Every transaction already runs through a data room, and the room does none of the work. Conclave automates all manual work and runs the process for your.































Upload in whatever structure the client sends. Conclave proposes the index, catches duplicates and near-duplicates, and flags figures that disagree across documents. It also runs the checklist, so you know what is still missing before a bidder asks for it.



Each counterparty is a company with its own groups and members. Release is per item and per group, cascading down the tree and phased across the process.



Incoming questions are triaged, routed to the documents that hold the answer and drafted with a citation to the source page. Anything that contradicts what another bidder was told is flagged. When the answer is not in the room, Conclave opens a request to the expert who owns it.
The overnight run takes the entire open queue, routes each question to the right documents and drafts an answer with citations. It checks each draft against what every other bidder has already been told. In the morning a coordinator approves, edits or rejects.
Every open question triaged and answered with a citation to the source page. Missing documents requested from the expert who owns them.
A coordinator approves and clears the draft answers' queue in the time it used to take to write one answer.
Every question, answer, release and permission change logged with full provenance.
Drop the documents in whatever shape the client sent them. The agent proposes a numbered index and files every document under it.
Files arrive as final(2), DEF and IMG_4471. The agent reads the content, not the filename, and proposes a name for every file and folder.
Every question is matched to the documents that hold the answer, inside what that bidder group may see, and checked against what other groups were told.
Questions and document requests are matched to the person who answered similar ones before, then sent out with a deadline and an upload slot.
Upload the buyer's information request list as it arrives in Excel. The agent matches every line against the room and turns the gaps into requests.
Most AI tools filter results after the model has already seen them. Conclave filters before. The entitlement check runs ahead of retrieval, so a document a group cannot access is never a candidate answer for that group in the first place.
Entitlements are resolved before the search runs. A restricted document cannot surface in an answer, a summary or an agent query, because it was never in the candidate set.
Every query, draft, approval, release and permission change is logged with its author, its timestamp and its source. The record is the disclosure history of the deal, exportable at close.
Hosted in the European Union. Your documents are not used to train models. GDPR compliant with ISO 27001 certification is in progress.
Unlimited documents, unlimited users, unlimited bidder groups, unlimited AI agents. Indexing, permissioning and Q&A drafting with citations are included, not add-ons. Billed monthly, sold in three, six and twelve month packages.
Unlimited rooms and unlimited deals across the firm, one licence. For advisers running more than eight processes a year. Priced on the size of the team.
Legacy rooms bill per page and per user, which is how a single process reaches five figures. Conclave is a flat monthly fee with nothing metered.
Every fund and adviser is deploying its own diligence agents. Automating the reading is close to solved. The unsolved problem is where those agents are allowed to work, on documents neither side can afford to expose.
Conclave was founded by two childhood friends, a banker who ran processes at Credit Suisse and Houlihan Lokey and an engineer whose last company was acquired by ESRI. It is backed by the Borromeo Family Office and ZEST S.p.A., and it is already live on European transactions.
Bankers and engineers, on the same side of the table.
No. Every draft is reviewed by a coordinator on your team before it is released. The room proposes, a human decides.
No. The entitlement check runs before retrieval, so restricted documents are not candidates for that group’s queries. This holds for the assistant, the Q&A drafts and any external agent connected to the room.
Yes, if the seller allows it. Buy-side teams can connect their own diligence agents. The agent operates inside the permissions granted to its group and every query it makes is logged.
In the European Union. Documents are not used to train models. ISO 27001 certification is in progress and we can share the current control set on request.
The same day. Upload the documents in whatever structure you have and the index is proposed for you.
You export the full record, every document, question, answer, release and permission change, with timestamps. That record is what an insurer or an auditor asks for later.
No. The room mirrors the way a process already runs, with companies, groups, phased release and a Q&A coordinator.
Boutique and mid-market advisory firms running sell-side processes, and the funds and corporates on the other side of them.
Twenty minutes, your own documents, a live process or a closed one.