Venture Bytes #113: Your Next Coworker? An AI Agent

Your Next Coworker? An AI Agent

‍

For decades, businesses have leveraged a blend of human skills and software to tackle complex workflows. While software has transformed certain tasks, human workers still remain the backbone of most business processes. However, AI agents are now emerging as the game-changer that could tip the scale in favor of automation.

While automation has long been a focus in enterprise technology, the natural- language processing breakthroughs of generative  AI  open  new  possibilities. By   combining   natural   language understanding with an ability to plan and act autonomously, these agents can tackle tasks that previously required human intuition and judgment.

NVIDIA CEO Jensen Huang envisions a future where Nvidia operates with “100 million AI assistants,” reflecting the high expectations  surrounding  AI  agents. A Deloitte study has suggested that AI agents outperform GenAI for enterprise automation and productivity. Microsoft’s Satya Nadella aptly describes these tools as “super macros” in a post-ChatGPT world, capable of automating complex workflows that were previously unimaginable.

While the AI agents category is still emerging,  enterprises  from  startups to Fortune 500 companies are already buying and leveraging these systems at scale. A Capgemini survey also highlights that 82% are planning to adopt them in the next three years. Another survey from  Forum  Ventures  showed  that 48% of companies are already using AI agents with another 33% exploring options. Companies like IBM, Coinbase, Microsoft, and Google are leading this shift by embedding AI agents within their platforms to streamline workflows and enhance customer experiences. Gartner forecasts that by 2025, 75% of enterprises will transition from piloting AI initiatives to fully operationalized AI, with AI agents playing a pivotal role.

The potential for AI agents is enormous. Over the next decade, AI could significantly automate 50 major job roles—ranging from  architects  and  construction managers to personal finance advisors, lawyers, insurance sales agents, desk clerks, graphic designers, and software developers. Together, these roles account for 25 million workers, or roughly one-third of the total U.S workforce, according to data from US Bureaeu of Labor Statistics and a16z.

AI agents are proving to be transformative assets in operational efficiency and workforce productivity. By reducing content review cycles by 20-60% (per McKinsey), these agents streamline processes that were previously time-consuming, freeing up valuable human resources for higher-priority tasks. Their potential to automate up to 70% of global work hours highlights the broad scope of impact across sectors. In customer support, AI agents enhance service effectiveness by increasing resolution rates by 14% per hour and reducing handling times by 9%.

Furthermore, deploying AI agents for customer care functions translates into tangible productivity gains of 30-45%, significantly lowering operational costs and maximizing output. For example, California-based DevRev’s AgentOS streamlines workflows in customer care operations and accelerates query resolution by up to 50%.

‍

These capabilities to enhance productivity and responsiveness are fueling a growing market trajectory. The AI agents market, currently valued at $5.1 billion, is projected to skyrocket to $47.1 billion by 2030, growing at a 44.8% CAGR, per MarketsandMarkets.

Major technology players as well as start-ups are quickly adapting their offerings and building new offerings to capture this opportunity. California-based Sierra, which develops AI- powered virtual agents for enterprise customer service and support, is a leader in AI agent space. The company has seen its valuation quadrupled to $4.5 billion within eight months when it raised Series B-II round in October 2024. Sierra’s multi-model AI approach reduces hallucinations and improve accuracy in customer interactions. Adept AI, a Series B start-up, is developing AI models that can interact with computer interfaces. Established in 2022, the company has raised $415 million, reaching a valuation of $1 billion, with backing from General Catalyst, Spark Capital, Microsoft, and NVIDIA.

Cognition AI, another Series B start-up, has built AI software engineer capable of executing complex coding tasks from text prompts. The company doubled its valuation to $2 billion in April 2024 with its $175 million series B round from its Series A round a month earlier. The company also secured a partnership with Microsoft to integrate Devin, its AI software engineer, with Azure cloud services, potentially expanding reach to Microsoft’s customer base

Canada-based Ada  provides AI-driven  customer  service automation solutions, enabling businesses to deliver personalized and efficient support. Founded in 2016, Ada has raised $191 million in total, with a current valuation of $1.2 billion, backed by investors including Spark Capital, Accel, and Bessemer Venture Partners.

A Billion-Dollar Market Opportunity No One is Talking About

‍

The World Economic Forum’s 2024 Global Risks Report identifies disinformation as a top global threat in the next two years. Also, with Gartner projecting that enterprise spending to curb misinformation to exceed $500 billion by 2028—consuming half of marketing and cybersecurity budgets—the scale of this challenge has transformed misinformation into a growth market. For startups in disinformation security, this crisis signals a ripe, high-stakes opportunity.

Disinformation has evolved into a sophisticated challenge, leveraging advanced technologies like artificial intelligence (AI), behavioral science, and social media platforms to spread misleading information rapidly. A 2018 MIT study revealed that falsehoods are 70% more likely to be retweeted on X (then Twitter) than truth, and fake news reaches 1,500 users nearly six times faster than the truth.

A significant portion of the American public — 73% — acknowledged encountering inaccurate news related to the 2024 presidential election, according to Pew Research, highlighting the extent to which misinformation has risen. The scale of this issue can be further underscored by global sentiment: over 85% of people express concern about the impact of online disinformation, and 87% believe it has already negatively affected    country’s politics, as noted in a 2023 United Nations report.

‍

In response, governments worldwide are taking proactive steps to mitigate this threat. The US government, for instance, introduced the ‘Educating Against Misinformation and Disinformation Act’ in 2022, which established a commission focused on media literacy and combating false information.

The high velocity and reach of disinformation underscore its ability to undermine public trust, destabilize markets, and fuel polarization. As AI-generated content becomes increasingly indistinguishable from reality, the risk multiplies as misinformation can now be algorithmically tailored to exploit psychological biases, targeting specific audiences with laser precision.

For enterprises, the stakes are higher than ever. Misinformation can rapidly erode brand credibility, disrupt supply chains, and even influence stock prices. This has opened a new frontier for startups specializing in disinformation security, promising robust demand from sectors where trust and credibility are critical.

In this high-stakes landscape, innovators who can combine technology with deep behavioral insights stand to become indispensable  allies  in  the  battle  against  disinformation. Blackbird.AI, a New York-based startup, shields organizations from narrative attacks fueled by deepfakes, misinformation, and disinformation. With coverage spanning 25+ languages, it analyzes risk across text, images, and memes on the dark web, social media, and news sources. Clients include major players like Bank of America, Amazon, and Pfizer, and recent partnerships with Databricks, DarkOwl, and Two Six Technologies bolster  its data processing, dark web insights, and government reach.

Georgia-based Pindrop focuses on voice security and identity verification, achieving 99.2% accuracy in detecting voice deepfakes, far outpacing human performance, which hovers around 50%. Its voice biometrics and liveness detection tech reach up to 98.3% accuracy. According to a recent Pindrop survey, 67.5% of U.S. consumers are concerned about deepfakes, especially in banking and finance.

New-York startup Reality Defender specializes in deepfake detection for cybersecurity, using AI to identify synthetic content across audio, video, images, and text. It has uncovered political misinformation in Canada and protected a Tier-One bank and an Asian public broadcaster from deepfake audio threats. With dark web demand for high-quality deepfake videos spiking, Reality Defender’s tools address a growing need.

ActiveFence, headquartered in Israel, protects online platforms from various risks and abuses, monitoring over 750 million daily signals across 10 million sources and in 100+ languages. With more than 3 billion users protected, the company has moved beyond proof-of-concept.

‍