Playbooks let you encode your negotiation positions once and apply them to every agreement you review. This article covers building and maintaining a playbook. For rule-writing technique, see Writing & Refining Playbook Rules — Best Practices.
Overview
In this article you will learn how to:
Find and manage playbooks from the Playbooks entry in the left navigation
Generate a playbook with AI from your own templates or redlined agreements
Understand the structure of a rule — a Starting Position plus optional fallback positions
Run a playbook against a document and work through the results
Amend a whole playbook by asking Aline AI, and review its proposals as tracked changes
Validate a new playbook before you rely on it
Maintain and version a playbook over time
Where Playbooks Live
Playbooks have their own entry in the left navigation. You can build, edit and review a playbook there without opening a document first.
You can also work with playbooks from inside a document: open the document and click the Playbooks icon in the middle menu bar — the book icon, just under the robot icon — to open the Playbooks panel.
Anyone with access to your Aline workspace can see the playbooks you build; they are shared, not personal.
Step 1: Understand How a Rule Is Built
A rule is a title plus an ordered list of positions:
The top entry is the Starting Position — what you ask for first.
Every entry below it is a fallback position — what you will accept if the starting position is rejected.
Fallback positions are optional. If you generate a playbook from a template you are happy with, you will usually get starting positions only. Fallbacks are most useful when you have real negotiation history to draw on.
💡 When a rule needs review and you accept one of its positions, the sibling positions are greyed out so you cannot also accept them.
Step 2: Generate a Playbook with AI
Aline can build a playbook from documents you already have — a baseline template, or a set of redlined agreements that show the positions your team has actually accepted. There are two entry points:
From Playbooks (left navigation):
Open Playbooks and click the + option.
Choose Create playbook with AI.
Attach the files you want the playbook built from.
From a document:
Open the document and open the Playbooks panel in the editor.
Click Generate a playbook.
Add any additional context in the chat before generating.
From the AI chat: type an @ symbol, scroll down the dropdown to Playbooks and tag an existing playbook or a reference contract, then ask the chat to generate a new playbook from it. Asking the chat to create a playbook with nothing attached produces a generic, industry-standard one.
Aline reads the attached files, checks whether a playbook already exists, and then asks you:
Which party you represent — an entity name ("Aline, Inc.") or a role ("Company", "Customer", "Vendor")
How balanced or aggressive you want the positions to be
Any additional instructions — for example, "do not generate any rules for assignment provisions"
You then get a draft plan to review and refine in chat before you approve it. Generation runs incrementally, and every rule and position is editable afterwards.
📌 Because the AI reads your redlined deals rather than a generic library, the playbook reflects positions your team has genuinely accepted. Pointing it at 10 to 20 redlined versions of the same agreement type produces better fallbacks than pointing it at one clean template.
⚠️ Say up front that you want a flexible playbook that does not over-redline. A generated playbook that is given no steer tends to produce too many rules, and too many rules means a document that comes back marked up far more heavily than you intended.
⚠️ Rules stream in one at a time while generation is running, so the rule count you see partway through is not the final count. Wait for generation to finish, then refresh the page before judging the playbook.
Step 3: Run the Playbook on a Document
Open the document you want reviewed. The quickest route is Documents — drag and drop the agreement straight in.
Open the Playbooks panel and select the playbook.
Click Run at the top right of the playbook.
Run the whole playbook rather than a single category — clicking Run evaluates every rule at once. A playbook you have not run yet shows as Not run; that only means it has not been executed, not that it failed to find anything.
Aline evaluates all positions in each rule and shows which rules passed and which need review, with its reasoning and a citation back to the source text.
📌 If results do not appear, refresh the page before re-running — a finished run occasionally needs a refresh to display.
Step 4: Work Through the Results
Each finding is laid out the same way, top to bottom:
The Starting Position — the rule itself, which is editable from here.
The result — Met, or Not met when the document does not satisfy the rule.
The AI reasoning, and under it the suggested edits.
Working through them:
Use the filters at the top of the results to separate passing rules from rules that need review.
Use Previous and Next to move between findings — the document scrolls to the relevant location automatically.
Accept or reject each suggested addition or redline. Accepting writes the change straight into the document.
If the document meets a fallback rather than the starting position, accept that fallback and the rule is resolved.
📌 Passing and resolved are different states. A rule that needed a redline is not passing — it is resolved, because the redline fixed the issue.
⚠️ A rule fires whether or not the document already addresses it. If you have a rule about a topic the agreement is silent on, the playbook will propose adding a provision for it. Write those rules on an if/then basis to stop it. See Writing & Refining Playbook Rules — Best Practices.
⚠️ Accepting or rejecting a redline does not teach the playbook anything. The playbook only changes when you change its rules — by editing them directly, or by asking the AI to amend them (Step 6).
Step 5: Edit or Re-Run a Single Rule
You can jump straight from a finding to the rule behind it:
Click the three dots next to the rule title, beside Needs review.
Click Edit rule — you are taken to that rule in Playbooks.
Use the pencil icon to edit the Starting Position, or add a fallback position below it.
💡 You do not have to re-run the whole playbook after editing one rule. The same three-dot menu has a re-run option that evaluates just that rule against the document.
⚠️ Editing a rule this way changes it for the whole playbook, not just the document you are looking at. That is usually what you want, but it does mean the change applies to every future run.
Step 6: Amend a Playbook by Asking Aline AI
You no longer have to open each rule to change a playbook. Ask Aline AI to update the playbook — from anywhere in the platform, not only from inside the Playbooks page — and it rewrites the rules for you.
Tell the chat what you want changed, in plain language. For example: this playbook is too strict and over-redlining — change the rules to be more flexible where possible, and update the rules that can be applied on an if/then basis.
You can also hand it a redline from a closed deal and ask it to fold the positions you accepted into the playbook.
Aline proposes every change as a tracked change over the rule, with its reasoning, and starts redlining the rules in front of you.
Accept changes one at a time with the check mark on the right, or click Accept all at the top of the playbook.
⚠️ Nothing takes effect until you accept it. That also means a proposal someone generated and never accepted is still sitting in the playbook as a pending tracked change — and it is easy to miss. If a playbook is not behaving the way you expect, open it and check for unaccepted proposals before you start rewriting rules.
💡 Once you accept, the change applies to every future review.
Step 7: Cover What the Playbook Missed
A generated playbook only knows the clauses that appeared in the documents it was built from. If your standard paper is silent on a topic, the playbook will have no rule for it — and a long third-party agreement will contain terms your playbook never looks at.
The two-step fix:
Run the playbook, then ask the AI chat to review the rest of the agreement that was not covered by the playbook.
Ask the chat what additional rules you should add to the playbook, then add them.
📌 The AI's reasoning can flag a problem even where no rule exists. If the reasoning mentions a clause but no rule covers it, accepting the suggested edit will not remove that clause — you need a rule for it.
Step 8: Validate Before You Rely on It
Generating a playbook is only half the job. Before you put it into production:
Read every generated rule and edit anything that does not match how you actually negotiate. Treat the first run as the one where you work out what needs editing — it will not be perfect.
Run the playbook once against a third party's paper — not your own template — so you can see what it flags on a document it has never seen.
Work in Suggesting Mode so redlines appear as tracked changes. Avoid Editing Mode, which applies changes in final form.
Adjust the rules that over-flagged or under-flagged, then run again.
Step 9: Version and Maintain
Playbooks are not versioned automatically. If you generate a replacement rather than amending the existing one, you will end up with two playbooks of the same name in the list.
Prefer amending the existing playbook (Step 6) so there is only ever one live version.
If you do generate a replacement, delete the old one once you are happy with the new one.
Until you do, use the created date to tell them apart — the list does not show you which is newest.
When Not to Use a Playbook
A playbook evaluates the whole document every time it runs. That is right for a first-pass review, and wrong for a second or third pass: run a playbook against a document that has already come back redlined and it will re-examine every clause rather than focusing on what the counterparty changed.
For reviewing returned redlines, use the AI chat, or build a separate review that looks only at the tracked changes. See Using the AI Chat to Review & Redline Contracts.
File Formats
To have the playbook apply redlines, the document must be a Word (.docx) file. A PDF can still be analysed — you will get flags, reasoning and citations — but Aline cannot write tracked changes into a PDF. Convert the PDF to Word first. See File Format Requirements & PDF-to-Word Conversion.
💡 Best Practices
Generate from redlined agreements, not just a clean template, when you want realistic fallback positions.
Ask for a flexible playbook that does not over-redline — say it at generation time, not after.
Edit every AI-generated rule before the first production run — treat generation as a first draft.
Amend playbooks through the chat and accept the tracked changes, rather than generating a second copy.
Check for unaccepted proposals before concluding a playbook is misbehaving.
Validate against a third party's paper before rolling the playbook out to the team.
Keep one concept per rule; merge duplicate or overlapping rules rather than stacking them.
Remember that runs consume tokens — see Understanding Token Usage in Aline.
📚 Related Articles
❓ Need Help?
Click the Support button in Aline to reach our team.