“Do we have to migrate anything?”
Nothing moves. Nothing gets re-filed.
Cogneros is an intelligence layer over the repositories you already run — originals stay where they are, under the permissions they already carry.
Engagements typically start at $35,000, and scale with what you connect and where it runs.
You are not replacing anything
Cogneros does not replace a document management system, and it does not ask anyone to move or re-file a document. Your approved repository stays the canonical source. Cogneros builds the retrievable layer over what is already in it, and every answer points back to the document in place.
This matters more than it sounds. The projects that fail in this category are usually the ones that begin with a migration — because a migration is a year of work that produces no answers, and because the moment there are two copies of a document, one of them is wrong and nobody knows which.
Why the AI feature in your repository is solving a different problem
Most repositories now ship some form of AI search. It is genuinely useful, and it is answering a narrower question: which documents contain these words.
What a firm needs underneath that is retrieval that understands the unit of work — a matter, an engagement, a project — that knows who is walled off from what, and that shows which document and which page an answer came from, so the answer can be checked before anyone relies on it. That is a different job. It also has to reach across every repository rather than the one it ships inside.
The hard part is entitlement, not ingestion
Ingestion is a solved problem. The part that sinks these projects is permission: a retrieval system that can read everything becomes, the first time someone asks a pointed question, a way around every wall the firm maintains.
Permissions on the source documents travel with them. The rule is that Cogneros never surfaces to a person something they could not already open themselves. Where walls exist they are mapped during the architecture phase, before anything is indexed, and then tested against questions whose answers are already known.
More than one repository, which is the normal case
Institutional knowledge is rarely in one place. It is in the document system, and on shared drives, and in ten years of correspondence, and in the folder a practice group quietly keeps for itself.
Cogneros indexes across them rather than asking you to consolidate first. Consolidation is usually a good idea and it is a separate project; it should not be the price of admission for being able to ask a question.
What has to be true on your side
Documents live in systems rather than only on individual machines. The permissions on them mean something — if everyone can already open everything, retrieval will reflect that faithfully, which is rarely what anyone wants to discover afterwards. And someone can say which material is authoritative and which is superseded: not for everything, but for the categories you want answered first.
None of this has to be perfect. It has to be true enough in the areas you start with, which is why the first phase decides which questions are worth answering before anything gets built.
How to decide
Four situations and what each one points at. If none of them describes yours, the answer is a conversation rather than a page.
- One well-run system, with permissions that mean something
- The fastest possible start. Architecture is mostly confirmation of what you already run.
- Several repositories, inconsistent structure, years of correspondence
- The normal case, and the one with the most to gain. Indexing reaches across them.
- Permissions are nominal and everyone can open everything
- Worth deciding deliberately first. Retrieval mirrors what it finds, faithfully.
- Material lives on individual machines and in email
- Get it into a system first. There is nothing yet to build a layer over.
Questions this raises
The ones that come next in nearly every evaluation, answered plainly.
Which document management systems do you connect to?
Repositories are connected during the engagement based on what you actually run. We describe them by what they are rather than by vendor name, because naming a product in marketing implies a certified connector, and that is a claim about a roadmap rather than a description of your environment. What connects, and how, is settled in the architecture phase and written down before anything is indexed.
Does it copy our documents?
An index is built from them; the originals stay canonical and stay where they are. Where that index physically lives is the deployment decision — a private cloud environment, hardware in your own rack, or a split between the two that you define.
What happens when a document changes?
Ingestion is continuous and managed as part of the service, so the layer tracks the repository rather than freezing at the moment it was built. A precedent bank that has to be curated by hand is exactly the thing this is meant to replace.
Will it surface a superseded version?
Answers name the document and its date, so currency is visible rather than assumed — the first reason citations are not optional here. Beyond that, which material is authoritative and which is superseded is a decision made with you during the build, for the categories where getting it wrong would matter.
The other questions
These three come up together, in roughly this order, in nearly every evaluation.
On your own servers
AI that runs inside your own network.
Where the reasoning happens is a decision you make rather than a default we set — private cloud, hybrid, or an appliance in your own rack.
Read moreConfidentiality
Three questions wearing one word.
“Is it confidential?” usually means three things at once: what trains on our material, who inside the firm can reach it, and what we can show afterwards.
Read moreCompare
Four ways to put AI into a firm.
Public assistants, point tools, private deployment, or building it yourself — what each is good at, and how to tell which one your situation calls for.
See the comparison
Every one of these is settled in writing during evaluation, before anything is indexed — so what your counsel reviews is an architecture rather than a summary of a sales conversation.
Your business has already built the knowledge.
Cogneros makes it available to the people who need it — under your permissions, traced to your own documents.
What an engagement costs
Engagements typically start at $35,000.
Scope follows the repositories you connect and where the system runs — cloud, hybrid, and on-premise are priced differently.
- Prefer email?
- hello@theravengroup.com
- Prefer to talk?
- +1 303-351-1691
- Based in
- Denver, Colorado