- Key Insight: Discover how AI-enabled continuous feedback compresses performance cycles from months to days.
- What's at Stake: Slow reviews risk outdated feedback, demotivated talent, and delayed promotion decisions.
- Forward Look: Prepare for HR platforms to require unified people data and AI governance.
- Source: Bullets generated by AI with editorial review
The traditional performance review process can be a challenge for managers and benefits leaders, often taking weeks to complete as managers balance review writing with other administrative responsibilities. As a result, employees frequently receive outdated feedback, and a process intended to improve performance can end up slowing the business down.
Leaders at Remote, a global employment and payroll platform with around 1,800 employees, recognized this problem and overhauled their approach, compressing a two-month review cycle into just 48 hours.
Instead of asking employees and managers to
Reviews are now completed 15 times faster, and leaders record fairer promotion and merit insights, said Madeline Grecek, director of people operations and transformation at Remote.
"The process was broken before it was slow," Grecek said in an email interview. "In our last traditional cycle, about a quarter of the company had self-ratings that didn't match what their manager thought. People genuinely didn't know where they stood."
Read more:
"We'd also doubled in size in a year during a real culture shift. A six-to-eight week cycle doesn't work at that pace, and it wasn't consistent with the no-surprises culture we were trying to build. You can't tell people there will be no surprises and then run a process that doesn't actually enable that."
Grecek recently spoke to Employee Benefits News about Remote's redesigned review process, how managers and employees responded to the change, and the broader implications
How did you maintain review quality while moving so much faster?
We went through every step in the process and asked: Does this actually help a manager make a better decision? If not, it got cut. That's how you move faster without sacrificing quality.
A lot of the pre-work was around calibration. We ran sessions with no individual names, just role levels. What does strong performance look like for this role? What's the baseline versus what pushes someone higher? Once every department had that mapped out, the actual calibration conversations got much shorter because everyone was already on the same page.
We also moved to monthly check-ins, but our mandate was to keep things simple — max three questions. The goal was a pulse check, not a mini performance review. Managers tracked a trending rating and had a short conversation each month. By the time review season rolled around, 70% of managers said those monthly touchpoints made the formal review much easier to complete.
The other thing that helped was integration. People could log notes and feedback throughout the year directly from Slack, and all of it flowed into the review automatically when the cycle opened. Nobody was starting from a blank page or trying to reconstruct months of work.
End result: 95% of employees submitted on time, 97% of managers did, and 85% said the overall experience was good or very good.
How much money or staff time did this actually save?
The first cycle saved over 7,600 hours across the organization, covering employees doing self-reviews, managers completing assessments, and the HR team running calibration. Merit and promotion decisions landed within two working days, where it used to take weeks after the cycle closed.
On the AI piece: 31% of people said they saved one to three hours using it, 13% saved four to eight hours, and some saved even more. Across almost 2,000 people, that adds up quickly.
How did managers and employees respond to the new process initially?
Manager habits were the hardest thing to shift. Getting people to trust that a faster, lighter process could still be rigorous took real effort.
We overcommunicated ahead of the first cycle: all-hands updates, the "people" newsletter, personalized nudges through our internal Slack bot. And then the results spoke for themselves. High completion rates, same-day calibration, outcomes delivered quickly. That converted a lot of skeptics.
The conversations that needed the most care were with people whose ratings went down from one cycle to the next. What worked was making sure all leaders were delivering the same message: the bar keeps moving. What earned a strong rating last year might be the baseline this year, because the business evolves, goals shift, and expectations grow. Once managers felt confident delivering that message consistently, those conversations went much better.
Is this something most employers could realistically replicate?
Yes, but
AI has also changed what's possible here. HR teams can now think and act more like product owners, identifying a problem and building something to solve it without needing a big engineering team or waiting on a vendor. But what you can build is only as good as the data underneath it. If your employment data is fragmented or unreliable, you hit a ceiling pretty quickly. What made it possible for our team to build fast and trust the output is that our global employment infrastructure is solid and accessible — payroll, compliance, people data across countries, all in one place and open to build on. For any HR team thinking about going down this path, getting that foundation right is as important as the AI layer on top.
What role did AI play, and where did humans still make the final decisions?
We used AI to cut admin and help people communicate more clearly. Every rating decision and final performance feedback stayed with the manager.
On day one, employees and managers could pull in notes, feedback and accomplishments from across the year, from Slack, Notion, one-on-ones, and use AI to generate a starting draft. Ninety-one percent of people found that helpful. Most used it to get something on the page and refine from there, not to replace their own thinking.
On day two, live dashboards helped the HR team and leadership spot patterns quickly, places where ratings clustered in ways that warranted a closer look. AI surfaced those, and humans resolved them.
The
What lessons does this offer companies struggling with productivity and workplace efficiency?
Stop
The other thing is to review your process before you touch the tooling. Most performance processes have steps that exist out of habit. Ask which ones actually help a manager make a decision or take action, and cut the rest. Then, and only then, think about where AI fits in. It won't fix a fragmented process, it'll just make the problems move faster.
What advice would you give HR leaders considering a major consolidation of their HR tech stack?
Audit before you shop. Map one end-to-end workflow and find where decisions actually happen and where work gets stuck. The cost is usually in the handoffs and the rework, not in the number of tools.
And ask whether you need to buy something at all. I built a travel safety tool for our team recently using Claude Code and Remote's MCP. It tracks our people in real time, surfaces live government advisories, and automates the outreach and escalation workflow when we need to act. It took about three hours to build and cost $216. We canceled a $60,000 contract because of it.
That's what's possible now for HR teams
