Imagine this scene: A founder demos their new AI feature: a chat interface, a smart dashboard, a workflow copilot… and receives a positive but slightly flat reaction. That founder sees a spark in the customer’s eye for a moment; a vision of a different way to run the business. But then the customer goes back to tallying hours and doing payroll in separate tabs, chasing invoices in another system, and manually reconciling inventory in a spreadsheet.
The AI didn’t actually do anything. It informed. It surfaced. It recommended. And then it handed the work back to the owner.
Digital platforms serving small businesses (SMBs) live this scene daily in early 2026. They’re living the defining limitation of first-generation AI: intelligence without execution. And for platforms serving SMBs specifically, whether the restaurants, home service companies, healthcare practices, or retail shops that make up the backbone of the U.S. economy, closing that gap takes more than AI. It will take new infrastructure.
The AI models available today are extraordinary. But every competitor can access the same models at roughly the same price per API call. From a differentiation lens, LLMs are commoditized. What isn’t commoditized: the data and workflow that allow AI to act on what it knows.
And for the thousands of digital platforms targeting SMBs, particularly the ~5,000 vertical SaaS platforms (Pitchbook as of Dec 2024), embedded fintech is the most consequential piece of that infrastructure a platform can own.
Embedded fintech connects the dots between AI, cash flow, and workflow. Executives, investors, and fintech thought leaders increasingly agree that only by connecting these dots will the next generation of platforms realize the potential of AI applied to the many of the challenges SMB owners and operators face today. The platforms that solve these pains for SMBs will benefit across three key dimensions: differentiation, user value, and business model.
1/ The Moat Is the Data, But Only the Right Data
Few frameworks have influenced vertical SaaS leaders more than the concept of data gravity. Data Gravity, as articulated by Tidemark founder Dave Yuan, refers to the unique advantages that can accrue to systems that manage critical business data and that those benefits compound over time. “Data gravity… it’s the biggest pool of data. The data that is most mission-critical,” Yuan has said.
Data gravity creates stark implications for vertical SaaS platforms in this new era: AI alone is not your moat. Competitors can access frontier models and fine-tuning open source models yourself yields diminishing returns. How your AI uses operational data from your customers’ businesses, i.e. the data that AI actually needs to do useful work, will determine if your AI products create durable value or get copied in six months.
Not all data has equal gravity. A restaurant’s Yelp reviews are data. Past reservations are data. But employee payrolls, sales data, and vendor costs? Those inputs determine the viability of the business, which makes them the most consequential levers to get right. According to Tidemark’s 2025 Vertical & SMB SaaS Benchmark Report, companies with fintech and back-office control points exhibit the highest gross and net retention of any category, precisely because this data captures the core cash flow or workflow (or both).
Building embedded fintech not only enables a new revenue stream. These products help generate data gravity by connecting the most mission-critical datasets in a small business. Connected data offers another level of value via the context graph: the dense, interconnected web of proprietary business data, workflows, and historical decisions that live inside systems of action.
As Matt Brown, the widely-read early-stage investor at Matrix, points out, the “…most valuable [context graphs] already exist inside vertical software.” Interestingly, by connecting key datasets, context graphs generate additional value (particularly for AI), by “…capturing the decision traces that systems of record miss… a queryable record of business logic: the reasoning, precedents, and decision traces that explain why things happened, not just what happened.” AI agents find the why particularly valuable: an agent reasoning about labor costs, cash flow timing, or staffing decisions reasons from a deeper ground truth versus approximating from incomplete signals.
The context advantage goes deeper still. Leaders in the developer and product management community have continually advanced new techniques to allow AI agents to perform better with limited context windows. The term of art is compaction and the core insight: what you feed an LLM matters as much as the model itself. As product thinkers like Tal Raviv have noted, agents that receive clean, structured, semantically rich context outperform agents drowning in raw, noisy data.
Compaction implies a massive, underappreciated advantage for vertical SaaS platforms over horizontal ones. When handed a small business’s financial life, a standalone chatbot or AI assistant has to parse a chaotic dump of PDFs, spreadsheets, and disconnected records. A vertical SaaS platform with both cash flow and workflow data can feed its AI something entirely different: “This employee worked 47.5 hours last week at $18.25/hour; California overtime applies above 8 hours/day; there is an active child support garnishment of $312/month; direct deposit is configured to two accounts.” That offers extraordinarily efficient, executable context.
And executable context separates an agent that can act with confidence from one that hedges, hallucinates, or fails.
The platform that owns structured cash flow data via embedded fintech products doesn’t just have more context. It has better context. In agentic AI systems, context quality equals leverage.
2/ Intelligence Without Execution Is Just Advice
Picture a specific Friday morning. A restaurant owner gets a notification at 8 a.m. that the Mayor just announced a major event downtown the following weekend; the kind of event that will bring double the foot traffic. Her vertical SaaS platform’s AI sees everything relevant: her inventory and turnover, upcoming cash outflows (ex. payroll next Friday), her current bank balance, her current staff schedule, the overtime risk from last week, and the labor law implications of adding Saturday shifts.
It surfaces an insight: “You may want to bring in three additional staff for next Saturday and increase the order for your most popular items by 30%. But that will leave a gap to cover Friday’s payroll run.”
Now what?
The AI has done its job if the platform does not own inventory, scheduling, or payroll execution. The owner logs into separate systems, manually changes the inventory order, starts texting employees and updating schedules, maybe calls her bank to draw on a line of credit, and hopes nothing everything comes together ahead of the weekend.
The insight was valuable. But the work was still hers. As was the stress that comes with the ‘Chief Everything Officer’ title that millions of small business owners have opted into.
Euclid Ventures has termed this the Dispatcher Problem. As AI agents become capable of routing tasks, attention, and money — i.e. deciding not just what to do but where to send the concrete tasks — the platform that owns the destination captures the economic value. If your AI dispatches the user to an external provider to complete the loop, someone else’s execution layer will monetize the value created by your intelligence layer.
Embedded fintech can help close this loop. The same AI that surfaces the insight can suggest an updated schedule, change the inventory orders across multiple wholesalers, apply the correct overtime for last week to calculate the cash gap, even source credit options… all with approval but without the owner ever leaving the application. The workflow begins and ends in one place.
This is the architectural definition of a system of action rather than a system of record. The scenario above illustrates why the most defensible vertical SaaS platforms of the next decade will not be the ones with the best AI model. The platforms that own both the intelligence layer and the execution layer will outperform platforms where AI just tells users what to do next but can’t actually do it.
The data bears this out at scale. Toast, the storied all-in-one solution for restaurants, built a business where financial technology solutions (embedded payments, payroll, capital, and more) now represent 82% of total revenue. Shopify’s “Merchant Solutions” segment, which includes embedded payments and capital products, accounts for over 75% of total revenue. While outliers today, they demonstrate what happens when a vertical SaaS platform owns both the workflow and the financial rails underneath it.
3/ Determinism Is a Feature, Not a Bug
This discussion has a technical tension at its core, one that deserves direct treatment: large language models are probabilistic. They generate the most statistically likely response, not the provably correct one. For brainstorming, drafting, or summarizing, that is fine. For executing transactions, it is not.
You cannot “hallucinate” a tax payment. A plausible-but-incorrect direct deposit amount is not a near miss; it’s a failed payroll run, a potential regulatory violation, and a significant breach of trust with every employee affected.
This tension resolves elegantly in the right architecture. In fact, the resolution offers a strategic advantage for platforms with embedded APIs. As Andrew Oved from Reformation Partners has argued, “determinism makes the value of traditional vertical SaaS observable: you can see the results immediately.” Probabilistic and non-observable, AI on the other hand, still represents an incredible opportunity but requires a very different value model according to Oved. The optimal combination will likely fall in the middle: “the most valuable vertical-specific companies will likely blend both paradigms, combining deterministic infrastructure (SaaS) with probabilistic intelligence (AI).”
Very few SMBs will trust a fully autonomous AI, even a vertical-specific AI, to manage payroll or take out a loan unsupervised. It seems more likely they would come to trust a hybrid system where the probabilistic layer and the deterministic layer each do what they do best:
- The LLM handles unstructured intent: reasoning that the owner wants to schedule additional employees and purchase more inventory because of a big weekend, flagging the employees that already crossed an overtime threshold, and recommending the new schedule.
- The embedded fintech platforms handle deterministic execution: executing the pre-approved working capital loan or applying the correct federal and state tax rates for every employee’s jurisdiction and generating the actual direct deposits with zero tolerance for error.
The best of both worlds: a forward-focused advisor with a reliable operations team behind them. The AI does the reasoning and the communication; the embedded finance platforms do the work with the precision that a regulated financial process demands.
This architecture matters enormously for SMB trust. Small business owners are not looking for an AI that promises to do everything. According to a 2025 survey by HiBob, 64% of employees have experienced stress or disruption from a payroll error, and research from Ernst & Young found that fixing a single payroll error costs an organization an average of $539.
These are not abstract risks to an owner. They are the kind of mistake that costs real employees a real paycheck and triggers a compliance audit. AI must earn their trust. AI can do just that by executing specific things reliably, within guardrails users understand, and demonstrating that the platform gets smarter over time.
4/ The Business Model Hiding Inside the AI Feature
Questions about future revenue models and monetization cut to the heart of why embedded fintech and AI belong in the same conversation.
Small business owners understandably hesitate to pay a meaningful premium for AI features above and beyond their current SaaS subscription. “Save time with AI” is abstract. “$30 more per month for a chatbot” feels questionable to a restaurant owner already managing thin margins. Per-seat AI upsells require the customer to perceive and attribute value to an AI layer that, by design, works best when it’s invisible.
Savvy investors and the operators they back have already begun experimenting with new value models. According to a B2B sales leader interviewed by Bain Capital Ventures, “Our goal isn’t to sell AI. It’s to sell outcomes. AI just happens to be how we get there.”
Many thought leaders in addition to the Bain Capital team have explored how the capabilities of LLMs change value delivery and monetization. Like Bessemer Venture Partners, many come to the conclusion that vertical AI agents (ranging from fine-tuned open source LLMs to post-trained proprietary models), have a unique advantage because they compete with labor vs IT budgets. Andrew Oved, linked earlier in this post, contrasts the deterministic benefits of “traditional” vertical SaaS with the probabilistic benefits of LLM: “Vertical AI represents a class of products built not just to enable workflows, but to make probabilistic judgments within them.”
This leaves us with a spectrum: vertical SaaS (deterministic) to vertical AI (probabilistic) with a range of combinations and value models in between. From this lens, vertical SaaS platforms (and perhaps eventually vertical AI solutions) can also use AI as a conversion and activation engine for the embedded financial products that already monetize through usage rather than charging for the AI itself.
In this model, the value chain works as follows: the AI spots that a business owner repeatedly makes the same time card edits and surfaces an in-product suggestion at the right moment. Further, the AI addresses the root cause of the user’s confusion by answering compliance questions. AI lowers the activation energy to migrate away from the standalone time tracking tool, payroll provider, and separate accounting solution; i.e. to move past disconnected workflows.
Once the owner decides to manage schedules, run payroll, and account for key transactions inside a single platform, the embedded financial product captures the economic value. This may range from per employee fees to tiers upsells to interchange on payments or spend management products and origination fees on capital products. Andreessen Horowitz has noted that SaaS platforms can increase revenue per user by 2–5x by adding embedded fintech products.
Tidemark’s 2025 benchmark data reinforce this: multi-product vertical SaaS companies grew roughly 21% faster than their single-product peers, and fintech-led companies held the strongest retention profile in the entire survey.
AI products may or may not get a line item on the invoice. They may earn their keep by making embedded financial products easier to adopt and stickier to retain. But the flywheel is real, and AI is increasingly the mechanism that gets it spinning.
5/ A Candid Note on Adoption
The shifts outlined above rarely happen on the timeline that series A pitch decks suggest.
As Medha Agarwal of Defy VC has pointed out, “There has been a lot of hype about AI agents and the ability to automate previously manual workflows. While I believe there is true potential to do aspects of this, I’d argue it is often hard to completely automate away large swaths of human tasks.”
Opus 4.5 dropped in November 2025 and GPT2.5-Codex in December 2025. Claude Cowork released in January 2026. Bleeding-edge power users have only had ~3 months to figure out how to best leverage the latest tools (as of this writing). So while AI may catch up to Mdeha’s words from the Summer of 2024 soon, adoption by the average SMB owner will still take time.
Most embedded fintech products still have a long way to go in terms of availability, which precedes adoption. From Tidemark’s 2025 Benchmark Report again: while 80%+ of respondents offer embedded payments, only 23% offer embedded credit products and no other embedded fintech product, including embedded payroll, banking, insurance, or expense management surpasses 15% of respondents. (See image below from Tidemark.)
That is not to say the AI + embedded fintech combo is dead on arrival. Both product areas require significant investment and time to do right. As Medha also points out: “AI has captured the average user’s imagination of what is possible… From my customer conversations with a broad swath of professionals from lawyers to service technicians to doctors, they are more open to new software than I’ve ever experienced.”
Beyond AI hype, there are several reasons to think that AI + embedded fintech can systematically lower the activation energy of each step towards greater adoption of all-in-one solutions, thereby delivering the data foundation for better AI and in turn generating AI trust via reliability in lower-stakes contexts before taking on the highest-stakes ones:
- AI lowers switching costs. Take payroll: switching payroll providers mid-year introduces operational risks for a business owner as well as data migration risks. The owner and/ or the payroll provider must correctly migrate historical employee records, prior tax filings, accrued PTO balances, garnishment histories, etc. Ditto for inventory management, expense management, accounting, and more. AI can compress this work dramatically: mapping messy historical data to new schemas, flagging discrepancies, and walking users confidently through setup that would otherwise require a human specialist. This is not a glamorous use case but it is an enormously valuable one.
- Trust compounds through smaller wins. The AI that earns trust on scheduling recommendations, overtime alerts, and cash flow projections eventually gets trusted to manage the cash flow itself. Platforms that think about this as a staged trust journey, not a single product launch, may move faster in the long run because they won’t ask owners to take a leap of faith. They’ll ask them to take the next obvious step.
- Control points remain sticky. A platform that owns scheduling, time tracking, embedded payroll, and the AI layer connecting them creates something nearly impossible to dislodge: the authoritative record of every client served, every hour worked, every dollar owed, and every business decision that flows from them. In Tidemark’s framework, vertical SaaS wins when platforms identify and solve for control points. Strong control points make for sticky customers; while the future remains unwritten, AI and embedded fintech appear poised to further compound the retention advantages of control points in the future.
6/ Conclusion: The Architecture of the Next Decade
The platforms that will define SMB software over the next ten years are not the ones with the best model fine-tuned on the most data. They will be the ones who close the loop between intelligence and action. When AI doesn’t just surface what needs to happen, but makes it happen, reliably, and within the regulated guardrails that zero-mistake environments demand, then the all-in-one solutions that many pitch start to look like reality.
Embedded fintech gives AI the hands it needs to become an agent.
Data gravity and the context graph can make AI reasoning trustworthy. Solving the dispatcher problem makes AI action possible, while deterministic features and embedded fintech APIs can make AI-driven financial workflows safe. And new business models, or contrarian takes on AI value, can make all of it profitable without asking an owner to pay a separate line item for intelligence she can’t yet fully evaluate.
The restaurant owner doesn’t want a smarter dashboard. She wants to open her app on a Friday morning, see that the weekend is going to be busy, and have the system ask: “Want me to increase inventory and bring in two extra servers on Saturday without increasing your overtime costs?”
That’s the product. Everything else is the path to build it.
