Behind the Road Map

AI can help software teams build faster than ever, but speed only matters when you’re solving the right problems. At Flyntlok, that means pairing AI-powered development with something decidedly human: ongoing conversations with the dealers who use the platform every day.

Meet Danielle Karr

Danielle’s four-year-old son, Nico, is sure his mom spends her days repairing excavators, tractors and snowblowers, and she hasn’t tried hard to correct him. It got even better when a dealer introduced her to a hi-rail truck, a vehicle that runs on roads and railroad tracks. Working with dealers has brought her close to the real machines, workflows and people Flyntlok supports.

From financial data to dealership data

Danielle spent much of her 20-year technology career at FactSet, working across product management, engineering, strategy and large volumes of financial data. A dealership generates just as much data, from part numbers and inventory to service, customer and financial records. The opportunity is turning that volume into insight people can act on.

Joining Flyntlok at an inflection point for AI

Soon after she joined, Danielle and several Flyntlok developers attended The Gauntlet, a two-week AI training program hosted by Main Street Partners, to learn how to put the newest AI tools into both the development process and the product.

What AI-powered development looks like

The team built an internal tool that takes an incoming support ticket, looks through Flyntlok’s code and the customer’s setup, and hands the developer its findings. Work that could take hours of investigation now starts with far more context. The goal is to bring the same kind of AI efficiency into the product for dealers.

When building gets easier, choosing what to build gets harder

Teams used to ask whether they could build something. As AI speeds up execution, the harder question is whether it’s the right thing to build. Dealers don’t need an endless stream of features just because they’re possible.

Six product areas, each with dedicated focus

Development is organized around six areas: Parts and Purchasing, Accounting and Finance, Rentals and Customer-Facing Portals, Sales and CRM, Service, and Mobile. Small groups of developers own each one and are expected to understand the dealer problems behind it.

Why direct dealer access matters

Before joining, one of Danielle’s biggest questions was whether she’d have real access to customers. For her, that was a requirement. At FactSet, customers who used the software every day and gave detailed feedback made the product team better.

The Rule of Seven and the Five Whys

Listening doesn’t mean building every request exactly as written, especially when dealers ask for different or conflicting things. The team aims to talk with at least seven customers on a topic, and keeps asking why (who does the work, where the handoff fails, what happens next) until the real operational problem shows up.

July office hours put the model into practice

Flyntlok’s first big round of office hours across the six areas drew 142 registrations from 63 users at 32 customer organizations. Dealers compared workflows, challenged assumptions and showed why a small inconvenience repeated hundreds of times becomes real friction.

From dealer conversation to product: expanding Smart Receive

Smart Receive started with invoices. Dealers pointed out that the packing slip often arrives first and goes to the person putting parts away, not the person paying the bill. That led the team to look at packing slips as their own document type, so parts can reach customers and the service department faster.

Dealer feedback is shaping Flyntlok’s AI support

In the CRM office hours, dealers spent about 18 minutes on a bigger question: how to learn and get more from everything Flyntlok can do. That connects to an AI chat Flyntlok is testing internally to answer how-to questions. It won’t replace the support team, and it won’t reach customers until the team is confident in its answers.

Four ways dealer feedback reaches the road map

Office hours and focus groups, which Flyntlok plans to hold every other month. Patterns in support tickets. Beta programs, especially for AI features. And one-on-one conversations with customers who use a specific area of the product.

Closing the feedback loop

Flyntlok publishes release notes every two weeks. When a change started as a support request, the team aims to update that ticket so the dealer knows it’s been addressed. The team is also exploring in-app announcements matched to each user’s work, like Smart Receive news for the parts team.

AI can speed up execution. It shouldn’t rush the thinking.

Danielle’s biggest misconception about AI-powered development is that AI can build on its own, indefinitely, and still produce quality software. Developers still validate architecture, quality and accuracy. Research, customer conversations, strategy and prioritization stay human, and customer feedback matters more, not less.

What dealers can expect next

The next office hours were planned for September, starting with what happened since July: some ideas already released, others shown as early designs that need another round of feedback. Part of each session stays open, because some of the best conversations in July were the unplanned ones.