Background and Context

Long Beach, California is a port city of about 450,000 residents. The municipal workforce numbers roughly 6,000 employees across close to 350 civil service classifications. The city’s service mix is unusually broad for local government. Open‑water lifeguards, port and airport operations staff, animal control officers, X‑ray technologists and other specialized roles all sit inside the civil service system.

For years the city managed people through two separate HR authorities. A professional Human Resources Department handled day‑to‑day operations. A stand‑alone Civil Service Commission, operating under formal public‑meeting rules, reviewed and approved a long list of personnel actions. Even routine changes to minimum qualifications or titles had to move through that public process. Hiring stretched. Bringing a new employee on board could take nine months from posting to start.

City leadership moved to streamline the structure. Voters approved a November 2024 ballot measure that merged the Civil Service Commission into HR. The merger became effective in April 2025. The change consolidated accountability for classification and compensation work in one chain of command and reduced procedural drag.

Key players in the modernization push included Human Resources Director Joe Ambrosini; Deputy Director Omar, who oversaw administration and the Classification and Compensation (Class and Comp) unit; and Brett, the lead class‑and‑comp analyst. Brett later brought in a second analyst, Marcelo, and served as Holly’s primary point of contact inside the city.

Challenges

At the analyst level, Brett and Marcelo had to coordinate with subject matter experts across many operating departments. Together they built and updated class specifications, minimum requirements, duty statements, and knowledge‑skills‑abilities sections. Department partners did not always understand the format or value of the class and comp system. Inputs arrived incomplete. Analysts spent hours revising, clarifying and coaching. No single analyst could possibly master the full range of municipal occupations.

Long Beach benchmarks compensation against a peer set of 10 to 11 jurisdictions. Not every peer is relevant to every job. The city operates a fireboat; inland comparators do not. Selecting the right comparators job by job took time and judgment.

Leadership also saw heavy analyst effort going into data gathering that was necessary for documentation but low impact relative to other work. Staff hunted for comparator data to support decisions and then tried to translate dense class specifications into job bulletins that would appeal to candidates. Both steps consumed scarce capacity.

Paperwork added friction. Subject matter experts were sometimes asked to complete a 14‑page Position Description Questionnaire (PDQ). Analysts then condensed that material into a two‑page class specification. The workflow encouraged information bloat and uneven quality in the first draft.

All this took place against a slow hiring backdrop. With nine‑month timelines, top candidates often accepted faster offers from other public or private employers before Long Beach reached interviews.

Ramifications

Classification uncertainty rippled into labor relations. During meet‑and‑confer, weak or disputed class spec data could stall negotiations for weeks or months. In some cycles the city commissioned outside consultants to rework a specification. One‑off studies could add as much as $10,000 to a bargaining round and push decisions further out.

Extended timelines hurt recruitment. High quality candidates drifted away. Remaining pools were thinner and harder to assess. Analysts lost time they needed for higher value work, including misclassification reviews and pay equity analysis. Document control deteriorated as redlined Word files moved by email. Version confusion raised compliance risk and slowed approvals.

Summary of Solution

Brendan Hellweg, Holly’s founder, explains the product succinctly: “Holly is an AI driven platform that helps local governments create more effective evidence based jobs.”

Holly ingests a jurisdiction’s existing classification and compensation data. It also pulls in the salary and class specification data from the comparator jurisdictions that HR teams already use. The system applies AI to map relationships across class specs and produces similarity scores based on duties. What once required slow manual comparator selection can happen in minutes and at scale.

With those mappings in place, Holly can generate improved class specifications and position descriptions. It can flag potential issues raised by legislation or accessibility rules. It also provides supporting evidence, showing which comparator data informed a recommendation.

The Long Beach engagement began as a six‑month pilot to analyze and produce updated drafts for more than 350 class specifications and to demonstrate the broader analytics capability. Alternatives looked expensive or slow. A consultant‑led rewrite of similar scope would typically cost in the range of three hundred thousand to three hundred fifty thousand dollars. It would likely take a year and result in a static report. As Hellweg put it, “If it’s just a PDF, the work is good as of the day it is delivered and then immediately starts to decay in quality.” Doing the work entirely in‑house would have meant reviewing 20 to 30 positions a year at best, given normal six‑to‑eight week cycles that can stretch to six months on hard jobs.

Outcomes Achieved

Through the pilot, Long Beach received updated drafts for roughly 350 class specifications. Each draft drew on peer jurisdiction data and documented the sources behind recommended changes.

Analyst workload shifted. Brett reported that a class spec update that once required two to four SME meetings usually required only one when Holly produced the first draft. Time gained flowed to misclassification complaints, pay equity reviews and other pressing work.

Subject matter experts who had struggled to translate operational knowledge into HR language now reacted to high quality drafts instead of facing a 14‑page PDQ. Feedback from the field was positive. Working with Long Beach also led Holly to build an analytic that tests whether a driver’s license requirement is legally justified by the job’s essential duties. The need surfaced as new California law took effect in January 2025.

Centralizing redlines and approvals in the platform created a clearer source of truth. Staff could see the current version, track changes and reduce the confusion that email attachments had created.

Future of the Partnership

Long Beach was Holly’s first customer and an early design partner. City feedback has shaped product decisions from compliance analytics to workflow features that keep reviews on platform instead of in offline Word batches.

As the post‑merger HR organization stabilizes, the focus is shifting from one‑time mass updates to daily use. Areas under discussion include additional rounds of classification modernization, support for labor costing, and conversion of improved class specs into more candidate‑friendly recruitment content. Long Beach leaders have expressed interest in deepening this operational integration in the coming year.