Agent by P&L Part II
Who are the Agents at each line of the P&L?
Please subscribe, and share the love with your friends and colleagues who want to stay up to date on the latest in artificial intelligence and generative apps!🙏🤖
In our last post (Agents by P&L Part I), we broke out a standard software P&L to explore where AI Agents could play a role and deliver impact. In this post, we will dig into the specific agents building at each layer of the P&L and where we believe the largest opportunities will be.
Let’s dive in!
Cost of Goods Sold
Cost of Goods Sold (COGS) typically makes up ~22% of the average public SaaS company’s revenue. We see opportunity for agents that can help boost margins by assisting humans and replacing manual, large-scale tasks.
SRE & Engineering → These are agents that act as co-pilots for site reliability engineers (SREs), by autonomously triaging and root causing alerts from production applications. (e.g. Cleric, Rootly, Resolve, Traversal)
Cloud Cost Management & Optimization → These agents and tools help manage spend on cloud hosting and servers (like your AWS and Kubernetes clusters), and autonomously or semi-autonomously optimize infrastructure performance. (e.g. ScaleOps, Cast AI, Sedai)
Customer Service Automation → Agents to intelligently route, triage, and resolve customer issues and inquiries. (e.g. Decagon, Forethought, Sierra, MavenAGI)
Sales & Marketing
Sales & Marketing (S&M) typically accounts for ~29% of average public SaaS company revenue. We believe this is one of the most natural segments for AI-driven disruption. As a key growth driver, sales can leverage agentic workflows to automate many functions entirely and increase top-of-the-funnel opportunities for businesses.
Sales:
Agentic BDR/SDR Solutions: Agentic systems are replacing traditional BDR/SDR roles by automating prospecting and lead generation. AI agents leverage data enrichment to identify target personas, personalize outreach, and generate top-of-funnel leads instantly. (e.g., 11x, Clay, UnifyGTM, Common Room)
AI in Closing Deals: These enhance AE’s ability to understand prospects, predict reactions, and tailor offerings. These agents may be offering AI-powered research on the customer, real-time proposal and contracts based on specifically what features the customer wants, and AI-created demos for end-customers. (e.g., Arcade Software, Tavus, HeyGen)
Next-Gen CRMs: CRM systems are shifting from static records to intelligent, real-time decision-making platforms. These agentic CRMs ingest data, integrate with workflows, and make smart recommendations for customer outreach prioritization and increasing sales efficiency. (e.g., Clarify, Day, Rox)
Marketing:
SEO: Traditional SEO is becoming less relevant as AI-driven search engines like Perplexity and ChatGPT gain traction. With Google no longer dominating through blue links alone, the future of SEO will focus on optimizing for AI-generated answers rather than traditional rankings. If AI agents become primary decision-makers we can see these replacing traditional blue links. (e.g., GrowthX, AirOps, Unusual)
AI-Driven Content & Branding: Marketing is shifting toward dynamic landing pages that adapt based on visitor intent and profile. AI agents could personalize marketing pages in real time, ensuring content resonates with each user. Instead of static web content, brands will need AI-powered systems that continuously update pages based on new offerings, trends, and customer data. (e.g., Gradial, Coframe)
Research & Development
R&D expenses typically account for 17% of revenue. AI agents are helping transform R&D from a cost center into a revenue center: accelerating innovation, reducing development cycles, and enabling companies to bring new, high-impact products to market faster than ever before.
Developer Tools: AI-powered code implementation and generation tools enhance developer productivity, especially for entry-level engineers. Low-code/no-code platforms streamline development, with tools tailored to fit an organization's specific workflows. (e.g., Bolt, Cursor, FactoryAI, Grit).
QA Testing: These agents offer end-to-end testing for companies including automating bug detection, regression testing, and deployment validation. Agents can be constantly running tests in the background. (e.g., QA Wolf, Ranger)
Product & Design: We are starting to see AI-driven prototyping and experimentation agents accelerate MVP creation. AI research and testing agents may also start to emerge and automate usability testing. We are also starting to see more UI/UX agents dynamically adapt interfaces based on user behavior and preferences. (e.g., Vizcom, Thesys, Framer, Inflight, Recraft)
Model Training, Development, and Evaluation: These are AI agents that help with model training, fine-tuning, RLHF, and model evaluation. Emerging tools optimize model performance, automate testing, and enhance evaluation, ensuring continuous improvement and deployment efficiency. (e.g., Galileo, AgentOps, Braintrust, OpenPipe)
General & Administrative
G&A expenses typically account for ~12% of revenue. The pejorative for G&A is that they are “back-office” cost centers. We see AI Agents as a channel for transforming this overlooked “back-office” into a competitive advantage for future-looking companies. By supercharging legal, finance, compliance, HR, etc., companies can offer a best-in-class experience to all of their constituents, including vendors and employees, not just customers.
Finance, Planning & Analysis → These are agents that supercharge the office of the CFO and finance teams to plan, budget, and forecast a company’s financial operations over the year. These tools can touch everything from procurement officers to accountants and we are seeing both fully automated and semi-automated solutions. (e.g. BRM, Basis, WiseLayer, Comulate)
Human Resources → While HR is vast, we think AI Agents can be most effective in recruiting, particularly in automating the sourcing and vetting of candidates at scale, a typically tedious effort that requires scrolling through endless LinkedIn profiles. (e.g. Mercor, Maki AI, Moonhub, Seekout)
IT → ServiceNow has been the dominant IT Service Management (ITSM), but we are increasingly seeing a new crop of Agents automating the help desk and modernizing IT operations. (e.g. Ravenna, Atomicwork)
Security → We’ve previously written about the intersection of AI and cybersecurity. Still, increasingly we are seeing agentic solutions deployed to help assist the security operations center (SOC) and in pentesting / offensive security. (e.g. Dropzone, XBOW, Torq).
Taxes & Interest
The last cost center we have to account for before arriving at net income is taxes and interest. Taxes and interest can vary wildly based on how much EBIT a company is producing, which geography they are in, whether they have debt on their balance sheet or not, etc. But we are beginning to see several agentic solutions that use AI to help monitor sales, manage tax collection settings, and automate remittances to streamline and optimize the tax process for businesses (e.g. Numeral, Black Ore, Numiro).
Conclusion
AI agents are transforming the P&L, driving revenue growth, cutting costs, and improving efficiency. These agents will function as semi-autonomous or fully autonomous ‘co-workers’, working toward the same goals as employees.
Revenue Growth: AI-driven sales, marketing, and customer engagement optimize outreach, boost conversions, and unlock new revenue streams.
Cost Reduction & Efficiency: AI automates R&D, customer support, and finance, reducing overhead while enhancing decision-making.
Margin Expansion: AI streamlines supply chain, logistics, and operations, minimizing waste and maximizing productivity.
It’s an exciting time to be an AI Agent!