How Highway Got Off QuickBooks to an AI-Native ERP in 60 Days

Highway, a B2B SaaS company for logistics and transportation, ran on QuickBooks for years before its growing transaction volume and reporting needs outpaced what the platform could do. Ben Hudson, Highway’s controller, sat down with Consero’s Tim Neville, managing director of the AI Fusion Lab, to talk through the move to Rillet, an AI-native ERP, and what’s changed since.

Hudson has run ERP implementations from both the consulting side and the client side, so he came with a clear list of what he wanted and what he wanted to avoid. He and Neville covered the trigger points that pushed Highway off QuickBooks, what a 60-day white-glove implementation looked like, how Highway’s team uses Rillet’s Aura AI to automate collections, and what Hudson would tell an auditor about trusting AI-generated numbers.

Watch the full conversation below, or read the breakdown by topic.

Why Highway Outgrew QuickBooks

QuickBooks worked for Highway early on, but growth exposed its limits. Hudson’s team built manual workarounds just to report effectively, and QuickBooks isn’t structured for SaaS metrics like MRR and ARR: revenue doesn’t exist in the system until an invoice gets generated, which made forward-looking reporting difficult.

Highway also lacked clean department and business unit level views into spend and profitability, and the AI tools bolted onto QuickBooks caused more problems than they solved, including a cash module that kept miscoding transactions as dividends the company didn’t have.

Highway evaluated NetSuite and Sage before its FP&A tool provider, Abicum, referred the team to Rillet. Hudson determined that: either Rillet could deliver on functionality, process improvement, and AI-native capability, or it couldn’t. If it couldn’t, Highway could switch again in a year. The upside on time savings and efficiency was clear enough to make the move.

“Most companies, if they’re fortunate enough to be growing, get to a point where the needs outweigh the functionality.” — Ben Hudson, Controller, Highway

Compliance vs. Strategic Finance: The Case for an AI-Native ERP

Compliance is non-negotiable for any accounting department. Hudson still needs accurate monthly reporting and a clean path through audits. But compliance is also the least interesting function his team performs for the rest of the business. When Hudson evaluated new systems, he weighed how much each one would free his team to work in a forward-looking, real-time way instead of staying stuck in the compliance lane.

“A huge benefit of AI tools generally, but especially in AI-native ERP, is that it allows the finance function to be more strategic within a company, as opposed to simply compliance-focused.” — Ben Hudson, Controller, Highway

Inside Highway’s 60-Day Rillet Implementation

Hudson has been through enough ERP implementations to know how they usually go: the client does nearly all the work while the implementation partner guides the process from the sidelines. Rillet’s rollout, supported by Consero’s Fusion Lab, worked differently.

The team built the historical data upload and reconciled it, and Hudson’s role was to review and spot-check rather than build from scratch. The drain on his team was close to none, and Highway didn’t have to pause its day-to-day work to get there.

Highway went live in 60 days, against a NetSuite quote of a six-month timeline through a third-party consulting firm. The actual project stretched to 90 days because Hudson pushed the timeline a month while managing Highway’s first-year Deloitte audit. The implementation team also delivered a first draft of Highway’s chart of accounts mapping, which Hudson adjusted rather than building from a blank slate, a detail that let him stay tactical instead of buried in setup work.

“The single most important thing, like, all the AI stuff is nice, but if I can’t trust my ledger, then it doesn’t really matter.” — Ben Hudson, Controller, Highway

One Central System: Automating Collections With Aura AI

Within weeks of going live, Highway stopped needing QuickBooks for anything beyond the occasional historical customer note. Consolidating onto one system paid off fastest in collections. Rillet’s native AI tool, Aura, now runs Highway’s entire dunning process. Aura drafts every past-due and end-stage collections email and routes them to Hudson to send.

What used to take his team four to six hours of scrubbing the aging report and drafting emails every Tuesday now takes about 45 minutes, start to finish.

“That’s a process that used to take us four to six hours, you know, scrubbing our aging, drafting all these emails every Tuesday, and now that’s about 45 minutes start to finish.” — Ben Hudson, Controller, Highway

Measuring AI ROI: Time Saved vs. Headcount Avoided

Hudson thinks about AI ROI in two ways:

1. Highest and best use of team time: hours freed up from drafting collections emails go toward higher-value work like customer churn analysis.
2. Headcount avoided, the more directly quantifiable measure.

Based on Highway’s transaction volume and growth trajectory, Hudson estimates he’d need to add two people within six to nine months on a traditional ERP, and a third within two years. His actual plan is to add nobody.

Real-time revenue reconciliation, built around Rillet’s contract and MRR structure plus workflows Hudson built in Claude, has changed how fast Highway closes its books. In August, with both of his team members out of office, Hudson and Highway’s director of FP&A closed the month by the end of business day two, because the reconciliation work had already happened throughout the month rather than piling up at close.

“I’ve got a team of two people underneath me. I don’t see a reason why we would need to add beyond that.” — Ben Hudson, Controller, Highway

Garbage In, Garbage Out: Prompting AI the Right Way

Hudson is candid that AI tools don’t save time automatically. The workflows he built took a couple of weeks to iron out, with trial runs to confirm everything worked before he trusted them. He applies the same discipline to how he prompts AI tools: being specific about the field-level language a system uses (billing address versus shipping address, for example), stating what to exclude, and setting a clear expectation of what the output should look like before asking the question.

It’s also why Highway hasn’t connected its Salesforce instance to Rillet yet. The Salesforce data is still messy enough that when Hudson tested the connection, it started creating duplicate products, so he turned it off.

“The garbage in, garbage-out problem is a real issue, right? The integrity of the data you’re working with is paramount in all of this, because the AI tool can only use what you’ve given it.” — Ben Hudson, Controller, Highway

Learning Opportunities

Hudson would have started the process sooner to cut over on January 1 instead of March 1, giving Highway a full clean fiscal year on the new platform. Beyond that, he stands by how he approached the project: he asked every question that came to mind, pushed back when something didn’t work the way he expected, and treated that as the software provider’s job to answer.

When Highway hit an early limitation around editing future billing schedules on contracts, Rillet’s product team turned around a fix in about two weeks and beta tested it with Hudson before rolling it out to the rest of its customer base.

“There’s no question you shouldn’t ask. If it doesn’t make sense, ask it. It is the onus of the software provider to explain how the software works.” — Ben Hudson, Controller, Highway

Staying Audit-Ready With Real-Time Data

Hudson trusts Rillet’s ledger integrity because he’s pressure tested it, but he’s genuinely curious how auditing an AI-native ERP will play out at scale. His Deloitte team hadn’t yet audited a Rillet client when the two spoke, though they were checking with colleagues who had. Hudson’s ideal state is giving a staff auditor read-only access to Aura to pull information directly, though he expects some early friction as auditors learn to query the system precisely. In the meantime, he’s already thinking about how to use Rillet’s AI tools to generate a first pass at PBC list items ahead of the next audit, on the logic that the faster the data comes together, the faster the audit moves.

“Your data’s never cleaner than it is right then.” — Ben Hudson, Controller, Highway, on the week after signing off on an audit

See What an AI-Native ERP Can Do For You

Consero’s AI Fusion Lab helps finance teams evaluate, plan, and execute a switch to an AI-native ERP, from vendor selection through a white-glove implementation. Talk to our team about what a move like Highway’s could look like for your finance function. It’s a 30-minute conversation, no commitment required.