Four teams.Four ways in.
One walkthrough for each kind of team we talk to — the situation, the documents genuinely involved, the workflow step by step, and the objection someone in that seat actually raises.
These are illustrative internal examples, not customer results. The portfolios, timings, and comparisons are hypothetical assumptions. We have not published measured outcomes or performance guarantees, and we would rather say so than invent a percentage. When customers are willing to be named and quoted, this page will say that instead.
A CAM reconciliation that has to survive tenant challenge
Six neighborhood centers, roughly 180 tenants
The situation
Reconciliations go out in Q1. Every year a handful of tenants challenge their share, and answering a challenge means finding the lease, finding the amendment that changed the pro-rata formula, and reconciling both against the general ledger. The work is not hard. It is that it is spread across three systems and nobody remembers which tenants have a cap.
Documents involved
- Executed leases and every amendment
- Prior-year reconciliation statements
- Operating expense ledger export
- Tenant correspondence about past disputes
Sevrel searches the lease set and returns candidates with the clause text and the document each came from. Retrieval can miss documents that are unindexed or unreadable, so this is a starting list to verify, not a final answer.
Pull the pro-rata share from the lease alongside the figure in last year's statement, with both sources attached to the comparison.
Ask for a summary by tenant as an Excel file. The output is a real downloadable file, not a chat message you have to retype into a spreadsheet.
The objection we get
“If it gets a pro-rata share wrong and we bill on it, that is our problem, not the software's.”
Correct, which is why every figure links to the clause it came from and why we say plainly that material terms need verification before billing. The value is not that review disappears; it is that review starts from the right paragraph instead of from a filing cabinet.
A deal room with 340 documents and eleven days
Industrial portfolio acquisition, four buildings
The situation
The data room opens and the diligence window is short. The team needs to know what is in the estoppels, whether any service contracts survive closing, and whether any lease has a right of first refusal nobody flagged. In practice, the analyst who reads fastest becomes the bottleneck for the entire deal.
Documents involved
- Tenant estoppel certificates
- Service and vendor contracts
- Environmental reports
- Rent roll and lease abstracts from the seller
Rights of first refusal, assignment restrictions, contracts that survive closing. Each answer arrives with the document and page behind it, so a second reader can check it in seconds rather than re-reading the file.
Seller-provided abstracts are a summary of the leases, not the leases. Comparing them surfaces the discrepancies worth asking about before closing.
Queries and the documents retrieved are written to an append-only audit log enforced by database triggers. When someone asks in eighteen months what diligence was performed, there is a record to point at.
The objection we get
“We cannot put a confidential deal room into a chatbot.”
Reasonable, and it is the right instinct about consumer AI accounts. Sevrel is deployed to your organization with roles your admins set, isolation enforced at the database, and a signed DPA available. The comparison worth making is not against no AI; it is against the analyst who already pasted a lease into a personal account.
Answering owner questions across portfolios you do not own
Managing 40+ properties for nine separate ownership groups
The situation
Nine owners, nine reporting formats, and a standing expectation that any question gets answered same-day. The hard constraint is not speed. It is that an answer for one owner must never be assembled from another owner's documents, and the person answering at 6pm is not always the person who knows the boundary.
Documents involved
- Management agreements per ownership group
- Property-level leases and rent rolls
- Monthly owner reporting packages
- Work order and maintenance history
Each ownership group is a separate scope with isolation enforced at the database rather than by convention. The person answering at 6pm cannot cross the boundary by accident, because the boundary is not a habit.
Occupancy, expiring leases, and expense questions answered with the source attached, in that owner's scope only.
Generated Excel and PDF output per owner, reviewed before it goes out.
The objection we get
“Our owners will ask whether their data is mixed in with a competitor's.”
They should ask, and it is a question a general AI connector answers badly. Isolation between organizations is enforced at the data layer, and our security page describes the control and its current verification status rather than claiming more than we can evidence.
The critical date nobody was watching
Mixed-use portfolio, small team, no dedicated lease administrator
The situation
Lease administration is somebody's fourth priority. Renewal options, termination rights, and rent escalations live in a spreadsheet that was accurate when it was built. The failure mode is not a wrong answer; it is a date that passes without anyone looking at the spreadsheet.
Documents involved
- Executed leases with option and escalation language
- The existing critical-date tracker
- Historical rent invoices
Recorded lease end dates, option windows, and escalation dates read from the documents, so the tracker can be checked against its own source.
Scheduled lease-expiry and rent-anomaly agents run on their own and write findings into your assistant conversation. Agents are read-only by design: they surface, and never send an email or book anything on your behalf.
What changed, what is approaching, and what looks off, assembled before the day starts.
The objection we get
“We are too small for enterprise software.”
Small teams are exactly where an unwatched date costs the most, because there is no second person whose job it is to catch it. Pricing is scoped to the portfolio rather than sold by the seat, so the question is the size of what you own, not the size of your team.
Run one of these on your own documents
Pick the walkthrough closest to your situation and we will set it up with your files — ideally on a question you already know the answer to.