Chris Bishop, CEO of BillingPlatform – Interview Series

Chris Bishop, CEO of BillingPlatform, is an enterprise software executive with more than two decades of experience spanning SaaS, revenue operations, customer success, professional services, and go-to-market leadership. Before joining BillingPlatform as CEO in 2026, Bishop spent nearly seven years at Conga, where he held senior roles including Chief Customer Officer, Chief Revenue Officer, and Chief Marketing Officer, helping lead the company through a period of transformation and integration. Earlier, he served as Group Vice President of Global Services at Plex Systems and founded MIPRO Consulting, an enterprise applications advisory firm specializing in PeopleSoft and Oracle technologies. His career also includes senior leadership roles at PeopleSoft and Oracle, where he led a North American supply chain management practice of more than 230 consultants, as well as earlier process reengineering work at Ford Motor Company.

BillingPlatform is an enterprise revenue lifecycle management and monetization software company designed to help businesses manage increasingly complex pricing and billing models. Founded in 2012, the company provides a cloud-based platform spanning order capture, subscription and usage-based billing, invoicing, payments, accounts receivable automation, revenue recognition, and financial management. Its technology supports recurring, consumption-based, hybrid, and other configurable pricing models, while its usage mediation capabilities can ingest and transform data such as API calls, transactions, seats, bandwidth, or other consumption events into billable records. BillingPlatform says its systems process more than 50,000 invoices per day and over $4 billion in monthly billings, positioning the platform as financial infrastructure for enterprises moving beyond traditional fixed-price and subscription models.

Your career has taken you from process reengineering at Ford and enterprise applications at PeopleSoft and Oracle to founding MIPRO Consulting and leading customer, revenue and marketing functions at Plex Systems and Conga. How have these experiences shaped your priorities as CEO of BillingPlatform, particularly when introducing AI into mission-critical financial operations?

I consider every stop in my career as mission-critical experiences. The manufacturing process at Ford, financials and supply chain management at PeopleSoft and Oracle, the consulting years between, ERP at Plex, and revenue lifecycle management at Conga. When systems like those encounter errors or a line stops, someone does not get paid, or a deal does not close. That is a much lower tolerance for error than most software operates under.

The other thread is that value gets determined by the outcomes you unlock, not by what you market. Owning the customer and revenue functions at Plex and Conga is where that became obvious to me, sitting in the gap between what got sold and what the customer received.

Both threads land in the same place for AI in financial operations. Being right 99.5% of the time sounds great in a demo, but it is still a failure. We run billing for companies with millions of subscribers. At that scale, a half percent error rate is tens of thousands of wrong invoices in a single month, tens of thousands of customer conversations nobody planned for, and a credit process that buries the team. The customer expected all of them to be right, so 99.5% did not deliver the outcome they bought.

That is why we are conservative about where AI touches money and aggressive about where it touches work. AI does configuration, analysis, and investigation. The calculation stays deterministic. The test I apply to every AI feature is whether a controller can explain the output to an auditor without saying the model decided it for him or her.

Many enterprise software vendors are adding AI assistants to existing products. BillingPlatform describes its AI as native to the revenue lifecycle and built on a unified, self-describing metadata model. What can this architecture accomplish that a conventional AI layer added to legacy billing software cannot?

Our platform describes itself. Entities, fields, pricing rules, workflows and approvals are all metadata the system can read at runtime. That means our AI is reading your actual configuration, not a generic product schema, and it can propose changes to that configuration because configuration is data rather than code.

An assistant bolted onto legacy billing software has a different problem. The customer-specific logic in those systems lives in custom code, stored procedures and integration middleware. The assistant can answer questions about the documentation. Anything specific to how your instance actually behaves has to be read out of that code, and changing it means a development cycle rather than a conversation.

The second issue is that metering, billing and revenue recognition usually sit in separate systems with separate databases. Any AI on top of these functions is reasoning across a reconciliation, so its answer is only as good as last night’s batch job. We run all three areas on one model, so there is a single chain from usage event to journal entry. The AI is answering from the same record the invoice was built from, which is hugely important.

As companies adopt usage-based, outcome-based and hybrid business models, how is AI changing the way products and services are packaged, priced and monetized?

Two things are shifting. The first is that the unit of value is moving away from seats toward work completed, so companies are pricing resolutions, documents processed, tasks finished or outcomes delivered. The second is that pricing is becoming something you revisit quarterly instead of annually, because the cost curve underneath an AI product moves that fast.

In our experience, AI helps on the analysis side more than the creative side. If your transaction history sits in one place, you can model a proposed pricing change against real customer behavior before you ship it in minutes, and forecast how that may impact margin or churn, for example. That used to be a multi-week finance exercise.

I would also add a caution. Most companies moving to consumption or outcome pricing cannot yet answer what a single unit costs them to serve. Without that number, a company is guessing at floors and discount authority. That is why most enterprises land on a hybrid approach, a committed platform fee for predictability, consumption for alignment, and an outcome component where the value is genuinely measurable.

AI products can generate highly variable costs based on tokens, models, infrastructure and the complexity of each request. What new metering and billing capabilities do companies need to monetize AI services accurately without exposing themselves to unpredictable margins?

I would suggest starting with metering at the granularity of the cost driver and keep the events. If you pre-aggregate to a daily total by customer, you have thrown away the ability to see that one workflow is running on an expensive model.

From there, four capabilities do the real work:

1. Attach the cost of goods to the same event that carries revenue, so margin by customer, feature and model is visible during the period instead of at close.
2. Support rate structures that change mid-period, because model substitution and vendor price cuts will not wait for a company’s renewal calendar.
3. Give customers commitments, prepaid balances, caps and drawdowns, since predictability is a product feature and it also protects a company from an unexpected invoice.
4. And finally, be able to re-rate history when a price changes or mediation was wrong, without a manual credit process.

Margin protection on AI products is a metering and pricing design problem. If you treat it as a finance reporting problem, you find out about it 40 days late.

BillingPlatform allows teams to configure products, pricing rules and billing workflows conversationally. Which decisions can safely be delegated to AI, and where should human review and approval remain mandatory?

I would say delegate the work but keep the decisions for humans to make.

AI can safely read and explain existing configuration, draft new configuration in a sandbox, generate test data and test cases, detect anomalies in usage or billing runs, investigate a rating dispute back to the source event, and build a first-pass mapping for a migration. That is a large share of the hours in a billing team’s week and holds almost none of the risk.

Human approval is mandatory on anything that changes what a customer is billed or what revenue gets recognized in production. These include price and rate changes, contract terms, credits and adjustments, revenue recognition policy and standalone selling price judgments, GL mapping, tax positions, and period close. Also, humans need to be involved in anything customer-visible for the first time.

The operating rule is that AI proposes, a person with actual authority approves, and the approval is recorded against the change. Approvals and role-based permissions were already in the platform, so agents inherit them rather than route around them.

Financial systems leave little room for hallucinations or unexplained decisions. How does BillingPlatform combine AI reasoning with deterministic execution to ensure that invoices, revenue calculations and accounting actions remain accurate and reproducible?

We separate interpretation from execution. The model interprets intent and produces configuration, a query or a proposed change. The rating and accounting engine then executes. That engine is the same code path whether a person configured it in the UI or an agent configured it through our AI, so the same inputs produce the same invoice every time, and you can rerun a closed period and get the identical answer.

No number on an invoice is generated by a language model at invoice time. What the AI produces is a configuration change, and that change carries a version, an author and a timestamp like any other change in the system. If an auditor asks why a charge is what it is, the answer traces back through the pricing record to the usage event, not to a prompt.

The last piece is that proposals are reviewable before they are applied. You see the specific configuration the AI wants to create, in plain terms, and you can approve or reject it.

BillingPlatform supports the Model Context Protocol, Agent-to-Agent protocol and connections with external enterprise AI tools. How can companies give autonomous agents access to billing operations while preventing unauthorized actions, data exposure or untraceable changes?

MCP and A2A are transport, and get an agent to the door. The control plane is your permission model, and that is where the actual work sits.

In our case, an agent authenticates as a user and inherits the same role, field-level and record-level permissions any employee would get. There is no agent bypass. Beyond that, a few practices hold up. Give every agent its own scoped service account with least privilege, read-only until proven otherwise, so the audit log tells you which agent did what. Gate writes operations through approval workflow with dollar and volume thresholds. You should also rate-limit them, because an agent in a loop is a different failure than a person making a mistake. It is also critical to log every call, not just the ones that changed something. And return only the fields the task needs, since the fastest way to leak customer data is an over-broad read.

The mistake I expect to see across the market this year is companies handing an agent a human administrator credential because it was easier than scoping one properly.

BillingPlatform says AI can reduce implementations from quarters to weeks. Which parts of requirements gathering, configuration, testing and migration can AI automate today, and which parts still require experienced billing and finance professionals?

AI is genuinely good at translating documented pricing into configuration, running gap analysis against a library of known patterns, generating test cases and executing regression, mapping and cleansing legacy data for migration, and producing documentation. Those are the areas where work is reduced from quarters to weeks.

However, decision-making still requires experienced billing and finance professionals. Some examples that still require experienced people include, getting a customer to agree on their own pricing rules and revenue policy, reconstructing terms from contracts, integrating to a system with undocumented behavior, and judging revenue treatment on a multi-element arrangement. In short, change management does not respond to software.

How can a unified model spanning metering, billing and revenue recognition improve compliance, auditability and financial reporting, particularly for organizations operatingunder complex revenue recognition requirements?

Auditors always inquire about two things: Show me how you arrived at this number and show me you applied the same treatment consistently. If your metering, billing and revenue recognition systems all live separately, answering that means digging through three databases plus a human-built spreadsheet to reconcile them all.

When it is all one model, that whole exercise disappears. The usage event, the charge, the invoice line, and the journal entry are all the same record, just followed through to its next step. You can trace it forward or backward without a reconciliation step in between because there is nothing to reconcile. When a contract changes, you are evaluating the same transaction data that generated the invoice in the first place. Your standalone selling price and allocation will come from what occurred.

This matters most for companies dealing with variable consideration under ASC 606 or IFRS 15, where you must show your work and be able to show it again a year later.

BillingPlatform has recently announced adding a new Chief Financial Officer, Chief Product Officer and Chief Customer Officer as it scales its AI-native strategy. How will these leaders work together to translate technical innovation into financially sustainable growth and measurable outcomes for enterprise customers?

As Chief Product Officer, Rob Zwiebach ensures our product is differentiated and shipped. He ran the product roadmap for financials at Workday and spent 17 years at Oracle, so he knows the systems our buyers already run and what it takes to compete with them. As Chief Customer Officer, Chris King owns whether customers realize the value, meaning time to value, delivery consistency, retention and expansion. He led services and success at Medidata and Salesforce and started in this industry at Zuora, so he has seen both good and bad enterprise delivery at scale. Steven Springsteel, our Chief Financial Officer, owns the economics, and he comes to us from the CFO seat at Recurly, so he has first-hand experience running the finance function of a billing company.

Where they really must work together as one group is on the number I care about most. That is, what it costs and how long it takes to get an enterprise customer live and productive, and the margin on that work. That is a product decision, a delivery decision, and a measurement decision, and it has historically been the place where enterprise billing projects go wrong. Saying you’re AI-native is a claim about your architecture. It only becomes a business when it shows up in implementation cost, expansion rate, and gross margin. Those are the three numbers Rob, Chris, And Steven each own a piece of.

Thank you for the great interview, readers who wish to learn more should visit BillingPlatform.