AI Knowledge Assistant Answers from your own SOPs, in seconds, with the source.

A private assistant in Slack, Teams or a web page that answers your team’s “how do we…?” questions from your SOPs, price lists, past jobs and policies. Every answer links to its source, it respects who can see what, and when your documents don’t cover something, it says so.

Free · 20 minutes · You talk to Syed · Scoped before anything is built

How it runs

Knowledge Assistant Running

Exceptions reach a person

  1. A staff member asks

    In Slack, Teams or a private web page, in their own words.

  2. The assistant finds the answer

    It searches only the documents that person is allowed to see and replies in short steps.

  3. Your documents stay the source

    Each answer links to the section in Google Drive, SharePoint or Notion it came from.

  4. Exceptions reach a person

    Pricing calls, judgement calls and gaps go to the right manager, with the question attached.

See it work

Watch it handle one.

A scripted walk-through of a typical knowledge assistant conversation — press play.

Simulated demo

AI knowledge assistant in Slack

Northline Heating · #ask-ops Slack · Tue 10:12 am · staff only
  1. Jordan · new tech

    First week here. A customer says her furnace igniter failed 14 months after we installed it. How do I handle a warranty claim?

  2. Searching the documents Jordan can access… 3 matches · field-tech permissions applied SharePoint
  3. AI knowledge assistant

    From SOP-Service-07 §3: 1) Check the install date and serial in the job history. 2) Photograph the old part and rating plate. 3) Submit the warranty claim form on site.

  4. Sources
    • SOP-Service-07 §3 · updated Aug 2026
    • Warranty claim form
    • Parts & labour warranty policy
  5. AI knowledge assistant

    Parts on registered units are covered for 10 years. Our labour warranty is 12 months, so labour on this visit is billable at the standard price-list rate.

  6. Jordan · new tech

    She’s a long-time customer and pretty upset. Can I waive the labour?

  7. AI knowledge assistant

    I can’t approve that. Waiving labour is a pricing exception, and the policy says the service manager decides. I’ve sent Priya the details; her answer will appear here.

  8. Sending the exception to the service manager… Job history, photos and question sent to Priya Slack
  9. Priya · service manager

    Approved: waive labour as a one-time goodwill credit. Note it on the job and on the claim form.

  10. Warranty call Manager approved
    Job
    Igniter replacement · installed 14 months ago
    Parts
    Manufacturer warranty claim
    Labour
    Waived · approved by Priya
    Answer based on
    SOP-Service-07 §3 + warranty policy
  11. Logging the question for the monthly review… Gap noted: no SOP yet for goodwill labour waivers SharePoint

Answered from 3 cited sources in 9 s. The pricing exception went to a manager.

Set this up for my business
What it replaces

Off your plate.

  • Stopping your own work to answer the same question for the fifth time this week
  • New hires guessing, or waiting until the owner is free
  • Digging through shared drives for the current version of a form
  • Policies that live only in one person’s head
How it works

Step by step.

  1. 01

    We gather what your team already knows

    SOPs, price lists, warranty terms, policies and notes from past jobs, plus the questions new hires actually ask, so we can see where the documents are thin.

  2. 02

    We connect it with your permissions intact

    The assistant reads from Google Drive, SharePoint, Notion or wherever your documents live, and mirrors who can see what. A tech never gets answers from the HR folder.

  3. 03

    Staff ask where they already work

    Answers come back as short steps with a link to the exact section and its last-updated date. If the documents don’t support an answer, it says “I don’t know” and names who to ask.

  4. 04

    Gaps turn into better SOPs

    Unanswered and escalated questions are logged. Each month we draft the missing SOPs for your approval and flag documents that are stale or contradict each other.

Works with the tools you already use

  • Slack
  • Microsoft Teams
  • A private web page
  • Google Drive / Google Workspace
  • SharePoint / OneDrive
  • Notion
  • PDFs and scanned manuals
  • Jobber or Clio history (read-only)
  • Anthropic or OpenAI business APIs

Connected through official APIs where available. Names and logos are trademarks of their owners; no affiliation implied.

Guardrails

You stay in control.

It cites its sources and says “I don’t know”

The assistant is labelled as AI, and every answer links to the section it came from. If your documents don’t support an answer, it says so instead of filling the gap.

Access controls carry over

It only searches what each person can already open. Folders you mark as off-limits, such as HR or payroll, are never indexed, and we test this with real accounts from each role before launch.

Your data stays in your accounts

Documents stay in your drives. We use AI providers whose business terms say they don’t train on your data, choose the processing region per vendor and document where everything lives, designed around PIPEDA in Canada.

Health and client-confidential material, handled carefully

For clinics, patient records stay out of the index unless the setup meets PHIPA in Ontario, with us acting as your agent, or HIPAA in the US under a business associate agreement. Law firms supervise its use, in line with law society guidance on AI.

Prices and exceptions go to a person

It can quote your price list, but discounts, warranty exceptions and anything outside policy go to a manager to decide. It never approves money or promises on its own.

In practice

What it looks like.

Example scenario · illustrative, not a client result

A 15-person plumbing and heating company

Ten techs, three office staff and two owners. Most questions about warranties, prices, policies and past jobs land on one owner or the office manager, by text or phone. The assistant takes them in Slack.

About 2–3 hours a week back for the owner and office manager (estimate)

The math
  • Staff questions the assistant answers: ~30 a week
  • 30 questions × ~4 min of the owner’s or office manager’s time (stop, look it up, reply) ≈ 120 min/week ≈ 2.0 h/week
  • Document hunts avoided in the office: ~12 a week × ~5 min ≈ 60 min/week ≈ 1.0 h/week
  • Total ≈ 3.0 h/week
  • Less 20% for overlap and ramp-up ≈ 2.4 h/week

Estimates use typical volumes and include a 20% allowance for overlap and ramp-up. Your audit measures the real numbers.

Why it matters

The same questions, all day, to the same person

In most small teams, one or two people hold the answers. How do we handle a warranty claim? What do we charge for a second visit? Each question takes a few minutes, but they arrive all day, by text, in the hallway and from job sites.

The answers usually exist somewhere. There’s an SOP in a shared drive, a price list in a spreadsheet, a policy in an old email. Nobody can find the current version quickly, so they ask the owner instead. New hires feel it most: they either interrupt someone or guess.

What changes with a knowledge assistant

Staff ask in their own words. The assistant searches the documents that person is allowed to see and replies in short steps, with a link to the exact section. If the documents don’t cover it, it says “I don’t know” rather than inventing a plausible answer.

Some questions need a person, not a document. Discounts, warranty exceptions and anything outside policy go to the right manager with the context attached, and the manager decides. The log of questions also shows which SOPs are missing. Once your content is current, the same approach can power a customer-facing website chatbot.

Where we start

We start with one team, one chat tool and the 30 to 50 questions your people ask most, and test against real questions before anyone relies on it. If your manuals are scanned PDFs or paper forms, document processing makes them readable first. When an answer should lead to an action, like opening a warranty claim in your job system, we connect it with custom workflows that you own.

How we roll it out

Audit. Build. Run.

Every engagement is scoped on a call. You’ll know the cost before anything is built.

  1. 01

    Audit

    • Document inventory: what exists, where it lives, who owns it and what’s out of date
    • The 30–50 questions your team asks most, mapped to the documents that answer them
    • Access map: which roles can see which folders, and what stays out entirely
  2. 02

    Build

    • Assistant connected to your drives, with permissions mirrored and tested by role
    • Slack, Teams or web access, with answers that cite the section and date
    • Escalation routes for pricing, warranty and policy exceptions
  3. 03

    Run & Optimize

    • Monthly review of unanswered and escalated questions
    • Draft SOPs for the gaps, sent to you for approval
    • Plain-English report: questions asked, answered and escalated, plus documents flagged as stale
FAQ

Questions owners ask about Knowledge Assistant

Last updated October 7, 2026

What happens when our SOPs are out of date or contradict each other?

Every answer shows the source’s last-updated date. When two documents disagree, it shows both and flags the conflict to the document owner instead of picking one.

How is this different from staff pasting questions into a public AI chat tool?

A general chat tool doesn’t know your SOPs, so staff paste documents into it, often from personal accounts. The knowledge assistant answers only from your approved documents, respects each person’s access and runs on business terms where your data isn’t used for training.

Can it write the SOPs we don’t have yet?

Yes, as drafts. Unanswered questions are grouped by topic, and we draft the missing SOP for you to edit. Nothing becomes a source until you approve it.

Can techs use the assistant from a phone on a job site?

Yes. Slack, Teams and the private web page all work on a phone, so a tech can ask from the truck. Answers lead with the steps and put the source link underneath.

What does an AI knowledge assistant cost to run?

It depends on how many documents and sources we connect, how many people use it and which chat tools you want. We scope it on a short call and give you a fixed build price and a clear monthly running cost before anything is built.

Let’s map Knowledge Assistant to your workflow.

Bring your call volume, your tools and the part that annoys you most. In 20 minutes you’ll know what it would take — and whether it’s worth doing.

Free · 20 minutes · You talk to Syed · Prefer email? muhammad@sadaat.ca

Talk about Knowledge Assistant