How do you give AI your price book, manuals and SOPs?
Written lesson · about 3 minutes to read
The short answer
Gather the current version of each document in one folder, delete the old copies, and connect that folder to an AI tool that searches it and shows the passage it used. Then test it with twenty questions your team really asks, because the tool is only as good as the files behind it.
Three things to remember
- 01Clear out old price lists and duplicate manuals first, or the tool will quote them confidently.
- 02Set who may see what before connecting anything, because the tool can read whatever its account can read.
- 03Test with twenty real questions from your team, and check that every answer names its source document.
What you are setting up
Your price book, your equipment manuals and your standard operating procedures (SOPs, the written steps for how a job is done in your business) hold most of the answers your team asks each other all day. Give an AI tool access to them and someone can type 'what do we charge for a water heater flush?' and get an answer taken from your own file, with the passage shown.
The tool searches your files at the moment of the question, reads the closest passages and answers from those. That is why the documents matter more than the tool: it can only be as right as what it finds.
Get the documents in order first
Gather the current version of each document in one folder. Then take out the old price lists, the three copies of the same manual and the draft nobody finished. Move the old versions somewhere the tool does not look, rather than leaving them beside the new file. A drive with three price lists gets a confident answer from the wrong one.
Decide which file is the single source of truth for each topic. If the price book exists as a spreadsheet and a PDF that disagree, settle that before anyone asks a question. If a procedure only exists in your senior tech's head, write it down now, even roughly.
Put them where the tool can read them
There are three common routes. A project or workspace inside a business AI assistant, where you upload the files and add standing instructions. A knowledge tool built for searching company documents. Or a set-up that searches the folders you already keep in your shared drive. The right one depends on how many files you have, how often they change and how many people will ask.
Whatever you pick, the tool can see whatever its account can see. Connect only the folders it needs, with permissions that match your people, or the newest hire may be told what is in the wage file. Instruct it to answer only from the documents, to name the document and passage it used, and to say it does not know when the answer is missing.
Test it with twenty real questions
Ask your team for the twenty questions they really ask: what a job costs, which part fits this unit, what the steps are for a callback, what the warranty covers. Put each one to the tool and check the answer against the document it names. Add a few questions you know are not in the files. A good set-up says it does not know instead of guessing. Fix the documents or the instructions, then run the twenty again.
Where to start this week
Pick one set of documents, usually the price book or the procedures for one job. Clear out the old copies, put the current ones in one folder and name one person who owns it. Set the tool up for a small group, run the twenty questions and keep a note of every wrong answer. Tell staff to read the passage shown as well as the summary.
Then keep it current. Each time a price or a procedure changes, replace the file the same day and remove the old one. Check that your tool re-reads changed files. Keep customer records and anything private out of the folder unless the plan's data terms clearly allow it.
More in part 5
- 1What data does your AI actually need?
- 2Garbage in, garbage out: cleaning up before you automate
- 3How do you give AI your price book, manuals and SOPs?WrittenYou are here
- 4What is a knowledge base, and does your business need one first?Written
- 5What is RAG?
- 6What is a vector?
- 7What are embeddings?
- 8What is a vector database?
- 9What is semantic search?
- 10What is chunking?
- 11What is re-ranking?
- 12What is AI memory?
- 13What is fine-tuning?
- 14Fine-tuning vs. RAG vs. prompting: which one do you need?
- 15How do you keep your AI's answers up to date?Written
Next step
Where does this fit your business?
Six quick questions about your business, then your AI plan: what to start with, what a person still checks, and the lessons that go with it. Or bring it straight to a free thirty-minute call.
