§ Digital services

How the knowledge system works

What we build, how a project runs from the first visit to the monthly service, and the limits you should know about before you call.

§01

The short version

We find where your working knowledge lives, merge it into one record your business owns, and give your staff a chat window that answers only from that record and cites the entry it used. Staff keep it current by telling it what changed, and a person you choose approves each change. The parts you approve for the public become answers on your website that AI search tools can read.

  1. Gather the sources
  2. Merge and settle conflicts
  3. Staff ask
  4. Staff report changes
  5. Approver signs off
  6. Approved answers go public

§02

The nine parts

You don't need all nine on day one. Most businesses start with the record, the chat window and the update process.

  1. 01

    A private knowledge system

    Discovery starts with finding every place your working knowledge lives and how each one is used: the price sheet one person updates, the vendor list on the shared drive, the email thread everyone searches for. We merge those sources into one record.

    Where sources disagree, we list each conflict and you settle it before anything goes live. Your old files stay untouched until you sign off, and then we archive them. You can export the record to a spreadsheet whenever you like.

  2. 02

    Ask in plain language

    Staff type questions the way they would ask a coworker. Every answer cites the entry it came from, so anyone can check it. If nothing in the record fits, the reply says so and stops.

    Every question is logged. The log tells you what people need most often and where the record has holes.

  3. 03

    Update it by telling it

    When a price, a rule or a contact changes, a staff member tells the system what changed and where they heard it. The system shows the proposed edit and saves who made it, when, and the source.

    The earlier version is kept. The people you name as approvers sign off before the change reaches everyone else.

  4. 04

    Faster training, and knowledge that stays

    A new employee can look things up on day one instead of waiting for the one person who knows. Your experienced people get interrupted less.

    When someone retires or moves on, what they recorded stays in the record with their name on it.

  5. 05

    Get found in AI search

    This is answer engine optimization (AEO). We turn knowledge you approve into question-and-answer pages on your website, one question per page, marked up with structured data so AI tools can read them and credit you.

    It takes months, and no one controls what the AI tools choose to quote. We show you the crawler visits and referrals as they arrive.

  6. 06

    An MCP server for your public knowledge

    MCP (Model Context Protocol) is a standard way for AI assistants to call outside tools. We can publish your public answers as an MCP server, so an assistant can look them up directly when someone asks about your field.

    Only published, approved answers are exposed. The internal record is never connected to it.

  7. 07

    Fewer repeat emails

    The questions customers ask every week go on your site as answers. For incoming email, the system can draft a reply from the record, and a person reads, edits and sends it.

  8. 08

    Websites handled

    We move sites off page builders and expensive hosting, rebuild slow or dated sites, and handle the small updates that pile up.

    Your domain, hosting account and content stay in your name.

  9. 09

    Managed, with numbers

    We monitor the system, back it up daily, and put an uptime target and restore times in the service agreement. Access uses multi-factor sign-in, a VPN where it applies, and the least access each person needs. Professional and cyber liability insurance is in place before any agreement is signed.

    You choose where it runs: on a machine your business owns, on site, or in a cloud account in your company's name.

§03

Engagement shape

  1. 01 · First

    Paid discovery

    We visit, look at how your team finds answers today, collect the sources, and write up what we found and what a first build would cover.

  2. 02 · About nine weeks

    First build

    We merge the sources, settle conflicts with you, set up logins and approvers, and train your staff. The timeline depends mostly on how many sources there are and what shape they are in.

  3. 03 · Ongoing

    Monthly managed service

    We run and monitor the system, back it up, fix problems, and keep improving the record and your public answers. You can export everything at any time.

§04

Limits

  • 1The assistant only knows what is in your record. If an entry is wrong, the answer will be wrong, which is why every answer shows its source.
  • 2AI can misread an entry. Anything that goes to a customer is checked and sent by a person.
  • 3Showing up in AI search takes months, and nobody can promise that a particular tool will quote you.
  • 4We do not publish prices yet. Discovery is quoted after the first call, once we know how many sources you have.

§05

Questions owners ask

What kind of business is this for?
Owner-run companies that run on specialized knowledge: part numbers, pricing rules, procedures and exceptions that sit in a few spreadsheets, a lot of email, and the heads of long-time staff. It fits best when your products or services take explaining, to customers or to new hires. Distributors, contractors, specialty shops and service firms are good examples. It isn't for every business. If your answers are simple and rarely change, we'll tell you so on the first call.
Will you work with our competitors?
Not in your region while we're working together. We take on one company per industry in each region at a time. If we're working with an HVAC contractor in Central Indiana, for example, we won't take on a second one here until that engagement ends. Your industry and area are written into the agreement.
Where does our data live?
On hardware your business owns or in a cloud account in your company's name. You decide during discovery. You can export the whole record to a spreadsheet at any time.
Does the AI learn from our data or share it?
Your record is not used to train public AI models. The system uses AI services whose business terms exclude training on customer data, or a model that runs on your own hardware if you prefer. We go through the options with you during discovery.
What happens when it doesn't know?
It says that nothing in the record fits, shows the closest entries if there are any, and logs the question so the gap can be filled.
How long until it's running?
A first build usually takes about nine weeks after discovery. Getting found in AI search takes longer, usually months.
How much does it cost?
We have not published prices. After a free 30-minute call we can tell you what discovery would cost for your business, and discovery ends with a written quote for the build and the monthly service.

Next step

Book a free 30-minute call, or have us come to you.

We work in person across Central Indiana. Monday to Friday, 9 to 5, and weekends by appointment.

Book a call

Call or text 317-662-3636