Make the business answerable without you — before a buyer asks whether it is.
What you and a few long-tenured people carry in your heads already sits, in pieces, in your files, mailboxes, chats, ERP and CRM. Knowledge Engine gathers it into one private knowledge base that stays with the company, where every answer cites its source. Your managers stop asking you. A buyer’s team can check the answers for themselves.
What is Knowledge Engine for a founder planning an exit? It is a private knowledge base that runs inside your company, on a box in your building or a server in your own cloud account. It reads your file servers, e-mail, chats, ERP and CRM, organizes what it finds, and lets managers, new hires and AI assistants ask questions in plain language. Every answer cites the exact source passage, shows whether that source is current, and is limited to what the person asking may see. For an owner-dependent business, that means the pricing logic, customer exceptions, supplier history and unwritten procedures stop living in one or two heads and become a documented company asset. Olyra installs it and operates it as a managed service that keeps it current, secure and supported. The data and the keys belong to the company and stay with it through a change of ownership.
The business works. It works because you are in it.
None of this is a failure of management. It is what happens when a company grows for twenty years around the people who built it. It becomes a problem on the day someone else has to run it.
“Ask Marta.”
Half the procedures in the building are a person’s name. Why that customer gets sixty-day terms, which supplier to call when the usual one is late, how the month-end really closes. It is all known, and almost none of it is written down.
“I am the pricing system.”
The price list is in the ERP. The logic behind the exceptions is in your head and in ten years of e-mail threads that only you know how to find.
“They will want me to stay for two years.”
A handover period that is really you answering questions. Transition terms and earn-outs are often tied to the founder staying, because nobody else can explain how things work.
“If two people resign, we are in trouble.”
Long-tenured staff carry customer history, the reasons behind decisions and the workarounds. When they leave, it leaves with them, and the timing is rarely yours to choose.
“The data room is going to be a nightmare.”
Contracts across shared drives, amendments in mailboxes, three versions of every policy and no way to say which is current. Each diligence question starts a manual search.
“I do not have time to write manuals.”
You are still running the company. A documentation project that depends on you and your best people sitting down to write is a project that does not get finished.
Buyers and their advisers look hard at owner-dependence and key-person risk, because they are buying a company that has to work on the first Monday without you. The more of the answer that sits in your head, the more of the deal depends on you staying in the building.
From “ask the founder” to “ask the company”
Knowledge Engine does not ask you to write the manual. It starts from the record your business has already produced and makes it answerable.
| Today | With Knowledge Engine in place |
|---|---|
| A manager walks into your office to ask why a customer has special terms | The manager asks the knowledge base and gets the answer with the contract clause, the e-mail that agreed it and the ERP record beside it |
| A procedure exists as “how Carlos does it” | SOPs, policies and procedures are read as versioned, cited steps with roles, systems, thresholds and approval points |
| Nobody is sure which price list, policy or contract version is current | Each document is marked verified, stale, superseded or disputed, and answers say so when they lean on an old one |
| The reasoning behind a decision is remembered by whoever was in the room | The reasoning is recorded once as a curated answer or a precedent, approved by a steward, and found by everyone after that |
| A new manager takes a year to learn who to ask | A new manager asks in plain language from the first week, within the permissions you set |
| A diligence question means days of digging through folders and mailboxes | The question is answered with citations, and the supporting documents can be collected and exported with a record of what was produced |
The product is the same Olyra Knowledge Engine described on the main page. This page covers what it does for a founder who wants the company to stand on its own before a sale.
Six capabilities, each chosen for an owner-dependent business
Every item below answers one of the problems above.
It reads what already exists
For “I do not have time to write manuals”
- Connects to file servers and shared drives, Microsoft 365 or Google Workspace, your ERP and your CRM
- Reads documents, mail, chats, spreadsheets and database records, including scanned PDFs through built-in OCR in the languages your documents use
- Shows ERP customers, orders and invoices next to the e-mails and contracts about them
- The first pass needs your access approvals and your questions, not your writing time
Every answer shows its source
For “how do we know that is right?”
- Citations point to the document, the version and the exact location
- Conflicting numbers, dates or revisions between sources are flagged, not hidden
- When the evidence is not there, it says so instead of guessing
- A successor or a buyer’s analyst can open the source and check it
The ‘why’ is recorded once
For “I am the pricing system”
- Curated answers approved by a steward take precedence for the questions that matter most
- Precedents capture how a case was handled and why, once a steward approves them
- Record the reason a customer has special terms a single time; it is found from then on
- Corrections your staff make to earlier work become visible to the next person and the next AI agent
Procedures become cited steps
For “ask Marta”
- SOPs and policies are read as versioned steps with roles, systems, thresholds and approval points
- Stewards are told when tickets or mail show a process running without its approval step — documents and counts, never people
- A “who owns this?” hint routes an unanswered question to the owning role
- Unanswered questions become a to-do list of knowledge gaps to close before a sale
The right people see the right things
For “who should be allowed to ask?”
- Permissions are enforced inside the database itself, so no screen or AI assistant can go around them
- Existing file-share and SharePoint restrictions can be carried over
- A new manager, an adviser or a buyer’s reviewer gets an account you approve, scoped to what you allow
- Every sign-in, search, document view and export goes to a tamper-evident audit trail
The knowledge stays with the company
For “what exactly am I handing over?”
- Runs on a box in your building or a server in your own cloud account
- The data, the knowledge base and the keys belong to the company and stay with it
- Olyra operates it as a managed service — updates, security patching, monitoring and support — and holds no copy of your data
- The whole knowledge base can be exported at any time
How the founder gets out of the middle
The aim is not a perfect manual. It is a company where the next question goes to the knowledge base first and to you only when it has to.
Capture what is already written
Years of contracts, quotes, e-mail threads, order history and procedures are read and organized against categories you define: customer, supplier, department, document type, sensitivity. Most of what people ask you about turns out to be on record somewhere.
Fill the gaps that only you can fill
Knowledge-gap reports show the questions people asked that the record could not answer. Those are the ones worth your time. You or a long-tenured colleague answer each once, a steward approves it, and it becomes a curated answer.
Let people and AI agents use it
Managers ask in a web console. AI assistants connect through the open MCP standard under each user’s permissions, and get answers fast enough to use in the middle of a task. Questions that used to reach your desk get answered without you.
You can watch the shift happen. The audit trail and the gap reports show which topics are being answered from the record and which still depend on a person, so you know where the company still leans on you well before a buyer finds out.
What a buyer’s team sees when they look
We will not put words in a buyer’s mouth, and no tool decides how a buyer values a company. What we can describe is what is in front of a reviewer when the knowledge base is in place.
A company that can answer for itself
Operational questions are answered from the company’s own record with citations, by a manager or by the reviewer under a scoped account, without booking time with the founder.
Procedures with versions and owners
How orders are approved, how exceptions are handled, who signs what. Each is shown as cited steps tied to the documents that define them, with the current version marked.
The reasons behind the exceptions
Why a customer has special terms or a supplier is preferred, recorded as curated answers and precedents with the supporting contract or e-mail attached.
Evidence of what is current
Each source is marked verified, stale, superseded or disputed. The knowledge base can also be viewed as it stood on any past date, which helps when a question is about what was known at the time.
Controlled access with a record
Who can see what, and why, for any document or person. What the reviewer was shown is logged, and you can end that access when the review ends.
Knowledge that stays with the company
The data, the knowledge base and the keys belong to the company and sit inside its own building or cloud account, not in a shared cloud service. Olyra’s managed service keeps it current and supported through the handover, so a new owner meets a working, maintained system and not an archive.
Knowledge Engine reduces key-person risk and supports a cleaner handover. It does not promise a valuation, a sale price, deal terms or the outcome of any diligence process. Nothing on this page is legal, tax or financial advice; your M&A, legal and tax advisers remain responsible for those questions.
An orderly data room, built from the record you already have
Diligence is a long list of questions asked against a short deadline. These parts of the product help your team and your advisers answer them in an orderly way.
- Collections and export. The eDiscovery tools open a matter, run collections across sources and produce a numbered package with a chain-of-custody manifest, so you can show what was gathered and from where
- Access reviews. See who can see what, and why, for any document or person, and fix old shared-folder access before outsiders are given accounts
- A tamper-evident audit trail. Sign-ins, searches, document views, assistant calls, label changes and exports are all on the record
- Freshness on every source. Superseded contracts, policies and price lists are marked, so the current version is the one that gets produced
- Search as of a past date. View the knowledge base as it stood on a given day when a question turns on what the company knew then
- Sensitivity and residency rules. Decide by source or by content what may leave the box, and preview what a rule would allow before switching it on
- Retention rules. Keep-for periods by label or source; the system proposes what is due and nothing is erased without approval
- Export at any time. The data and the knowledge base are yours to export whenever you ask, and a factory release issues a signed certificate that records the export
These tools support the work of your advisers. They do not replace a virtual data room chosen by your deal team, and what is disclosed, to whom and when remains a decision for you and your counsel.
How it compares with what founders usually try first
Each of these is a reasonable choice in some situations. Here is how they tend to play out in an owner-dependent business of roughly 50–500 people.
| The alternative | Where it helps | Where it tends to fall short for an exit | What Knowledge Engine does |
|---|---|---|---|
| Write the manuals yourselves (an SOP project, a wiki) | Forces useful conversations; costs no new technology | Depends on the time of the founder and the busiest people; the result starts ageing on the day it is finished and covers only what someone thought to write | Starts from the existing record, marks what is current, and asks people only for the gaps that questions reveal |
| Hire consultants to interview and document | An outside view; produces a structured set of process documents | A snapshot of what people said in interviews, separate from the contracts, mail and ERP records that prove it | Answers come from the source material itself and cite it; the knowledge base keeps updating as the business does |
| Tidy the shared drive and assemble a data room late | Necessary at some point in every sale | Organizes documents but does not explain them; the questions still come back to the founder | Makes the documents answerable in plain language, with the reasoning recorded beside them |
| A copilot bundled with your office suite, or a chat tool over uploaded files | Quick to start; already in tools people use | Centered on one suite or a handful of uploads; follows today’s folder permissions, oversharing included; your content sits in a shared service; little link to the ERP or to an audit trail | Puts ERP and CRM records beside files and mail, enforces permissions in its own database, and runs privately inside the company, with its data under its own control |
| Rely on a long transition period | Needs no preparation now | Keeps you tied to the business after the sale, and leaves the same dependence on any long-tenured person who decides to go | Moves knowledge into a company asset ahead of time, so less of the handover rests on any one person |
Knowledge Engine does not remove the need for a good second-in-command, a clean set of accounts or experienced advisers. It deals with one specific risk: knowledge that exists only in people.
The knowledge stays with the company
The knowledge a successor will rely on has to stay with the company: its data, its knowledge base and its keys. Olyra’s managed service keeps that knowledge current, secure and supported, before and after the handover.
| The company’s | Olyra’s ongoing service |
|---|---|
| The data, the knowledge base and every export, at any time | Setup: installs and hardens the system, connects the sources, and builds the categories and residency rules with your data steward |
| The recovery keys, handed over at go-live | Runs updates in your maintenance window, with automatic rollback if a health check fails |
| The box, or the server in the company’s own cloud account, where it runs | Monitors and security-patches the system, and supports it under an agreement with published response times |
| Its own third-party accounts, including any cloud AI service it chooses to use | Adds connectors and improvements, and delivers evaluation reports on your own questions so you can see answer quality hold |
| The decision to approve every support session and every sensitive change | Works on the box only in sessions you approve, which you can watch, end and review afterwards; holds no copy of your data and hosts no AI inference for you |
AI models are reached at addresses the company controls: its own GPU machine, a cloud service under its own account, or the box itself. Your data trains no one’s models. The knowledge base lives on encrypted storage, and sensitive settings need two people to approve.
The ongoing service is what keeps the knowledge base accurate, current and secure while the business changes hands. And if the business grows after the sale, the same system grows with it: a single box, then an automatic failover pair, a cluster at one site, several offices as one master with replicas, and a disaster-recovery standby. As a point of trust, the data can be exported at any time, and a factory release exports it in full if the company or its new owner ever decides to stop. More detail is on the main Knowledge Engine page and in the Trust Centre.
Start with one source and ten of the questions only you can answer
Ideally you begin 12 to 24 months before an intended sale, so that the knowledge base has a track record of use by the time anyone outside looks at it. A shorter runway still helps; it simply covers less.
Install and connect one source
Sign-in is tied to your identity provider and the single richest source is connected first. For most founders that is the mailbox and file share where customer and supplier history lives, or the ERP.
Your own questions, answered
Ten real questions, the kind your managers bring to you, asked and answered in front of you with citations that open the source.
The first managers stop asking you
A starter set of categories is in place, more sources are connected, and a small group of managers uses the console for daily questions. The first gap report shows what only you can answer.
A measured baseline
An evaluation on a golden set of your own questions, a trained data steward, and a short list of curated answers to record next.
First ingestion is real work: your data is read, extracted, classified and built into a fast database in several formats, which takes time and AI processing. That is why there is no free trial on your own data, and why the engagement starts with one high-value source and shows cited answers on day 3. You can see a demonstration box running on sample data before you commit. Remaining sources, classification review, residency rules and roll-out follow over the next four to six weeks.
In the months before a sale
- Record the reasoning behind your largest customer and supplier exceptions as curated answers
- Work down the knowledge-gap list until routine questions no longer need you
- Have long-tenured staff confirm or correct the procedures in their area, one click at a time
- Run an access review and fix old shared-folder permissions before any outside account is created
- Agree with your advisers what a buyer’s team will be allowed to see, and set that up as a scoped role
If you are not yet sure where knowledge risk sits in your operation, an AI Operations Audit is a sensible first step.
Who this is for, and who it is not for
- Founders and owners of manufacturing, distribution or services businesses with roughly 50–500 people, in Brazil or the US
- Companies where pricing, customer terms and supplier relationships depend on what a few people remember
- Owners who expect to sell, bring in an investor or hand over to a successor within the next few years
- Businesses with years of history spread across file shares, mailboxes, an ERP and a CRM
- Owners who want this run for them, not another system to staff
Who it is not for. If you have no intention of selling or stepping back, and a small team with a handful of documents, a shared folder and a lighter tool will serve you well. If a sale is closing in a few weeks, there is not enough time for the record to build, and your effort is better spent with your advisers on the data room. If the real obstacle is that there is nobody to hand the business to, a knowledge base will not fix that. We will tell you if any of these apply.
What founders ask before they start
Will this increase what a buyer pays for my company?
We cannot promise that, and nobody credibly can. Price and terms depend on the buyer, the market, your financials and your advisers. What Knowledge Engine does is reduce a specific risk that buyers examine: dependence on the owner and on a few key people. It makes the company’s knowledge visible, cited and transferable. Nothing here is legal, tax or financial advice.
I do not have time to document anything. How much of my time does this need?
Far less than a manual-writing project, because it starts from what already exists in your files, mail, chats, ERP and CRM. You are needed for access approvals, for the ten real questions on day 3, and later for the gaps: the questions the record could not answer. You answer each of those once. The company also needs a data steward who can give a few hours a week during the first two months.
Can we try it on our own data for free first?
No. Bringing your data in means reading, extracting and classifying it and building a fast database in several formats, which takes real time and AI processing. We can show you a demonstration box running on sample data, and the engagement itself is built to prove its worth early: one high-value source first, and cited answers to your own questions on day 3.
Much of what I know was never written down. What happens to that?
More is on record than most founders expect, in e-mail threads, quotes, contract amendments and order history. For the rest, the gap reports show which questions people asked and could not get answered. You or a long-tenured colleague record the answer once as a curated answer or a precedent, a steward approves it, and it is found from then on. Knowledge Engine cannot capture what was never recorded and never asked about.
My long-tenured people may see this as replacing them. Is it a way to monitor staff?
No. It tracks documents, not people. Evaluative questions about an individual employee are refused and results for very small groups are withheld. For experienced staff the practical effect is fewer interruptions: the questions they have answered a hundred times get answered from the record.
We want to keep a possible sale confidential. Who can see what?
Permissions are enforced in the database itself, so each person and each AI assistant sees only what they are allowed to see. Sale-related material can be given its own label and limited to a few people. Olyra holds no copy of your data, and a support engineer sees content only in a session you approve.
Can a buyer’s team use it during diligence?
If you and your advisers decide so. You can create scoped accounts limited to the material you choose, every search and document view is logged, and you can end the access when the review ends. Many owners will prefer to keep it internal and use it to answer diligence questions faster. That is a decision for you and your counsel.
What happens to it after the sale?
It stays with the company. The data, the knowledge base and the keys belong to the business, so they go with it, subject to whatever your sale agreement says. Olyra’s managed service can continue for the new owner, keeping the system updated, secure and supported, which is what keeps the knowledge current after you step back. The data remains exportable at any time.
How fast is it, and will it keep up if the company grows?
Your data is prepared in advance, so answers come from what is already built, fast enough to use in the middle of a task. Lookups can return in as little as a couple of milliseconds. Results vary with hardware, data volume and query type. For growth, you add machines: a failover pair, a cluster, or replicas at other offices that send local data to a master site, which ingests it and replicates it back.
Find out how much of your business is still in your head
In a 30-minute briefing we look at where your company’s knowledge lives, which single source would make the best starting point, and what the first thirty days would look like for you. You will also see how the Knowledge Engine answers with citations on a demonstration box.
Find out what AI can actually do for your operations.
Start with an AI Operations Audit — a fixed-scope diagnostic that maps your workflows, scores your AI readiness, and hands you a prioritized 120-day roadmap. If we don't find real opportunities, you'll know that too.
Request an AI Operations Audit → Prefer to talk first? Book a 30-minute scoping call.