On a Tuesday morning in San Francisco, somewhere inside a brokerage you have never heard of, a young analyst opens her 11th spreadsheet of the day. The columns are unlabeled. The roof types are stored as adjectives. A building in Houston is listed twice, once as a warehouse, once as a "mixed-use facility (see notes)." Renewal is in six weeks. This is the problem Archipelago wakes up to.
01 / Who they are nowThe data layer for commercial property risk
Archipelago is not a chatbot, not a model, not an insurance carrier. It is a B2B platform that sits between two unhappy parties - the corporations that own buildings and the underwriters who insure them - and tries to make their data agree. The pitch is small and tidy: clean the schedule, enrich the attributes, surface the gaps, send the submission. The implication is enormous: more than two trillion dollars of insured value, on a single platform, finally legible.
The company runs out of San Francisco with roughly 120 people, a Series B in the bank, and a customer list that includes both the buildings (Prologis, Alexandria, Extra Space, JLL, LVMH) and the brokers who place coverage on them (Gallagher, Alliant, EPIC, WTW, Brown & Brown). For a four-year-old in insurance years, that is unusually balanced. It is also the point.
02 / The problem they sawTrillions of dollars on stale spreadsheets
If you have never sat through a commercial property renewal, here is the short version. Owners maintain a Schedule of Values - a multi-thousand-row record of every building they own, with construction type, occupancy, protection, exposure, and a half dozen other fields the industry collectively calls COPE. Some of those rows were last touched in 2014. Some of them were touched by an intern. All of them get sent, every year, in an Excel attachment, to a broker, who reformats them, sends them to underwriters, who reformat them again.
The data is then used to determine the price of risk on assets worth, in some cases, more than the GDP of small nations. It is, to put it gently, an industry crying out for a database.
Archipelago's founders looked at this and saw two things at once: a workflow problem and a worldview problem. The workflow problem was the spreadsheets. The worldview problem was that the insurance industry had quietly stopped trusting its own data, and was pricing risk against models that were only as good as the inputs. Climate change made that worse. Hurricanes that used to be once-in-a-hundred-year events are showing up on a more flexible schedule.
03 / The founders' betThe encore problem
Hemant Shah is not new here. In 1988, as a Stanford graduate student, he co-founded Risk Management Solutions - RMS - the firm that effectively built modern catastrophe modeling for the global insurance industry. He grew it past 1,200 employees before stepping back. Then he did the unusual thing, which is to look at the same industry twenty-five years later and conclude that the bottleneck had moved.
It was no longer the models. It was the data feeding them.
Shah co-founded Archipelago in 2018 with Roger Bodamer, a serial CTO with stops at Oracle, 10gen (MongoDB) and Upthere. The combination is suggestive: someone who knows the insurance industry to its plumbing, and someone who has built a lot of plumbing. In 2023 Shah handed the CEO seat to Alex Lyashok and moved to Executive Chairman, the kind of succession that tends to read either as a graceful handoff or a slow goodbye. With Archipelago, it reads as the first.
The encore, in one line
RMS told insurers what could go wrong. Archipelago tells them what they are actually insuring.
04 / The productSOV Manager, PreCheck, and a hub
Most enterprise software is named for the thing it replaces. Archipelago's products are named for the people they replace - or, more politely, the people they unburden.
SOV Manager is the headline AI assistant. It ingests messy schedules of values - PDFs, Excel files, broker templates, the spreadsheet your colleague swears is final - and extracts the COPE attributes that underwriters need. It reconciles them against industry hazard data, flags inconsistencies, and produces a clean schedule. The company's pitch number is that a 15-day workflow becomes a 15-minute one. Even halved, it would still be unreasonable.
PreCheck is a quieter assistant aimed at the underwriter side. It runs over a submission, identifies the gaps that will get the submission rejected or down-priced, and suggests fixes before anything leaves the building. Property Data Hub is the broader workspace these tools live inside - the substrate where ingestion, enrichment, lineage and governance meet.
The unsexy genius is in the lineage tracking. When an attribute changes - the roof was steel, now it is composite - Archipelago remembers who changed it, when, why, and from which source. That sounds like accountancy. In an industry where the same building might be modeled by three different carriers with three different exposures, it is a small revolution.
05 / The proofNumbers, names, and a $2 trillion pile
Source: Archipelago disclosures, Series B announcement, BusinessWire. Approximate; growing weekly.
The proof is in the customer list, which is unusual in two ways. First, it includes both sides of the insurance transaction - the owners and the brokers - which means the same dataset has reasons to be trusted by both. Second, several customers are also investors. Prologis Ventures participated in the Series B. Stone Point Capital, a financial-services specialist, came along too.
Funding by round
Investors include Scale Venture Partners (lead), Canaan, Ignition, Zigg Capital, Stone Point, Prologis Ventures.
A short, honest timeline
06 / The missionThe resiliency dividend
Shah has written about something he calls the resiliency dividend - the idea that an owner who can prove their building is hardened, well-maintained, and well-documented should pay less for insurance than a neighbor who cannot. The logic is unobjectionable. The execution requires data nobody had on hand. Archipelago is the bet that, once you do, capital will reward you for it.
This is the line that separates Archipelago from a hundred other workflow tools: it does not want to make insurance faster so much as it wants to make insurance accurate. The faster part is a by-product. The accurate part is the point.
07 / Why it matters tomorrowThe climate balance sheet
The next decade of commercial insurance will be defined by two pressures pulling in opposite directions. Climate-driven losses are rising. Capacity to underwrite them is not, at least not at the same speed. Carriers are exiting whole geographies. Premiums in Florida and California are doing things that would have read as typos a few years ago.
Inside that squeeze, the owners who can present clean, current, defensible data on their portfolios will pay less than the ones who cannot. The brokers who can move submissions faster will win the accounts. The underwriters who can trust what they are looking at will write the business. None of that happens on a stack of unlabeled spreadsheets.
Back to that Tuesday morning. The analyst with eleven spreadsheets has, somewhere on her desktop, a tab open to a tool that ingests them, cleans them, reconciles them, and gives her the afternoon back. She uses it. The renewal goes out on time. The building in Houston is listed once. The pricing reflects what is actually there. This is the part where the company has changed the scene.
Not a revolution. Just an industry, finally, doing its own homework.