Work record / Development / 301 W Osborn

301 W Osborn: permit path first, unit count second

301 W Osborn is LÏEF's own project, and I run it directly. I settled the zoning and the permit classification first, and only then let anyone count units.

The seat

I am principal of LÏEF Development and run acquisition, entitlement and pre-development on this project directly. It started in December 2024 and is still going. The work runs through LÏEF Development.

What I was brought in to decide

301 W Osborn, marketed as Lief Midtown, is a four-story office tower from 1971, built in cast-in-place concrete, on a corner parcel in Midtown Phoenix. It has a large subterranean garage. A broker reached out about it and I walked the building. In December 2024 LÏEF Development put it under contract at $3.85M, and we closed in September 2025. The plan: 22 luxury condominiums for sale, each with twelve-foot ceilings, sitting over a garage of 51 stalls.

This is LÏEF's own project. The usual route is a full-service architecture firm and the standard review queue, and I judged that wrong here for three reasons. The parcel's C-1 H-R zoning cast doubt over the entire program. Outside quotes for the same scope were so far apart that the number no longer described the work. And when you convert a building, waiting in the city's plan-review queue is a bigger risk than any single code question.

What we did

• Zoning came first. In April 2025 the city confirmed the zoning, and in May 2025 a written plan-check recap said the same: multifamily by right, with no rezone, no variance and no hearing. Design had not started yet.
• I fixed the permit classification before the unit count. Under the IEBC it is a Level 3 alteration of an existing building, and the occupancy changes from Group B office to Group R-2 residential. Phoenix built its self-certification program for exactly that case. A registered architect of record certifies the set, and the permit comes out in one to five business days, bypassing the standard queue. After that, we locked the program. Each of the two lower floors holds six residences and each of the two upper floors holds five, and they range from 871 to 2,101 SF.
• The city's October 2025 site-plan comments raised fire access, so we filed that appeal in December 2025, months ahead of the certification packet. That kept it from stalling the one step the whole path depends on. We built the 83-page packet with fill, page and assembly scripts instead of by hand, which took days where it would have taken weeks.
• The architect of record still does the stamp and the certification. The Revit model we built ourselves. Fourteen specialist skills run through a 138-tool connection into Revit, and the model grew to 12,684 elements. The agentic tools did roughly 40 percent of the architectural work. Outside quotes for that scope ran $300,000 to $400,000. Our in-house direct spend was about $12,000, a gap of roughly $150,000 before counting the team's own hours.
• We did not skip the quality check an outside firm would have supplied. We rebuilt it. The model's geometry is exported, and agents check it against code parameters and the project's design criteria. Then a person clears each flagged item. It caught two errors, a wrong building setback among them, that a lone human reviewer would have had to find unaided. An earlier version of the model had picked up geometry from an unrelated building's set, and we ruled it out before it shipped.
• A July 2026 project meeting showed that no complete code study had ever been finished. The next day I commissioned an independent owner-side one. It ran 89 pages, with every finding tied to a code section and a sheet, and the critical findings went to the architect already cleared. They included another parcel's address in the plan set's title block, an exit-stair separation that looked short of the code distance, a construction type stated three inconsistent ways, and an elevator cab below the ambulance-stretcher minimum.
• In June 2026 we handed the architect of record a packet of 182 files in nine sections, and then turned it into a ten-item template to reuse for the next handoff.
• We sold the building's rooftop telecom lease, after I turned down the buyer's opening request, which was an 85-year easement over the whole roof. The grant shrank to the carrier's equipment footprint and closed in April 2026. Proceeds were paid at closing, and the rest of the roof stays ours for the conversion's equipment, amenity space, a future carrier or solar.
• Outside counsel spotted a perpetual 2.5 percent match payment on investor distributions tucked into an architect's profit-sharing note. We never signed that contract.
• We built the project's data room in June 2026 and checked every link by hand. A July completeness audit staged what the room lacked and corrected three claims that had drifted from the record. We audit it again whenever the project moves.

Results

• The city confirmed multifamily by right. There is no rezone, no variance and no hearing.
• The self-certification packet is assembled and ready for the architect of record to certify. On that path a permit issues within one to five business days.
• We replaced the outside pre-development scope with an in-house model and review pipeline. It came in roughly $150,000 under the outside quotes. We kept the stamp and cut nothing that carries liability.
• The rooftop grant is scoped to the equipment footprint and was paid at closing, and we kept the rest of the roof.
• The residential floors run at 89.5 percent efficiency, and projected sellout is $30M+. Entitlement and design work is finished and temporary construction is done. Capital is still being raised.

The lesson

Lock the classification before the unit count. Anything downstream of zoning is a guess until that answer is fixed.

Public record

• LÏEF Development
• Lief Midtown
• Lief Midtown Look Book

Stories from this work

Questions about 301 W Osborn, answered

Firm record

Common Ground keeps its own account of this engagement: 301 W Osborn on the Common Ground wiki.

Back to all projects