BILL's AI Strategy: How the SMB Finance Automation Leader Understands Its Customers in 2026

TL;DR#

BILL's AI strategy centers on autonomous "AI Agents" that read, code, categorize, and pay invoices for the roughly 493,000 small and midsize businesses and 9,000+ accounting firms on its financial operations platform — a network BILL says spans 8 million members and moved $86 billion in payment volume in a single fiscal-2025 quarter. In May 2026, CEO and founder René Lacerte told investors that AI is "no longer one priority among three — it is our No. 1 priority," and the company announced a 30% workforce reduction to reorganize around agentic finance. BILL's Invoice Coding Agent has automated 1.2 million invoices, its Accounting Agent has saved customers 1,100-plus hours of manual work, and over 100,000 customers have adopted its agents. But BILL's core audience — busy SMB owners and accountants — is exactly the cohort that ignores surveys, which means the "why" behind feature adoption and abandonment hides in usage logs no dashboard fully explains. This is where conversational AI interviews close the gap: instead of forcing a survey-averse audience into rating scales, an AI interviewer asks an accountant why they trusted the Invoice Coding Agent but turned off auto-categorization, and follows up in their own words. That is the model Perspective AI is built on.

What is BILL's AI strategy?#

BILL's AI strategy is the company's plan to embed autonomous AI agents across accounts payable, accounts receivable, spend, and cash-flow workflows so that financial operations for SMBs become "touchless" — invoices get coded, categorized, and paid with minimal human input. Announced as the centerpiece of fiscal 2026, the strategy reframes BILL from a bill-pay tool into what Lacerte calls a platform for "intelligent financial operations," and in May 2026 the company elevated AI to its single top priority and cut roughly 30% of its workforce to fund and focus the effort.

This post is written for product, CX, and research leaders at fintech and SMB-software companies who want to understand how a category leader is operationalizing AI — and where even a data-rich platform like BILL still struggles to hear why its customers behave the way they do. If you build for small businesses or accounting firms, the listening problem BILL faces is the same one you face. Our guide to AI-powered customer experience from first touch to renewal maps the full arc this post zooms into.

BILL by the numbers: the scale behind the strategy#

BILL operates one of the largest financial-operations networks serving small business, processing tens of billions of dollars in payments across hundreds of thousands of customers. The scale is what makes the AI bet rational — and the customer-understanding gap expensive. The headline figures, drawn from BILL's fiscal-2025 reporting:

According to BILL's fiscal 2025 fourth-quarter results filed with the SEC, the company crossed $1.5 billion in annual revenue. Those numbers describe a business with enormous quantitative visibility — BILL can see every invoice, every payment, every late vendor. What the ledger cannot tell BILL is the reasoning: why a 12-person agency abandoned the spend module after onboarding, or why an accounting firm rolled the Accounting Agent out to three clients and then paused. That gap is the subject of this article, and it mirrors the pattern in our Stripe AI customer research breakdown and our PayPal customer-discovery analysis.

Where BILL uses AI today: the agent stack#

BILL uses AI today primarily through a stack of task-specific "AI Agents" that automate the most tedious parts of bookkeeping and accounts payable. Rather than a single chatbot, BILL has shipped discrete agents that each own a workflow. Based on BILL's product announcements and reporting in early 2026:

• Invoice Coding Agent — extracts and codes complex, multi-line invoices automatically. BILL reports this agent has automated 1.2 million invoices, removing one of the most error-prone manual steps in AP.
• Accounting Agent — automates the manual parts of bookkeeping and month-end close. BILL says it has saved customers 1,100-plus hours of manual work since launching in October.
• W-9 Agent — autonomously emails vendors to collect and enter W-9 tax forms, eliminating over 80% of the manual steps in vendor compliance.
• Auto-categorization and receipt matching — AI matches receipts to transactions and categorizes spend, reducing data entry and "all but eliminating incomplete expense reports."
• Predictive vendor search and invoice prediction — the platform learns from how each business uses BILL to surface faster, more accurate vendor matches and predict line items for repeat vendors.

As CPA Practice Advisor reported in February 2026, these agents are aimed squarely at accountants — eliminating busywork "from complex invoice coding to vendor compliance collection." The strategic logic is sound: automate the drudgery, free the human for advisory work. It's the same play we documented in our Workday HR-and-finance-cloud customer research piece and across the Block / Square seller ecosystem.

Why understanding SMB customers is uniquely hard#

Understanding SMB customers is uniquely hard because small business owners and accountants are time-starved, skeptical of vendor outreach, and structurally averse to surveys — the exact failure mode that makes traditional voice-of-customer programs collapse for their segment. A platform can instrument every click, but instrumentation answers what happened, never why.

Consider the audience BILL serves. A bookkeeper closing month-end for fifteen clients does not have ten minutes for a 15-question NPS survey. A solo-operator landscaping business owner will not narrate their decision to stop using spend cards in a comment box. The pattern on survey fatigue is unambiguous: survey volume keeps rising while response rates keep falling, which is exactly why the annual customer survey is dying, and drop-off climbs sharply with each additional question past 15. For a customer base defined by scarce attention, surveys don't just underperform — they actively select for the wrong respondents, capturing only the rare power user with time to spare.

The consequence is a distorted picture. Quantitative tooling tells BILL that auto-categorization adoption stalls in a cohort, but not whether the cause is distrust of the AI's coding accuracy, a fear of audit exposure, or a workflow conflict with the firm's existing close process. Each cause demands a completely different fix. This is the core argument of our foundational post, why AI-first customer research cannot start with a web form: the form forces the customer to translate a messy "it depends" into a dropdown, and the nuance — the part you actually needed — evaporates on contact. For a deeper treatment of the trade-off, see AI vs. surveys: why conversations win for real customer research.

The customer-understanding bottleneck in SMB finance automation#

The customer-understanding bottleneck in SMB finance automation is the gap between usage data, which shows that a feature was adopted or abandoned, and the lived reasoning of the SMB owner or accountant, which explains why and is rarely captured at scale. BILL's AI agents widen the surface area where this gap matters.

Every new agent is a new adoption decision. When BILL ships an Accounting Agent that touches month-end close, the firm has to decide whether to trust an autonomous system with reconciliations that carry real financial-statement risk. The telemetry will eventually show the adoption curve. But by the time the curve is visible, the product and CX teams have already shipped — and they're guessing at the cause. Three bottleneck symptoms recur in SMB-finance products:

1. Trust is invisible in the data. An accountant who reviews every AI-coded invoice "just in case" looks identical in the logs to one who trusts the agent blindly — until something breaks. Usage data cannot distinguish confidence from caution.
2. Abandonment has no transcript. When a business turns off an agent, the event fires, but the reason — a single bad categorization, a partner's veto, a billing surprise — is never recorded.
3. The advisory promise is unmeasured. BILL's pitch is that AI frees accountants for higher-value work. Whether firms actually feel that shift, or just feel displaced, is a qualitative question no dashboard answers — and it matters enormously given BILL's own 30% workforce reduction tied to AI, which signals to the market exactly how seriously the company takes automation.

This is the same listening bottleneck we traced through the Notion roadmap-decision case study and the HubSpot CRM customer-research breakdown. The pattern is consistent across categories: the more a platform automates, the more its product decisions depend on reasoning its instrumentation cannot see.

How conversational AI reaches a survey-averse audience#

Conversational AI reaches a survey-averse audience by replacing the static questionnaire with an AI interviewer that conducts hundreds of natural, adaptive conversations simultaneously — meeting busy SMB owners and accountants in the channel and moment they're already in, and following up on vague answers the way a human researcher would. Instead of demanding effort up front, it earns the "why" through dialogue.

The mechanics matter for this audience specifically. A short, conversational prompt that opens with a single open question — "What made you turn off auto-categorization last month?" — clears the attention bar that a 15-question survey never could. When the accountant types "the codes were off for one client," the AI interviewer agent doesn't move to the next dropdown; it probes: which client type, how often, what would have made you keep it on? That follow-up is the entire value, and it's the one thing a form can never do. For BILL's intake-style flows — collecting W-9s, onboarding a new firm — a concierge agent can replace the static form with a conversation that captures intent while it captures data.

Done at scale, this turns research from an occasional, low-yield survey blast into a continuous listening habit. A CX or product team can run an always-on study against the cohort that just abandoned a feature, get fifty real explanations by morning, and route the findings to the roadmap. This is the model we lay out in the complete guide to voice-of-customer programs in 2026 and how to build a voice-of-customer program from scratch. It's also why we argue the survey layer itself is being rethought without the survey pattern.

What this means for fintech and SMB-software teams#

For fintech and SMB-software teams, BILL's strategy is a template for agentic automation — and a warning that automation outpaces understanding unless you instrument the "why" alongside the "what." The teams that win the SMB segment will pair quantitative telemetry with qualitative depth, not choose between them.

A practical checklist for product and CX leaders building for this audience:

• Trigger conversations at the decision moment. Don't survey quarterly; ask the moment someone adopts, abandons, or downgrades a feature — while the reasoning is fresh.
• Lead with one open question, not fifteen fields. Survey-averse audiences respond to dialogue, not interrogation. Let the AI follow up instead of front-loading the form.
• Segment by firm type and role. A solo bookkeeper, a 50-person accounting firm, and an in-house controller adopt AI agents for different reasons. Conversations surface those segments; aggregate scores bury them.
• Close the loop visibly. Uncertainty over whether feedback gets acted on is a top reason customers abandon surveys — show the cohort what changed.
• Treat research as continuous. Pair always-on conversational research with your existing dashboards so every adoption curve arrives with its explanation attached.

Whether your team owns product, growth, or support, the tooling stack matters — our roundup of the customer research tools modern product and CX teams actually use and the ranked guide to AI customer-insight platforms for enterprise both map where conversational research fits. If you're a product org specifically, Perspective AI is built for product teams; if you lead CX, it's built for CX teams, too.

Frequently Asked Questions#

What is BILL's AI strategy in 2026?#

BILL's AI strategy in 2026 is to embed autonomous "AI Agents" across its financial-operations platform so that accounts payable, accounts receivable, spend, and bookkeeping become largely "touchless" for SMBs and accounting firms. CEO René Lacerte named AI the company's No. 1 priority in May 2026 and announced a roughly 30% workforce reduction to reorganize around agentic finance, signaling that automation is now central to how BILL plans to grow.

What are BILL's AI agents?#

BILL's AI agents are task-specific automation tools that each own a finance workflow. They include an Invoice Coding Agent that has automated 1.2 million invoices, an Accounting Agent that automates month-end close and has saved customers 1,100-plus hours, a W-9 Agent that eliminates over 80% of vendor-compliance steps, and AI auto-categorization and receipt matching. More than 100,000 customers had adopted these agents as of early 2026.

How does BILL do customer research for SMBs?#

BILL relies heavily on platform usage data — invoices coded, payments made, features adopted or abandoned — to understand how its roughly 493,000 business customers and 9,000+ accounting firms behave. This quantitative visibility is enormous, but it captures what customers do rather than why, and the SMB audience is notoriously survey-averse, which limits how much qualitative insight traditional questionnaires can recover.

Why are SMB owners and accountants hard to survey?#

SMB owners and accountants are hard to survey because they are time-starved and skeptical of vendor outreach, and survey fatigue has worsened industry-wide — survey volume rose roughly 71% since 2020 while response rates fell, with sharp drop-off after about 15 questions. As a result, surveys tend to capture only atypical power users, producing a distorted picture of why the broader customer base adopts or abandons features.

How can conversational AI improve SMB finance customer research?#

Conversational AI improves SMB finance customer research by replacing static surveys with an AI interviewer that runs hundreds of adaptive conversations at once, asking open questions and following up on vague answers in the customer's own words. This reaches survey-averse audiences in the moment of a decision, captures the reasoning behind feature adoption or abandonment, and turns occasional survey blasts into a continuous listening habit.

Conclusion#

BILL has built a genuinely impressive AI strategy: a network of roughly 493,000 businesses and 9,000+ accounting firms, $86 billion in quarterly payment volume, and a stack of autonomous agents that have already automated 1.2 million invoices and saved customers more than 1,100 hours. By naming AI its No. 1 priority in 2026, BILL is betting the company on touchless finance for the Fortune 5 Million. But the harder a platform leans on automation, the more its roadmap depends on understanding why its survey-averse SMB and accountant customers trust, adopt, or abandon each new agent — and that reasoning lives in conversations, not dashboards.

That is the gap conversational research is built to close. Perspective AI runs hundreds of AI-led customer interviews simultaneously, follows up on vague answers, and captures the "why" behind every adoption decision — reaching exactly the time-starved, survey-averse audience that BILL and every SMB-software company struggles to hear. Start a new research study to put a conversation in front of your customers the moment they adopt or abandon a feature, or explore Perspective AI's plans and pricing to see how continuous, AI-first customer research fits your team.