Better Call Bots: AI Agents

AI agents are redefining “work”. AI agents have rapidly transitioned from experiments to mainstream in the last 12 months. These tools autonomously analyze, take decisions, and execute tasks, enabling individuals and organizations to hand over entire workflows or tasks to AI. The impact is already tangible. Hebbia, an AI agent for knowledge work, is increasing efficiency across sourcing to diligence and even recommending the next CEOs. Orby AI's agents cut processing time by 70% and labor costs by 60% for a tech giant's contract migration, completing the task 3x faster than manual efforts. The Big Tech bang in AI agents is coming. OpenAI’s Operator, slated for a January 2025 debut, will manage workflows with near-zero human input — analyzing tasks, resolving calendar conflicts, and make end-to-end travel bookings. Anthropic’s “computer use” agents can type, click, and navigate the web on behalf of users. Google’s Project Mariner, powered by Gemini 2.0 will autonomously browse the web and perform actions like clicking and filling forms – effectively handling tasks like shopping, booking, and summarizing articles. The future is automated autonomous.  

Are AI Agents just fancier LLMs? Not quite. While LLMs like GPT process and generate text, AI agents take it further — they act autonomously. Think of LLMs as the brain, and AI agents as the brain with hands. A typical LLM might draft an email, but an AI agent could not only write the email but decide the best time to send it, schedule it, and follow up autonomously. OpenAI’s Operator is rumoured to go even further — handling schedules, resolving conflicts, and notifying stakeholders without human intervention.  Consumer interest in AI agents has surged over the past 24 months, mirroring the rapid growth of startups in the space. 

Market Map: A Booming Landscape with 200+ players building agentic solutions Interested in these startups? Download a curated list of 75+ AI Agent startups, with founder profiles, funding, HQ and more. 
 Fast-growing, competitive space200+ startups are actively building agentic tools, targeting areas with manual, repetitive workflows. The abundance of open-source infra for agents, e.g. LangChain, AutoGen, Dify AI have lowered barriers to entry, creating a fertile ground for innovation and faster product launches. Agents are addressing the problem of labor-intensive tasks in industries like healthcare and legaltech – driving operational efficiency by taking over end-to-end workflows. Oracle Health’s Clinical AI Agent has reduced time spend on documentation by 41%. Spellbook’s Legal AI Agent “Associate” is capable of planning and executing projects across various documents and applications, check its work, and adapt to accomplish larger scope assignments.
 Big Tech is loomingBig Tech has made bold moves in AI agents – OpenAI and Anthropic are building foundational solutions, while Microsoft, Google and Salesforce are integrating agentic capabilities into their ecosystems. These moves validate AI agents as core to future software, and we expect this to set off partnerships and acquisitions as smaller players push innovation and Big Tech seeks scale.
 Early adoption is prominent in coding, knowledge management, and financeMature AI agents have found their footing in domains with repetitive workflows. Coding: Coding agent Devin has been widely popular some time now, delivering 12x efficiency and 20x cost savings. Magic.dev, Copilot Workspace, Factory, Cosine are all essentially coming up with AI software engineers that can autonomously handle entire workflows from writing code to debugging, code optimization, and routine maintenance. Knowledge work: Hebbia enhances knowledge work by automating complex tasks and providing advanced search capabilities across large datasets. Their AI agents are tailored for industries such as finance, law, and consulting, enabling professionals to transform vast amounts of information into actionable insights efficiently. Orby AI, Wokelo are also facilitating similar research and workflows, helping companies automate due diligence, sector research, and compliance tasks. Fintech: Another hotbed for AI agents, where tools are automating accounting tasks, risk assessment, fraud detection, and compliance. Tools like Sema4.ai, Parcha are building AI compliance suites, taking over tasks like invoice reconciliation, receivables matching, AML screening. Norm AI is helping Fortune 100 companies with critical regulatory assessments.
 Despite the hype, challenges remainAI Agents are yet to tackle technical issues such as unreliable performance in long tasks, hallucinations, and compounding errors in multi-step processes. Agents also struggle with precise digital interactions, leading to issues such as unnecessary scrolling and dragging. Unpredictable behaviors also pose risks – for example, Anthropic’s Claude “suddenly took a break” from coding and began to check out photos of Yellowstone National Park. Enterprise adoption of AI Agents is likely to face operational challenges like integration hurdles, data privacy risks, and a lack of robust safeguards for tasks like payments or sensitive data handling. 

Funding Landscape: Investor enthusiasm and mega-Rounds defined 2024  Investor dollars are pouring in for AI agent startupsThere have been 150+ funding deals in this space in 2024 (81.4% YoY increase). AI agent startups have collectively raised $1B+ this year alone, proving as one of the hottest areas in tech. Mega-rounds and robust investment activity marks the 2024 landscapeSeveral companies like DevRev (100M Series A), Hebbia ($130M SeriesB), Sierra AI ($175M Series A) have seen huge rounds in 2024. Both DevRev and Sierra AI achieved unicorn status. Y Combinator has 130+ deals in the AI agents space. Andreessen Horowitz (a16z), Sequoia, Abstract Ventures, and Lightspeed Venture Partners are also highly active. Google Ventures has been a strategic investor with participation in rounds for Hebbia, Harvey, Flank AI, Moonhub and Pecan. 

Website Visits: Wordware seeing outstanding traction; Hebbia, Cosine capturing significant interest as well  Wordware, a YC-backed no-code platform for building AI Agents, shattered records with a viral Product Hunt launch in August, drawing 15M+ visits in a single month. Wordware has set a benchmark for driving organic user adoption with its intuitive platform enabling users to create AI agents using natural language, simplifying AI development for non-technical users. Notable clients include AiSDR and Runway. Hebbia AI has seen ~300% QoQ growth in web visits. Hebbia is gained popularity as one of the mature AI agents in knowledge management, designed to analyze large datasets and provide actionable insights. Hebbia's clientele includes major asset managers, law firms, and Fortune 100 companies using its AI agents for tasks such as due diligence and data analysis. Dify – an open-source, no-code AI agent building platform – got 2M+ web visits in Q3’24. Dify is also hugely popular amongst developers – with 50K Github stars in Q3’24. Dify's intuitive interface combines AI workflow, RAG pipeline, agent capabilities, model management, observability features and more, letting the user quickly go from prototype to production. 

Looking Forward AI Agents are poised to redefine work, with an incoming Big Tech Bang
AI agents are set to bring a new era of productivity for individuals and businesses – marked with high consumer interest, rapid startup activity and robust investments. BigTech moves are likely to trigger partnerships, acquisitions, and increased competition from smaller players.
 Agent building will become accessible to everyone, not just developers
Building AI agents will not limited to developers in the future. Open-source frameworks like LangChain, Dify, and Autogen are making the infrastructure widely available, while startups like Wordware, AgentGPT, and Autogen are creating tools that allow anyone to design their own AI agents. This shift empowers non-technical users to build agents for everyday tasks like calendar management, travel booking, and studying – potentially transforming how we handle routine activities in our daily lives.
 Investments are fueling growth, expect mega-rounds to continue
Frequent funding rounds and high investor confidence have cemented this space as one of the most dynamic in tech. Wordware, Hebbia, Dify are likely candidates for upcoming rounds, given their strong traction and value delivery. We expect Hebbia to achieve unicorn status soon!
 Impact on integration software ecosystems will be transformative
AI agents will reshape how software integrates and interacts with enterprises. The API economy might fundamentally change, with agents bridging the gap between systems. Beyond external integrations, agents could potentially redefine how internal functions like sales, operations, and compliance interact seamlessly within organizations. Multi-agent systems that facilitate such interactions will be interesting to watch out for.