Agents by P&L (Part I)
Where are the large opportunities for Enterprise Agents?
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!🙏🤖
The buzz around Enterprise AI agents is louder than ever, and for good reason - they’re set to transform how businesses operate. But amidst all the excitement (and vague handwaving), it’s critical to focus on where the biggest opportunities for AI agents are. One way to approximate opportunity size is by looking at an enterprise’s P&L and what drives its revenue, costs, and profitability.
Why?
The best places for an enterprise to derive real ROI from AI Agents will likely align with its largest revenue, cost, and labor pools.
AI agents are reshaping industries in two fundamental ways:
Acting as a "co-pilot" for humans, enhancing productivity and enabling employees to work smarter.
Taking over entire workflows, automating tasks and driving efficiency by replacing human efforts where possible.
While there will be large opportunities for both #1 and #2, we will likely see more “co-pilots” supercharging humans in the workforce before fully automated systems. Importantly, while tech/software companies are likely to be at the forefront of adopting enterprise agents, the even larger opportunity will be in non-software verticals (e.g., real estate, oil and gas, healthcare, discretionary consumer, etc.). Think about when Costco and Chevron start adopting agents!
This will be the first of a multi-part series. In this post, we’ll break out a standard software P&L to explore where AI agents can play a role, and identify the areas where AI can deliver the greatest impact and unlock new value. In Part 2, we’ll dig into which specific agents we see going after each of these line items and how we believe the space will evolve.
Let’s dig in!
Starting At the Top(line) - Revenue
Revenue is in many ways the lifeline of every business and represents the total income generated from the sale of goods or services. As companies deploy AI agents they want to see what tangible ROI they can drive, either through increased revenue or decreases costs (COGs & OPEX) which we’ll dive into next.
So where could agents help with revenue?
Product intelligence: In most cases, better products and services result in higher sales. Imagine an AI agent that analyzes customer feedback, social sentiment, and behavioral data to identify what customers truly want, or why they might be unhappy. Generating insights to address key issues will allow businesses to refine their offerings, leading to increased revenue.
Forecasting demand: Understanding market trends and customer needs is pivotal to capturing revenue opportunities. Today, businesses pay consultants large sums to analyze macroeconomic factors and industry trends. An AI agent, however, could automate this process - monitoring data in real-time, identifying emerging patterns, and forecasting how external forces may impact your business. This would give companies an edge against the competition.
Pricing strategy: Determining the right price for a product or offering is a complex challenge for any business. Companies often spend millions on consultants to help optimize pricing and packaging strategies. What if an AI agent could do this instead? By running simulations, testing different price points with select user groups, and analyzing performance, AI agents could optimize pricing to maximize revenue while maintaining customer satisfaction.
Revenue is unique in that its drivers - like sales, marketing, product development - are often categorized as expenses in financial statements. We see enormous opportunity for Agents to drive meaningful revenue, particularly in areas where companies have historically turned to consultants (e.g. McKinsey) or BPOs (e.g. TCS). AI agents have the potential to democratize this expertise, providing scalable, data-driven insights that are faster and more cost-effective.
Cost of Goods Sold (COGS)
Cost of Goods Sold (COGS) refers to the direct costs associated with producing goods or services that a company sells during a specific period. It includes all expenses directly tied to the production process, from raw materials to labor, and other costs required to deliver the final product or service to the customer.
Software businesses: Software companies typically boast high gross margins (~78%, per Meritech’s SaaS index) due to low marginal costs. This means COGS make up ~22% of revenue, primarily driven by:
Customer Success/Support: Typically includes assisting users with onboarding, troubleshooting, and ongoing user assistance. We see AI agents can play a key role here by automating repetitive tasks, providing 24/7 support, and delivering personalized, efficient solutions to enhance the customer experience.
Hosting & Compute: Includes cloud and infrastructure costs by hyperscalers. We see emerging AI agents and agentic workflows are optimizing backend infrastructure management, helping not only to reduce infrastructure and compute costs, but delivering insights in how to make overall systems more performant.
Non-Software businesses: Non-software businesses (e.g., business services, healthcare, CPG, real estate) tend to have higher COGS, with gross margins typically ranging from ~30% to 50%. Given the lower GMs within non-software businesses, we see opportunity for AI agents to thrive here. Key drivers include:
Direct materials: The raw materials or components used to produce goods. AI agents can reduce sourcing costs, optimize supply chains, and work alongside robotics to automate processes like packaging and sourcing.
Direct labor: Includes wages and benefits for employees directly involved in production, implementation (e.g., installing tires), or service delivery (e.g., consulting or book-keeping). Many industries are currently facing significant labor shortages (e.g., accounting, tax & audit, automotive, consulting, law). AI agents can augment human capacity, enabling one person to oversee many processes, thereby addressing talent gaps and driving cost efficiencies.
Operating Expenses
Operating expenses are the ongoing costs a business incurs to maintain its operations, including expenses like salaries, rent, utilities, and marketing. These are distinct from costs directly tied to producing goods or services which show up as COGS above.
For most companies, Opex is broken into three main types of expenses: 1) Sales & Marketing; 2) Research & Development; and 3) General & Administrative.
Sales & Marketing
As the name suggests, sales & marketing costs include any variable expenses incurred to promote a company’s products or services or drive growth. This includes everything from salaries and commissions for sales and marketing teams, promotional campaigns, trade shows, and customer acquisition costs (CAC).
For most mature and growth-stage software companies, this is typically the largest bucket of Opex. According to Meritech’s SaaS index, the average spend on S&M across public software companies is 29% of revenue. That figure can vary significantly, with faster-growing companies (e.g. Rubrik, Monday) spending 50-60% of revenue on S&M, and category-creating private companies spending upwards of 150% here!
So where do we see opportunities for AI agents to play a role?
SDR & BDR automation → Sales Development Reps (SDRs) and Business Development Reps (BDRs) are typically at the front-end of a sales process. They focus on generating and qualifying leads for a company’s pipeline, either through inbound harvesting or outbound prospecting. These reps are usually right out of college and have to work through significant volume.
Pipeline generation → We see opportunity for AI to massively expand and streamline how pipeline is generated. While this work was previously manual and required a good deal of guesswork, AI can help stitch together multiple different data sources and signals to create higher-converting, personalized pipeline.
Research & Development
R&D costs are investments a company makes in creating new products, improving existing ones, or developing innovative technologies and processes. These expenses include items like salaries for engineers, product development and design costs, data analytics, and testing.
R&D is usually the second largest expense bucket in Opex, averaging to 17% of revenue across the SaaS index. Highly product-oriented companies like Atlassian and Asana spend 30-40% of revenue on R&D, while startups that are just starting to build product or have large research labs can spend 50-100% of sales on R&D.
Areas we believe Agents can be impactful include:
Coding → Undoubtedly this is perhaps the hottest area of investment for AI agents, as the spend on engineering and coding tools within the enterprise can be massive. We expect to see plenty of demand and competition here.
Product Design → While not as in the spotlight as front-end and back-end coding, design has increased in importance and relevance over the past ten years, producing generational companies like Figma. We are seeing an increase in “AI for design” tools, or agents that can supercharge existing product and graphic design processes.
Quality Assurance (QA) → QA is an underappreciated but important part of the software and product development lifecycle, ensuring the end product or service meets specified standards of quality before being delivered to customers. AI Agents can help increase the number of releases and lessen the time between cycles, improving functional, performance, and regression testing.
General & Administrative
G&A is the “boring” bucket of Opex but necessary to the operations of a well-functioning business. These expenses typically include accounting, financial, legal, and IT.
The average SaaS company spends 12% of its sales on G&A, and most CFOs aim to keep G&A to a minimum, as they view this area as a cost center (vs. S&M and R&D which can help drive new revenue). This doesn’t mean AI agents can help here though. In fact, we believe AI agents can be massively helpful in G&A to help streamline costs, build more efficiencies, and help the entire business operate more smoothly overall.
Where do we see the biggest opportunities?
Accounting & Tax → Every company, particularly public ones, must have their accounting under control. But managing accounts receivable, accounts payable, financial reconciliation, reporting, and audits is no easy feat, particularly as enterprises and customers get more complex. With access to structured data and automation of repetitive tasks, there is a wide berth for agentic processes.
Recruiting → People are the lifeblood of every business, and recruiting is an often-overlooked area. HR teams are often understaffed and don’t have enough resources to meet every candidate they would like to, particularly in a tight labor market.
IT & Security → Cybersecurity risks are on the rise, and ~60% of security teams are understaffed. This is an ominous combination and will undoubtedly lead to more costly cyber attacks. The IT and Security Operations Center (SOC) is becoming a more prominent area within the enterprise, and we are already seeing AI Agents deliver results to these teams.
Conclusion
Enterprise AI agents are poised to revolutionize how businesses operate, offering transformative opportunities to optimize revenue, reduce costs, and address labor challenges. By analyzing an enterprise’s P&L, we can pinpoint the areas where these agents can deliver the most impact - whether by acting as co-pilots to augment human productivity or automating entire workflows to unlock new efficiencies.
While tech companies are often early adopters, the largest untapped potential lies in non-software verticals such as healthcare, real estate, and consumer goods, where AI can tackle significant inefficiencies and labor shortages. As we continue exploring this space, it’s clear that AI agents are not just tools, they’re catalysts for redefining industries and unlocking unprecedented value.
In our next post we’ll dive deeper into some of the AI agent companies we’re seeing within each functional group and why we’re excited about the promise of what they’re building. Stay tuned!