> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.dioptra.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Playbooks
Our playbook experience uses agentic tooling to make creating, maintaining, and running playbooks easy.
### Overview Video
# When Should I Build a Playbook?
Playbooks are one of four pathways Vera Engage supports for redlining contracts. Playbooks are the highest-effort option and are worth the investment only when you need repeatable results without per-contract expert oversight. If a lawyer or contracts professional will still review every redline, a lighter-weight pathway will usually get you strong results faster.
Still deciding? See the [Redlining page](/redlineagreements) for the full comparison of pathways and step-by-step guides for each.
# Playbook Building: What to Expect
Building a playbook is a collaborative process between your team and the AI agent. The agent handles the heavy lifting (drafting rules, rewriting positions, and iterating on the rule table). Your subject matter experts (SMEs) provide the source materials, validate the positions, and guide refinements. The five phases below describe what to expect at each stage.
**What happens**
Before gathering materials, decide whether a playbook is the right Vera Engage tool for the task and, if so, how many playbooks you need for the contract type. This upfront thinking prevents rework later.
**Do you need a playbook?**
Playbooks are most appropriate for **high volume contracts** with low reviewer discretion on redline responses, where your goal is **automation** or **delegation** — the same repeatable-results rationale described above under "When Should I Build a Playbook?"
If you need redlines grounded in your context, but those redlines will be reviewed by a lawyer or a contracts professional, other Vera Engage capabilities may give you strong results with less upfront effort and faster results. These include:
* Attaching a human playbook, standard form, or prior redline to the chat and asking it to redline to align with the attached.
* Creating a review skill with redline guidance specific to the contract type and attaching context to that skill.
**Do you need one playbook or several?**
Some contract types have regional, business line, or other variations in terms. Our playbooks support flexible choices (for example, only if the counterparty is in Quebec, agree that the software interface will be in French). However, the more determinations that need to be made at runtime:
* The more time a run may take to process
* The more questions a user will have to answer before the playbook run starts
If a single contract type has many variations, splitting into multiple targeted playbooks often produces faster, cleaner runs than one heavily conditional playbook.
**Your team's role**
A short strategic conversation to confirm the use case and scope the playbook count.
**Estimated effort**
Typically under an hour of discussion.
**What happens**
The agent builds your playbook from the source materials you provide. Quality is significantly more important than quantity. The quality of those materials directly determines the quality of the initial playbook output. This phase is about identifying and assembling the best available sources before the build begins.
**Examples of source materials**
* **Negotiation guides or fallback matrices** - shows your specific requirements and tiered positions (Word, Excel, or PDF; no specific organizational structure is required)
* **Preferred redlines on a counterparty form** - shows how you negotiate away from your standard
* **Your standard form agreement** - establishes your baseline positions
**Your team's role**
Identify the above materials. The more the materials are representative, consistent, and clear, the less refinement will be needed in Phase 5. Even a single redlined document is a strong starting point.
**Estimated effort**
Variable, typically under 30 minutes to locate existing materials. No new drafting is required at this stage. If you don't have an easy way to find example redlines, we recommend asking your top reviewers to save their next 1 or 2 first pass redlines in a central folder.
**What happens**
You create a workspace and add the identified materials. The agent reads your source materials and asks you around half a dozen questions. Then you can walk away while the agent constructs an initial rule table, a structured set of playbook rules covering the key issues for the agreement type. Each rule captures the issue, your preferred position, acceptable fallbacks, and the reasoning behind the position.
**Your team's role**
Workspace set up and the provision of context. The agent does the rest.
**Estimated effort**
| Who | Activity | Time |
| --------------- | ------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- |
| Your team (SME) | Upload and intake conversation with the agent | Under 10 minutes |
| Agent | Reads materials, drafts rule table with positions, fallbacks, and rationale per issue | 30 minutes to a few hours (depending on number of issues, number of documents, document length, and complexity) |
At the end of Phase 3, you will have a complete draft rule table ready for review.
**What happens**
Before testing the playbook against real contracts, it is worth pausing to validate that the positions in the rule table reflect your desired negotiating strategy, not just historical practice. Source materials often capture what your team *has* accepted, not necessarily what they *should* accept going forward. You can use our playbook strategy skill to help you navigate the questions and create materials for other stakeholder groups.
**Example questions to address**
* Do the positions in the rule table reflect current business and legal strategy?
* Do the positions reflect general market practices?
* Have the right internal stakeholders signed off? (Legal, Finance, Procurement, Business)
* Are there positions that need to be updated to reflect recent market shifts, regulatory changes, or internal policy updates?
* Are fallback positions appropriately calibrated, not too aggressive, not too permissive?
* Are there issues missing from the rule table that your team regularly negotiates?
**Your team's role**
A conversational exchange with the agent followed by strategic cross-functional conversations.
**Estimated effort**
Variable, dependent on the current state of your materials and the complexity and nature of your organization. Typically at least a few hours of SME review time, spread across one or two rounds of feedback. Effort scales with the number of rules and the number of stakeholders involved.
**What happens**
The playbook is tested against real contracts, ideally a representative sample of agreements your team actually reviews. The agent applies the playbook rules to each contract and flags issues. Your SMEs review the output, identify misses or misfires, and instruct the agent on corrections. The agent rewrites the affected rules and the cycle repeats until the playbook reaches the desired accuracy level.
**Factors that determine effort**
| Factor | Impact on Effort |
| --------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Desired consistency level | Higher consistency requires more testing rounds. A playbook used for initial triage or to support expert contracting professionals needs less refinement than one used for full automation or non-contracting self-help. For the highest consistency we recommend a 10x10 testing model: 10 different contracts, run 10 times each for validation. |
| Number of rules | More rules means more content to test and validate. A 10-rule playbook takes less time to review than a 100-rule playbook. |
| Average document length | More pages means more content to test and validate. Longer contracts take more time to review per test cycle, increasing the time SMEs spend validating each round. |
| Quality of source materials | Strong source materials (Phase 2) produce a more accurate initial build, reducing the number of refinement rounds needed. |
**Your team's role**
SMEs validate the agent's output on test contracts, identify errors, and instruct the agent on corrections. The agent handles all rewriting; your team's job is to review and direct, not to redraft. Expect two to four rounds of testing for most playbooks.
**Estimated effort**
| Playbook Complexity | Approx. Rules | Estimated SME Time (Testing & Refinement) |
| ------------------- | ------------- | ----------------------------------------- |
| Simple / focused | 5-30 rules | 4-8 hours total |
| Standard | 30-60 rules | 8-24 hours total |
| Comprehensive | 60+ rules | 24-50+ hours total |
Effort is spread across multiple testing rounds and does not need to happen all at once.
### Summary: Effort at a Glance
| Phase | Who Does the Work | Your Team's Time | Agent's Time |
| --------------------------------- | ------------------------- | ------------------------------------ | ------------------------------------------ |
| 1. Playbook Planning | Your team | Under 1 hour | None |
| 2. Materials Gathering | Your team | 30 min | Minimal |
| 3. Initial Build | Agent (with brief intake) | Under 10 minutes | 30 min to a few hours |
| 4. Strategy Assessment (optional) | Your team + Agent | A few hours to weeks (SME review) | 30 min (rewrites on instruction) |
| 5. Testing & Refinement | Your team + Agent | 4-60+ hours (scales with complexity) | 30 min to 1 hour (rewrites on instruction) |
**Key Principle:** The agent does the drafting and rewriting at every stage. Your team's role is to **provide the right inputs, validate the positions, and direct the refinements**. The more complete your source materials and the clearer your team's feedback, the faster the playbook reaches production quality.
# Step-by-Step Guide to Creating, Maintaining, and Running Playbooks
## Creating Your Playbook
Creating a playbook walks through the steps below, from opening a workspace to sharing the finished playbook.
See [Creating a workspace](/workspaces#creating-a-workspace) for how to name and set it up.
See [Creating a workspace](/workspaces#creating-a-workspace) for what materials to attach.
The agent reviews your materials and asks a short set of questions to confirm context, including:
* **Confirms your role and contract type** (for example, vendor, SaaS subscription agreement)
* **Paper type** the playbook should support (first-party, counterparty, or both)
* **Redline authorship** (only if attached files include redlines) confirming which redlines represent your positions
* **Discrepancy Precedence** (if there is more than one type of file) which file should control if there are discrepancies in positions
* **Redlining Style** (do you prefer to work with counterparty language, push your preferred language, or somewhere in between)
* **Comment preference** (comment on every redline by default, or only where your materials specify)
Building a playbook requires significant reasoning by the agent and includes agentic self-audits. Agentic build time will depend on the size and volume of the materials you attached, but can take more than an hour. No need to sit and watch the agent work, continue with your day and check back in a while for a completed playbook. The completed playbook is called your Rule Table and is available as a workspace artifact.
Run your playbook. Review the results. If the results don't match your expectations, tell the chat to modify your rule table so that if it ran again the results would be X instead of the results you received.
Share with your entire organization or with named individuals — see [Sharing a workspace](/workspaces#sharing-a-workspace) for what editors and viewers can each do.
## Maintaining Your Playbook
New template? New negotiation positions? Revise your playbook at any time by simply chatting with the agent, or if you prefer editing fields directly.
In the workspace chat, describe the change. For example:
* `Change all my aggressiveness scores to 2.`
* `Update my playbook to reflect my new standard form.` (after uploading the new standard form)
* `Add a rule addressing [new issue].`
* `My company just decided we need to have an arbitration clause. Please add it to my playbook.`
The agent revises the playbook.
Click **Edit** on the playbook. Every field, including the negotiation guide, agentic instructions, fallbacks, approval paths, aggressiveness, and evidence, is editable.
## Running Your Playbook
You run playbooks from the Word Add-in.
Open the document you want to review in Word.
Launch Vera Engage inside Word.
Click the paperclip, choose **Add Workspace**, and select the playbook workspace. Filter by the `playbook` tag (applied automatically whenever you use `create playbook`) to find it quickly. Attach any other deal context you want the agent to consider, such as counterparty emails, an issues list, or approval emails.
Kick off the playbook run by typing "run playbook". You can also attach any additional context that may be relevant to the review of this turn of the contract - negotiation notes, escalation emails, RFP responses, term sheets, etc.
If your playbook contains fallback (or variable) positions, the agent will assess the available context to determine if any of the available facts require the application of a fallback. For most companies, the types of facts that trigger variables are deal size, negotiation round, counterparty region, or the nature of the agreement such as the goods or services.
For any question where the answer is not in the available context (your conversation, the document being reviewed, the Icertis platform data about the contract), the agent will ask the user for the answers before running the playbook.
Each issue card shows the deviation score, the rule, the AI explanation for why the rule triggered (or did not), the proposed redlines, and the proposed comment. Apply on a rule-by-rule basis or use **Apply All** to accept everything at once.
## Reading Playbook Results
### The deviation score
A deviation score sits at the top of the results and tells you at a glance how far this contract is from your standard position. Alongside it you get the counts behind the number: how many high, medium, and low risk rules triggered, and how many came back compliant.
The score updates in real time as you apply, concede, or skip rules, so it doubles as a progress indicator through a review.
### Risk color coding
Every triggered rule is color coded, and the results bar at the top summarizes the whole contract:
| Color | Meaning |
| :--------- | :------------------------------------------- |
| **Red** | High risk. Your first priority. |
| **Orange** | Medium risk. |
| **Yellow** | Low risk. |
| **Green** | No deviation from your acceptable positions. |
Click any color bar to filter and sort the results to that risk level. On a contract with dozens of triggered rules, this is the fastest way to work top-down by severity.
### What is on an issue card
| Field | Notes |
| :-------------------- | :--------------------------------------------------------------------------------------------------------------- |
| **Rule Name** | A customizable label for the issue. |
| **Internal Guidance** | Context for the human reviewer. **The AI does not read this field.** Use it to tell reviewers why a rule exists. |
| **Reason Triggered** | The agent's explanation of why the rule fired, or did not. |
| **Proposed redline** | Tracked changes ready to apply. |
| **Proposed comment** | Explanatory comments. Not every rule has them. |
Internal Guidance is the field teams most often misuse. It is written for people, not for the agent. Guidance you want the agent to act on belongs in the negotiation guide or the agentic instructions.
### Actions on a rule
| Action | What it does |
| :----------------------------- | :-------------------------------------------------------------------------------------- |
| **Apply** | Inserts the redline or comment as a tracked change under your name. |
| **Apply All** (within a rule) | Applies every redline and comment for that one rule. |
| **Fallback** | Applies a pre-set acceptable middle-ground position, defined when the rule was built. |
| **Concede** | Accepts the counterparty's position, removes the flag, and adjusts the deviation score. |
| **Apply All** (whole playbook) | Applies every recommended redline and comment across the playbook. |
**Concede is scoped to this contract.** Conceding turns the rule off for future turns of **this contract only**. The rule still runs on every other agreement. Use it when you have decided to give way on a point in this negotiation, not when you want to change your position generally. For that, edit the rule.
**When two rules conflict.** If two rules collide (for example, one deleting a paragraph another is editing) the agent flags the conflict and leaves the bar solid-colored so you can navigate to it and resolve it manually.
**After Apply All.** The redlines are standard Word tracked changes. Accept or reject them individually in Word as you normally would.
# FAQs
## Sunsetting of Older Playbook Experience FAQ
Q: What happens with my existing playbooks in the old format?
A: Your playbooks will still be available to run and edit in the near term. For a better run experience, we encourage you to migrate them to the agentic playbook experience.
\_\_\_\_\_\_\_\_\_\_\_\_
Q: How can I migrate my existing playbooks?
A: It's easy!
1. **Create a new workspace** in the Vera Engage webapp
2. **Type "migration" in the workspace chat.**
3. **Answer a few questions** — the agent will take it from there
That's it. Once you've answered the questions, go do something else — the migration runs in the background and may take a while to complete. You'll have a fully built agentic playbook waiting for you when it's done.
\_\_\_\_\_\_\_\_\_\_\_\_
Q: Will migrating my existing playbook, delete what I have today?
A: No. The Migration skill only *creates* the new playbook format — your existing playbook remains exactly as it is. We encourage you to compare the results and make any tweaks you'd like in the new version.
\_\_\_\_\_\_\_\_\_\_\_\_
Q: Do you have a date by which you will sunset access to my older playbooks?
A: Not yet. We want to make sure our customers with older playbooks have a chance to migrate and explore. That being said, having two playbook experiences in the platform at the same time can be confusing to your users and we encourage you to move to the new experience as soon as possible.
***
Q: Can I still create new playbooks using the older experience.
A: No. We no longer support the creation of rule-by-rule playbooks.
## Troubleshooting
**Q. When I try to run the playbook from the chat, it says it can't find it.**
A. Check that you've attached your playbook workspace using the paperclip at the bottom of the chat. If you have, check that your playbook saved as a workspace artifact in the workspace. If there is no workspace artifact, navigate to the conversation in the workspace where you created the playbook and look for the rule table in the conversation artifact and promote it to a workspace artifact.
**Q. I clicked on the playbook menu and don't see my playbook.**
A. The playbook menu today houses the older playbook experience. Your workspace created playbook is available in your list of workspaces. If you don't see it, try filtering in your workspaces to just the playbooks.