish. This is where AI can be particularly valuable; workers can speak about their day in Spanish for three minutes, and their professional report in English is ready. The app integrates with Procore, automatically sending reports to the popular project management platform. Construction contractors aren't alone in needing field documentation. Structural engineers must document observations throughout a building's life cycle—during construction, after completion and over the years to ensure safety. California's SB 721 and SB 326 mandate periodic safety checks of balconies, decks and walkways in thousands of multifamily buildings. We worked with a licensed structural engineer based in San Dimas who is qualified to inspect balconies for both regulations. He captured about 1,600 photos and 800 bullet-point notes at a Long Beach condominium in three days. Previously, he would have spent another week renaming images, matching them to units and assembling a 400-page report. Now, AI can generate the report in 10 minutes, and he can review it in an afternoon. After adopting AI-driven report writing, the same engineer can take on more inspection jobs each week—potentially doubling income—and spend more time with family instead of writing reports. Despite AI's clear benefits for documentation, it isn't yet mainstream in the construction industry. Construction professionals traditionally aren't the most tech-savvy. Even with mobile technology available for nearly 20 years, many still write notes on paper. The efficiency gain must be significant enough to drive behavioral change and require an open mind toward technology. The current AI wave might revolutionize traditional industries faster than past technological advances. With AI, interacting with technology by just talking feels natural. One of our customers reported that their 70-year-old employee with a heavy Southern accent found the app easy to use and was impressed by the technology. However, new technology brings risks. Construction projects might contain highly sensitive data—you don't want AI using nuclear plant information for external model training. Additionally, there's a liability factor in inspections, which is why many require licensed professional engineers. With careful development of AI-enabled construction software, reports can be based on inspectors' substance and input, not fabricated information. AI serves as an assistant, not a replacement for engineers' technical expertise. With perhaps 5% oversight from AI, engineers should always review final reports before submission, as they remain ultimately liable. I believe AI-driven approaches will gradually become the mainstream of documentation in construction. As someone said, AI doesn't replace your job, but competitors that use AI might replace you. Players are motivated to adopt tech due to competition. For makers of AI construction software, we must deeply understand the industry and responsibly make this transformation happen. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Forbes
5 days ago
- Business
- Forbes
Agentic AI In Enterprise QA: Powering Intelligent, Autonomous Testing At Scale
Pradeep Govindasamy is the Co-Founder, President and CEO of QualiZeal. We're at the beginning of a new era in quality engineering, one shaped by agentic AI. While generative AI has captured global attention, the real transformation in software testing is only just beginning. I believe we're now entering a phase where AI isn't just assisting people in testing tasks. It's becoming autonomous, goal-driven and capable of acting with intelligence across the lifecycle. At QualiZeal, we're witnessing this shift firsthand. As someone who has spent years in the testing space, I can confidently say that AI is not a far-off future. It's here, being built into our processes today, and it's already beginning to disrupt how we think about quality at scale. Software development and testing are the two most critical pillars in any IT application lifecycle. To get a product into the hands of customers, you first build it, then test it and only then can you ship it. We've seen how tools like GitHub Copilot have revolutionized development. Now, that same level of AI adoption is happening in software testing. This is no small market—it's a $100 billion global industry. And just as smartphones once disrupted legacy devices like BlackBerry, AI is poised to transform testing in a similar way. Every phase of the software testing lifecycle—test case preparation, test design, test data management, performance testing, site reliability engineering—is now being infused with AI to increase efficiency, productivity, and ultimately software quality. Before we talk about agentic AI, we need to understand the evolution. The first step in embracing AI is automating repetitive, rule-based tasks. Once you have robust automation in place, AI capabilities can be layered on top to improve every phase of testing. But agentic AI goes one step further. With standard AI, we build prompts, define logic and teach the models how to behave. With agentic AI, we create systems that learn, adapt and act autonomously. These agents follow instructions and understand intent. They can analyze changes in the system, adjust automation scripts accordingly and execute tests without human intervention. For example, imagine a scenario where a company updates its checkout process, maybe tariffs or payment options change. In the past, a QA team would have to manually identify changes, rewrite test scripts and rerun tests. With agentic AI, the system learns what's changed, modifies the scripts, self-heals when errors occur and continues testing. It even generates a report outlining what it changed and why. This self-healing, self-optimizing capability sets agentic AI apart from traditional automation. And it's a game-changer. We're seeing both technical benefits and measurable business outcomes. With agentic AI, the cost of quality is decreasing. From my observation, the industry average today is about 18%, but with AI-infused testing, we anticipate a 5% drop, driven by reduced manual effort and increased efficiency. In maintenance alone, we've seen a reduction from 20% of team capacity to less than 5%. Even more importantly, release cycles are accelerating. Time to market (TTM) has gone from quarterly to weekly, and now, with agentic AI and DevOps practices, to daily releases. The entire production throughput is becoming faster and more reliable. And decision-making is more seamless because agentic systems provide full transparency through real-time reporting, eliminating the need to compile data across disparate systems. Organizations looking to lead in this space must prepare now. I always say this moment is not just about catching up—it's about disrupting yourself before you get disrupted. Companies that wait too long will miss the opportunity to lead. Those who invest now will be in a position to capture market share and build the next generation of testing capabilities. This preparation requires both a top-down and bottom-up approach. Leadership must allocate budgets, not just wait for client-driven funding, and teams must be empowered to get trained, certified, and exposed to different AI models. AI isn't just a CIO or CTO conversation anymore. It's happening at the board level, and for good reason: this is the foundation for long-term competitiveness. I recommend organizations push their teams to reach at least level three in AI readiness: basic execution. Core functions like engineering and QA need to go further, while ancillary teams like finance and marketing should also gain exposure. Of course, with great power comes responsibility. We need to ensure agentic systems operate ethically, transparently and securely. Especially in regulated industries like healthcare, insurance or banking, any AI-driven decision, no matter how small, can have massive consequences. That's why testing the AI itself is just as important as using AI for testing. There's a growing demand for AI-specific test engineers who can validate agentic systems through high-end exploratory techniques. Traditional testing models like equivalence partitioning or boundary analysis must now be complemented with new approaches tailored to AI behavior. In the near future, eight to 10 new job roles will emerge specifically to test and validate agentic AI systems. These won't be optional. They'll be mission-critical. We estimate that full-scale AI maturity across the testing lifecycle will arrive around 2027. Between now and then, we're in the planning and education phase, training models, customizing LLMs and building the necessary infrastructure. Implementation will accelerate in 2026, and by mid-2027, I expect the majority of enterprise QA environments to be agentic by design. This is a once-in-a-generation opportunity for testers, developers and technology leaders. Gen Z professionals, especially those raised in a digital-native world, will have an edge. They can adopt these tools faster, and many will find themselves building careers in entirely new domains. We're not just building testing systems anymore. We're building trusting systems. Platforms that learn, adapt and support business continuity without human babysitting. That's the future of QA. That's where agentic AI takes us. And the companies that embrace it today? They'll be the ones defining quality tomorrow. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Forbes
02-06-2025
- Business
- Forbes
Multifamily Real Estate: How AI Is Powering Smarter Investments And Asset Management
Matias Recchia is Co-Founder and CEO of Keyway, the AI- powered real estate investment manager. getty If you're not using AI to power your multifamily real estate investments and operations in 2025, it's not too late to start. From market opportunity identification to tenant sourcing and retention to optimizing rental pricing, I've found that AI can provide a competitive advantage for real estate owners and operators through improved decision-making and meaningful investment returns. AI is not only streamlining operations but also fundamentally changing the way real estate stakeholders acquire, manage and scale real estate portfolios. Based on my own experiences as a real estate investment manager, here are three ways you can use AI to empower your team: AI can provide you with access to real-time, actionable data. For example, you can use it to identify a neighborhood with high growth potential early—based on income trends, job prospects, school quality and supply-demand characteristics—providing investors with an early look at where to allocate capital. AI can also analyze large datasets regarding demographic trends, economic indicators and competitor developments to provide a holistic view to investors. For example, you can use AI tools to ascertain in which neighborhoods rental demand is growing, which amenities will win more tenants in a particular city, and which neighborhoods are undervalued relative to those in comparable rental assets. Optimizing rental pricing is about maximizing occupancy and profitability for landlords in any market environment while simultaneously providing fairness and transparency to tenants. AI and machine learning can play an integral role in improving comparable property data, or comps, for commercial real estate. Traditionally, real estate teams have relied on manual processes to manage comps based on old data. But by leveraging AI and machine learning, you can reduce human error, manual labor and time. Furthermore, by relying on public data rather than landlord-reported private data, you can ensure that your real estate stakeholders have access to unbiased data with full transparency. Rent optimization also means landlords can price their units based on real-time data, reducing the likelihood of underpricing or overpricing. Real estate teams need multiple data points beyond gross rent to inform their rental pricing decisions. For example, rent trends, fee and concession trends, neighborhood dynamics, competitor pricing and economic indicators represent several components that comps platforms should analyze. You can also train your AI platform to incorporate cap rates and long-term appreciation potential to garner a more accurate picture of investment returns. Tenant screening is another component that can optimize both rent and occupancy. Consider using AI and machine learning to review applications, employment history, credit and other factors to predict whether a tenant will be reliable and financially responsible. With more reliable tenants, landlords can reduce turnover, which translates to higher recurring cash flow and occupancy. If you're a property manager, consider incorporating AI-powered energy management systems and offering automated tenant communications. You can also leverage AI to drive predictive maintenance. For example, AI can monitor HVAC equipment and elevators to determine when the next maintenance or repair is required so managers can avert major delays of services. By analyzing usage, useful life and prior maintenance work, you can proactively avoid tenant inconvenience, reduce complaints and save operational costs. Asset managers can leverage AI to deploy capital more systematically. For example, by sharing financial goals, investment returns and time horizons with your AI platform, you can better determine when to refinance assets, which assets should be prepped for sale and what's the optimal way to reinvest proceeds. Capital deployment optimization can be done in Excel, but I've found that AI reduces reliance on individual decision-making and takes a more objective approach. For example, my company uses an AI-powered program to manage our document workflow. Keeping track of leases, ensuring consistent lease terms and identifying inconsistencies in legal language are critical must-haves for any real estate team, and using AI tools instead of relying solely on your legal team can help reduce human error. When you have a clear lease management strategy, you can reduce legal costs and maintain consistency across your portfolio. I believe the future of multifamily real estate will be marked by dynamic pricing, predictive maintenance and AI-powered energy optimization. AI can give property owners and asset managers a real-time full view of their portfolio, which is important in a real estate market that is unpredictable and subject to economic and market fluctuations. By reducing manual errors, removing bias and creating uniformity in decision-making, real estate teams can have greater control over underwriting, operational roadblocks and tenant satisfaction, leading to an operationally sound multifamily market powered by technology, efficiency and data. Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
