How BambooHR built the release process behind its AI ambitions
BambooHR’s Salesforce team was rebuilt from scratch, and inherited no release process at all: change sets, manual deployments, and changes made straight in production. They chose Gearset as a single platform for their entire DevOps lifecycle. A year on, the team has reclaimed capacity and control. They ship through a governed pipeline with automated code reviews and proactive monitoring, they’ve cut hours of manual work per admin per week, and that same pipeline is how BambooHR deployed Agentforce agents into production safely, and successfully. Their DevOps process with Gearset became the foundation their AI ambitions run on.
A team rebuilt, and a process that didn’t exist
When Amanda Sanchez joined BambooHR four and a half years ago she inherited a big backlog and no defined way of working. “We were using change sets, our developers were using the IDE and CLI to deploy things, and there were even some people just making changes right in production.”
As the newly-formed team grew, the process got even messier. People working on the same metadata overwrote each other’s changes, flows failed in front of users, and bugs only surfaced in huge streams of error emails.
Eric Raymond, BambooHR’s Director of Business Technology, saw another issue when he arrived. “There was no defined Salesforce backup and restore solution. That scared me pretty badly.”
Eric had spent 25 years in IT and knew what good looked like for an Enterprise organization. BambooHR’s own engineering teams ran a mature delivery pipeline for the core product. The Salesforce team deserved the same standard. “The software development lifecycle was broken, or non-existent,” he says, agreeing with Amanda that a solid process, without mistakes happening live in production, was critical to future success.
Choosing one platform for the entire DevOps process
Amanda ran the search for a DevOps platform, discovering Gearset through the Salesforce community, and their implementation consultant who’d rated it highly.
They’d looked at Copado and AutoRABIT, but chose not to continue with them due to lacking functionality. “Copado’s features just weren’t comparable to Gearset,” Eric says. “After the initial discovery call, the comparison made it a pretty easy decision to explore Gearset more deeply.”
For Amanda, the difference was breadth. “The amount of solutions included in Gearset’s platform meant that it had pretty much everything we wanted to put in place. Backup, automated release pipelines, error monitoring.”
Consolidating tooling was the ideal scenario. BambooHR were paying for Own (backup), so rather than justifying two vendors with the same offering, they were able to streamline budgets along with operational efficiency. One vendor accountable for the whole process, rather than time spent chasing support across several.
“Any time we can combine tools into a platform, it just makes sense,” Eric says. “We talk about this with Salesforce all the time, that 360 operational view. Looking at it holistically, Gearset gave us that view for our Salesforce DevOps process”
Gearset worked through an extended proof of concept with the team in order to validate the solution properly. From the moment Gearset went live, the team began to see the benefits right away.
When we launched, we saw tangible results immediately.
Quality at scale
Today, a mixed team of nine contributors work through one pipeline. Everyone builds in their own developer sandbox, deploys to integration for testing, and raises a PR through Gearset’s GitHub integration. Automated code review runs on every change before senior admins release to production.
The team made a deliberate choice about how strict that review should be. “We discussed what level of restriction we wanted early on,” Eric says. “We chose the moderate level because we wanted a little friction, to help clean up the tech debt and teach our admins and developers how to deploy clean code.”
“At first it was slightly annoying for some team members, going in and cleaning up errors before they could push to production,” Eric says. “But as time has gone on, everybody appreciates that we’re releasing much cleaner code, and we’re fixing tech debt that’s been there for years.”
For Amanda, the main impact is gaining trust in their process. “The automated code review has increased my confidence in releases pretty exponentially. We know there’s nothing going to production that will cause an error.”
This has helped level up the team over time – developers see exactly why something has been flagged, with the reasoning laid out, and write better code the next time. The improvements won’t stop there – next, the team plans to tighten the rule set further and use Gearset’s org health dashboard to prioritise the most impactful tech debt work first.
Ultimately now we save a lot of time not trying to figure out every little piece we have to include in the deployment. Each admin is saving at least four hours a week, maybe more now but it’s pretty significant in terms of time savings
From user tickets to real-time alerts
Before Gearset, production problems usually announced themselves through end users, and would take a significant amount of time to diagnose. “We’d get a ticket saying I can’t save this record”, sometimes with a screenshot, sometimes not,” Amanda says. “Then we’d have to work backwards and figure out what was wrong and how to recreate it.”
Now Gearset’s Observability automatically posts to a Slack channel the whole team is in, flagging Flow errors, Apex errors, and org limit warnings as they happen. Amanda triages from there, raising and assigning tickets before most users have noticed anything.
“That’s a big benefit of Gearset’s wider platform,” she says. “We didn’t really know half the time until somebody opened a ticket. Now we see it first.”
The pipeline that makes AI safe
BambooHR moved early on AI, having already launched two Agentforce agents through Gearset: a chat agent and an email agent, both doing real work for the business. “They’re providing a ton of value, doing the work of probably close to 20 people right now” Eric says.
Every version of those agents ships through the same governed pipeline as everything else. As BambooHR moved to the agent script architecture, Gearset’s handling of Agentforce metadata made releasing new agent versions dramatically simpler.
“Gearset has made it a lot easier to deploy new versions of the agents and validate that everything needed is included in a PR,” Amanda says. “Not just the Agentforce metadata, but anything the agent runs in its sub-agents too. It speeds up the process a ton.”
Amanda applies the same discipline to AI-generated changes as human ones. Whatever builds it, every change goes through the governed pipeline and automated code review before it ships. “AI speeds things up, but only if you can validate the output,” she says. “Everything still goes through our process.”
For Eric, the deeper value is governance. AI at BambooHR has to satisfy GRC, legal, and security teams who are rightly wary of AI without guardrails.
“Gearset comes packaged with all the governance checkboxes ticked,” he says. “We know we’re using a platform that has those safeguards in place already. If we’d been told to go and DIY build something ourselves in Claude with all the required governance and security layered in, I don’t think it would’ve been possible.”
I can’t imagine doing these big AI initiatives without Gearset, whether it even would’ve been possible. We could probably have done it, but there would have been a load of problems along the way.
Quality work, shipped quicker
Salesforce underpins how BambooHR sells, supports, and serves its customers, so the team building on it can’t risk changes that break things, and can’t afford to ship slowly either. BambooHR found that with Gearset as a partner, they can deliver high quality change, quickly. Every change travels the same governed path; built in a sandbox, version controlled, reviewed automatically, and monitored in production, with built in checks that protect quality as work moves. Nobody has to remember to run them, and nothing reaches production without passing through them.
That’s what lets a still growing team ship quickly without lowering the bar on what goes live, and it’s the same foundation their AI work now runs on.
“Having Gearset gives me peace of mind as we add to the team,” Amanda says. “There’s a solidified process that new people know exactly how to follow from day one.”
“The value of Gearset is enabling our team to really function as a team, and ultimately release quality work quicker.”
The quality shows up, not only in the team’s output, but internally too. “Our engineering teams that build our core product have had a flawless pipeline for years. We felt like we were lagging behind on the Salesforce business side. Now I feel like we’re up there with them, we’ve made huge strides in the last year with Gearset’s help.”
“Before, we were individual people doing our work and sometimes stepping on each other’s toes,” he says. “Now the team is more cohesive. When there’s a problem, we solve it together. You can’t put a price on that.”
The partnership behind the pipeline
Underpinning all of it is a true partnership between BambooHR and Gearset, beyond just support. Gearset works closely with the team to shape the process BambooHR runs today, is there when something needs troubleshooting, and gives the team a voice in where the platform roadmap goes next.
“We seem to work with the same support people, and they know us now,” Amanda says. “They’ll jump on a Zoom call to walk through why we’re seeing an error.” Her newer team members use the in-app chat directly, which has made onboarding easier. “In terms of customer support, they’re always willing and able to bend over backwards to support us.”
A year on, with everything from automated releases to production AI agents shipping through that governed pipeline, Eric sums up the impact himself. “Gearset isn’t just another vendor that offers a tool, we know plenty of those,” he says.
I’ve worked with a lot of vendors in 25 years, and Gearset has to be near the top, if not the best. It’s a true partnership.
