2025 AI for a Busy Executive: All You Need to Know

You’re too busy to fully dive into AI. I’m too busy running Techery, where we build AI solutions for our clients every day. So, I wrote this article.

After reading it, you’ll be fully caught up on everything you need to know about the current and near-future applications and expansion of AI in enterprise and beyond. Let’s go!

The Big Shift

Artificial Intelligence is leaving its cozy sandbox of chatbots and copilots. It’s venturing into a new world of autonomous agents—bots that don’t just answer questions, but actually get things done. In 2023, a fresh idea took hold: AI agents, powered by large language models (LLMs), working on their own, reasoning about tasks, and executing them without a human constantly whispering in their ear. By May 2024, Sam Altman, CEO of OpenAI, was calling these agents the “killer feature” of AI. Everyone took notice. Over 50 startups have launched since 2022, each focused on agents and the infrastructure around them.

This isn’t just idle talk. Experts are calling 2024 the year of the capable AI agent. In 2025, we expect swarms of them, interacting with one another to reach complex goals. These systems promise not just to respond, but to perform—driving a fundamental shift beyond the old notion of a Q&A machine and ushering in a future brimming with action-oriented digital workers.

Gartner’s forecasts for next year’s AI trends point directly at this: by 2025, fully autonomous AI agents will become a top strategic technology trend. They’ll run tasks without handholding.

By 2028, at least 15% of our daily work decisions could be entrusted to such agents. It’s not just about answering questions—it’s about taking action. Vice President at Gartner, Gene Alvarez, sees them not merely boosting productivity, but saving time. They’ll start with routine tasks, then scale up to monitoring, decision-making, even training new hires. Alvarez calls it both “cool and scary”—a revolution that might displace some professionals, while freeing others to work in roles better suited to their talents.

Gartner also predicts that by 2028, we’ll have 15 billion connected products behaving like consumers, making purchases on our behalf. NVIDIA’s CEO Jensen Huang envisions AI agents soon crafting multi-step plans to solve intricate problems, privileged with access to crucial data, working with other AI systems and humans alike. In 10 to 20 years, as we push deeper into Web3, these agents might run the show, with humans more like helpful sidekicks than central players.

Look closely, and you’ll see a curious pattern. Some projects let agents collaborate with people—or even hire them.

Consider Payman AI, a startup that gives AI agents a budget and access to a freelancer marketplace, paying humans to do the tasks the bots can’t handle. It might be a creative request, an on-site visit, a data check requiring a human eye, or influencer marketing. The AI agent posts the gig, a human picks it up, and upon task completion, the agent pays out. Payman AI’s approach spans everything from product management to software development to HR. Similar approaches show up in hiring processes, too: Pymetrics uses AI and neuroscience-based games instead of old-school interviews to assess candidates.

Follow the money

Money is rushing in. MarketsandMarkets says the AI agent market, $5.1 billion in 2024, could reach $47.1 billion by 2030, a staggering CAGR of 44.8%. CB Insights has taken a close look at this gold rush.

The theme: we’re shifting from chatbots and copilots to self-sufficient agents that tackle complex assignments on their own. These analysts detail the current capabilities, market landscape, top companies, investment trends, and key use cases—customer service, sales, software development, and more. The big question they’re asking: how will this agent revolution transform the workforce and business processes, potentially replacing employees and demanding new infrastructure solutions?

Real Capabilities

Let’s pull back the curtain on these agents. AI agents represent a new stage after chatbots and digital assistants. Unlike their simpler cousins, these agents can handle multifaceted tasks, engage with various programs and services, and make informed decisions with data at their disposal. They’re powered by LLMs, enabling them to read, understand, and generate text. They connect to the outside world through APIs. They remember past interactions, building context over time.

The possibilities are already here: info retrieval, booking travel, scheduling meetings, writing code, performing market research. In healthcare, finance, education, manufacturing, and even the arts, these agents find their niches. Yet there are challenges: reliability wavers when agents juggle multiple tools and websites.

Complex reasoning and strategic planning are still hard. Integrations through APIs remain tricky. One standout example: Devin, a cutting-edge developer-focused AI agent, handles only 14% of real tasks without human help. Devices like Rabbit R1 and Humane Ai Pin struggle with even simple jobs like hailing a taxi or playing music. Trust is fragile, a barrier to mainstream adoption.

Still, we’re moving forward. Language models like GPT-4 have become more powerful and context-aware, enabling agents to write articles, produce code, and serve as chat-based assistants. The next wave of models promises even more versatility, pulling from interdisciplinary data sources to solve harder problems.

Meanwhile, multimodal systems let agents process text, images, audio, and video all at once. DALL-E can generate images from text. Sore can generate realistic videos. New Gemini can alter your drawings.These capabilities open up vast frontiers in marketing, design, and media.

Agents are becoming adaptive and personal. They learn user preferences, offer recommendations, and adjust their behavior in real time. This personal touch matters a lot in consumer-focused industries like retail and online services. As we advance, ethics and safety loom large. Developers must minimize bias and misuse. We’ll likely see tighter regulations to ensure transparency and accountability.

The journey toward true autonomy also leads us to robotics and IoT. Autonomous supply chains, self-driving delivery drones, and AI-run logistics hubs are emerging. Eventually, these agents might manage entire city infrastructures or global supply networks. For now, horizontal applications—universal business processes cutting across sectors—dominate. Soon, vertical applications will catch up, customizing agents for finance, healthcare, and more.

The future of human jobs

Does any of this actually help? Klarna, the “buy now, pay later” giant, revealed that after one month of training, their AI agent can do the job of 700 support staff, saving $40 million annually. Sierra, a startup focused on AI customer support agents, raised $110 million. Their clients say the bots match human performance and slash costs. BP, a global energy firm, cut its need for developers by 70% thanks to AI.

Companies that started with older AI for customer service—Ada, Forethought, Intercom—now pivot to autonomous agents. Cognigy, a contact-center automation outfit, raised $100 million in June 2024 to boost its agent capabilities. In the near future, expect voice-based solutions to bloom. Fixie, originally chat-based, now focuses on real-time voice interaction.

As these agents take over tasks once done by human support reps, large swaths of the labor market may shift. According to the U.S. Bureau of Labor Statistics, about 3 million Americans worked as support agents in 2022. AI is coming for that workload, not necessarily to eliminate all jobs, but to reassign human effort to more uniquely human challenges.

#1 Investment

On the investment front, Adept ($415 million), Imbue ($232 million), and Cognition AI ($196 million) lead the pack. Amazon scooped up Adept’s team and licensed its tech in June 2024, a bold statement from Big Tech.

VCs like Y Combinator and Abstract Ventures back multiple agent startups. Corporate investors—Citi Ventures, New York Life Ventures, Workday Ventures, Atlassian Ventures—are in the game too. Citi Ventures and New York Life invested in Norm Ai, a compliance-focused player. Workday Ventures and Atlassian Ventures put money into Adept. Microsoft, Google, and Amazon are busy building their own agent fleets.

In May 2024, Microsoft expanded its Copilot tools, framing them as “team members” and unveiling new agent-building features. DeepMind introduced Project Astra, a multimodal assistant that converses via video, images, speech, and text.

Alphabet CEO Sundar Pichai called it “a glimpse of the future.” Amazon licenses Adept’s technologies for its AGI group, accelerating their digital agent roadmap. Google Ventures invested in Hebbia and Cognosys, while the Amazon Alexa Fund poured money into MultiOn and backed Imbue’s B round.

Tool-building startups form another crucial layer. LangChain’s open-source framework helps craft LLM-based applications, including agents. Zep AI integrates with LangChain to give agents long-term memory. Emergence AI, spun out of Merlyn Mind and funded with about $100 million, builds an “orchestrator” agent that routes tasks to the best LLM or agent available.

Then there’s Anon, created to solve the authentication problem. AI agents navigating the web need secure ways to log in. Anon’s SDK allows agents to authenticate on behalf of users—a crucial piece for reliable agent performance.

Why invest so heavily?

Efficiency, productivity, cost savings. Agents automate the routine and tedious, freeing people to be more creative. They run 24/7. They deliver personalized, round-the-clock customer service. They handle sensitive back-office tasks at scale, trimming costs. Such claims aren’t empty hype; sales development representatives (SDR) powered by AI represent one of the hottest segments.

Clay, with $62 million raised, launched Claygent for data enrichment and sales leads. Qualified, backed by $163 million, introduced Piper, an AI-driven SDR tool. 11xAI reached $2 million ARR by March 2024 with Alice, its AI SDR that charges per completed task—like researching accounts or scheduling meetings.

Software engineering

Programmers, too, face disruption. Engineer-assistant bots backed by LLMs from GitHub (Copilot), Meta (Code Llama), and Amazon (CodeWhisperer) get smarter every day. Cognition and Magic aim beyond snippet suggestions, hoping to build agents that handle entire engineering projects.

Though quality remains an issue—customers complain about lack of originality, data privacy concerns, and “fluff” in the code—the potential is huge. The U.S. employs 1.8 million software engineers, each costing around $130k a year, totaling over $230 billion. Given these stakes, the push for coding agents won’t slow down.

And yes, we at Techery came to this realization earlier this year as well. That’s when we decided to invest in developing what we call Pipe/ines—QA automation and software engineering custom AI agent platforms. Based on our tests with real projects, these platforms save enterprise IT teams at least 30% of their time.

Cybersecurity

Cybersecurity is another frontier. Security operations rely on specialists to investigate breaches, write reports, and patch vulnerabilities. Nullify wants an “AI security engineer.” Dropzone AI raised $16.85 million to build AI SOC analysts that resolve threats in minutes instead of hours. Autonomous threat hunters could become standard for large organizations.

Corporate workflow

General corporate workflow automation is a reliable bet, too. Some startups promise universal AI employees, from research to sales and HR. Ema, backed by Accel and featured in CB Insights’ AI 100 of 2024, pitches itself as a “one-stop AI worker.” TrueLayer and Moneyview are among its clients. Respell helps enterprises weave together workflows from models like OpenAI, Cohere, Anthropic, and others. Zeta Labs, Reworkd, and Lutra AI focus on research, data extraction, and knowledge management tasks.

More niche solutions loom on the horizon. As with other areas of generative AI, specialization is inevitable. For financial services, the focus lands on compliance and investment research. Norm Ai’s compliance agents are backed by Citi Ventures, New York Life Ventures, and TIAA. Parcha AI, founded by ex-Coinbase and Brex staff, zeros in on fintech compliance. Hebbia, fresh with $130 million in a Series B led by a16z, builds agents for financial services, enabling them to synthesize information from documents into spreadsheets and Q&A dashboards.

Insurance

Insurance might see AI agents underwriting policies, processing claims, and doing the grunt work that slows the industry down today. Roots Automation, which developed an LLM trained on unstructured insurance data, raised funds from Erie Strategic Ventures, hinting at a future where claims and compliance become the domain of AI agents.

Healthcare and Manufacturing

Healthcare’s autonomous agents remain an aspiration more than a reality right now, but don’t bet against them. Already some general-purpose agent startups hint at healthcare use cases like AI pharmacy assistants. In manufacturing and industry, Composabl launched an autonomous agent in May 2024 that can tweak and optimize industrial equipment in real-time. Partners like Rockwell Automation and RoviSys confirm its viability. Long term, as humanoid robots advance, they might carry agents inside, making dynamic decisions in unpredictable environments.

Gaming

Gaming provides a promising sandbox. Many companies build AI-driven NPCs (non-player characters) that feel more alive. In March 2024, DeepMind published research on a “general-purpose AI agent for virtual 3D environments,” navigating various digital worlds and tasks. Altera, a research lab backed by $9 million in seed funding, works on “digital beings” that can co-play Minecraft with humans. Games are perfect testing grounds for agents: rich environments, real-time changes, flexible goals. Success here could spill into more mainstream applications.

Risks

Risks and challenges remain. The competition is fierce. As more players crowd the market, consolidation could follow. Regulations are on the way. Governments may pass laws that reshape how we build and deploy these agents. Ethics and accountability will matter more than ever.

Conclusions

Despite all this, the future looks bright. In 5 to 10 years, expect agents woven into everyday life. They’ll tie into IoT devices so that smart homes adapt seamlessly to residents’ needs. They might evolve into virtual doctors diagnosing patients with full context. They’ll partner with or supervise entire teams.

The investment surge proves that belief in these technologies runs deep. Companies that conquer the trust gap, streamline integration, and ensure data security will claim a giant competitive edge.

These AI agents, once raw prototypes, are marching quickly toward the heart of our day-to-day reality, reshaping work, business, and society itself.

Do you need to be all-in? Absolutely YES.

Happy Holidays!