Getting Your Legacy Contracts Ready for Metrics: A Practical Guide
If you attended our recent webinar on proving legal’s value through metrics, you know that comprehensive contract data is the foundation of meaningful analytics. But here’s the challenge many of you mentioned: your most valuable insights are often locked in legacy contracts that were uploaded before Smart Import, executed outside workflows, or sitting in file folders waiting to be processed.
This guide walks through practical strategies for using Ironclad’s Smart Import and Run AI features to extract the metadata you need for robust contract analytics.
Understanding the two features
Smart Import is for new uploads—contracts you’re bringing into the Repository for the first time through the Imports tab, and it automatically detects properties as documents are uploaded.
Run AI is for contracts already in your Repository that you can access from the Dashboard. Use it when you need to extract metadata from contracts uploaded without AI detection, apply newly created custom properties to existing records, analyze contracts you’ve already negotiated in Ironclad, or evaluate how changes to your custom AI clauses or properties affect your contracts.
Both tools use the same AI model and detect the same 194+ out-of-the-box properties. The difference is timing and workflow.
Pre-work: Setting up for success
Before running your first bulk import, take these steps to improve accuracy:
Add your subsidiaries
The AI detects counterparty names and addresses. If it doesn’t know your subsidiaries, it may incorrectly tag your own entities as counterparties. You should add any subsidiaries and their addresses to Ironclad before doing a Smart Import in order to teach the AI to ignore those values for Counterparty Name and Address predictions.
Where to add them:
From the Repository: In the Imports tab, click on a Smart Import record you want to add a subsidiary to. From the panel on the right, select Counterparty Name or Counterparty Address. Click Add to Subsidiary Information indicated by a gray circle with a plus sign. Enter the subsidiary’s information and click Submit.
From Company Settings: Click on your profile icon located in the top right corner of Ironclad. Navigate to Company Settings > Profile. Click the Show Subsidiaries button located towards the bottom right, and then click Add Subsidiary. Enter the subsidiary’s information and click Update Profile.
Pro tip: After you add subsidiary information, you can refresh your records by selecting the records you want to refresh, and then click on the three stacked dots located in the top right. Click Refresh Company Properties. AI will refresh the predictions on imported records to recognize the added subsidiaries and disregard them from the counterparty name.
Review out-of-the-box properties
Navigate to Company Settings > Data Manager to see which properties Ironclad detects automatically. By default, Ironclad AI automatically detects properties, or specific pieces of information that Ironclad extracts automatically from contracts. This usually includes dates, numbers, or strings that describe key items of the contract, such as the counterparty name, agreement date, or contract value.
Knowing what’s detected out-of-the-box helps you identify gaps where you might need custom properties.
Create custom properties (if needed)
If you’re tracking metrics specific to your organization, you can set up custom AI properties before your bulk import. Custom AI properties use AI to automatically extract unique contract details using Smart Import and Run AI. You can train Ironclad AI properties to improve their accuracy using examples.
Best practices:
Typically, 5–10 varied examples are sufficient for most properties. Diversity is key, not quantity. Include versions of the property you might encounter to improve detection accuracy. Keep a record of both the paragraph where the property is located and the property itself. Include your standard terms for that property, redlined versions you’ve seen in past negotiations, and versions of the clause you’ve accepted from third-party contracts if applicable.
Smart Import: Uploading legacy contracts
Batch by contract type
Don’t upload everything as one massive import. Group contracts by type and set the record type for each batch. We recommend grouping your uploads by contract type and setting the record type with each group. For example, upload NDAs together and set the record type to be NDAs. This will save you the effort of setting the record type for each individual record later. The default record type is Imported.
Choose your upload method
You can import files from your local machine or from cloud storage providers OneDrive, Dropbox, and Box.com.
From local machine: Gather your documents on the local machine that you use to upload your files into Ironclad.
From cloud storage: If you want to upload records from your cloud storage provider, select the cloud storage provider from which you want to import. You will be prompted to login to your cloud storage account and select the desired files for import. You can import from multiple providers or locations within a single import.
Via email: You can use Smart Import to import completed contracts by emailing them directly to your company specific email address. You must send the email from the email address associated with your Ironclad user account. Click on your profile icon located in the top right corner of Ironclad. Navigate to Company Settings > Profile > Repository email. Copy the Repository email and send an email to the email address with the contracts that need to be imported. They must be in PDF, DOC, or DOCX format.
File requirements and limits
Each file must be 100mb or less and one of the following file types: .pdf, .docx, or .doc. Smart Import has a limit of 100 pages per file. You can upload up to 2,000 documents at a time.
Monitor your upload
The Ironclad application needs to be open during the entire upload process, so we recommend breaking up large uploads into smaller portions if you’re not sure if you will have an uninterrupted, stable internet connection.
What happens next: Initially, the metadata will be empty as the AI scans and generates the predictions. Smart Import runs each uploaded document through our AI model to suggest contract metadata to tag as properties.
Run AI: Processing existing records
When to use Run AI
You can run AI on your existing records on the Dashboard. This is especially useful for analyzing contracts you’ve already negotiated in Ironclad and evaluating how changes to your custom AI clauses or properties affect your contracts. You can select up to 2,500 documents at a time or select a view containing up to 2,500 records for AI analysis.
How to run it
When the analysis is complete, you’ll receive an email with the results in a CSV file.
Important limitations
If a record was previously added using metadata import, AI suggestions will not be gathered on the record. AI predictions will not be provided for records where data was manually added through metadata import or record editing, or where you’ve selected “don’t show again”. If you have previously uploaded a contract using Smart Import and run AI predictions on it, you will not receive new results if the underlying models haven’t changed.
New predictions may be generated when Ironclad updates its underlying model or when you make changes to AI Clauses or AI Properties.
Verification best practices
AI suggestions need human review. Here’s how to verify efficiently:
Use the “Needs verification” view
We have added a new default Dashboard view in the Repository section called “Needs verification,” which includes records with unverified AI predictions.
Create metadata groups
Group your AI-detectable properties together in Workflow Designer using the Reorder Metadata option. The goal is to create a condensed view into AI-detectable properties that can be viewed during the Review step, Archive step, and in the Dashboard.
Track accuracy patterns
Currently, the AI has limitations in accurately processing certain elements from tables, handwriting, watermarked paper, low quality scans, and overlapping fields.
Use these patterns to refine custom properties or adjust your scanning quality.
Getting metrics-ready data
Once your legacy contracts are tagged, you can build the metrics dashboards discussed in the webinar.
Create your baseline views
Set up filtered views for key contract segments:
- Contracts expiring in next 90 days
- High-value agreements (Contract Value > $X)
- Contracts by counterparty or business unit
- Agreements pending renewal decisions
Build Insights reports
Use detected properties from Smart Import to create reports on cycle time metrics, value analysis, risk indicators, and renewal pipeline tracking.
Set up automated tracking
With metadata extracted, enable email reminders and dashboards using the properties detected by AI.
Common issues and solutions
AI doesn’t detect a property you know is there:
When non-detection occurs (AI doesn’t tag the text as the property): Add the missed text as an example for that property. If the property appears in a table and table detection is off, add table detection. If the property appears in a new clause, include the clause where the property appeared as an example in one of the associated AI Clauses.
Counterparty keeps getting tagged as your company:
Add that entity to subsidiaries and refresh predictions on affected records using Refresh Company Properties.
Upload fails partway through:
The Ironclad application needs to be open during the entire upload process. We recommend breaking up large uploads into smaller portions if you’re not sure if you will have an uninterrupted, stable internet connection.
AI credits running low:
Every customer has a certain number of AI contract credits included in their base Ironclad package that can be used without additional charge. Additional uploads are subject to a fee. Please reach out to success@ironcladhq.com if you anticipate needing more than your allotted AI contract credits. If you run Smart Import on documents and the prediction fails, no AI credits are deducted from your total balance. Credits are only consumed when a prediction successfully completes.
Action plan
Here’s your step-by-step approach:
Week 1: Prepare
- Add subsidiaries to Company Settings
- Review out-of-the-box properties in Data Manager
- Create 2-3 critical custom properties (if needed)
- Test on 25-50 sample contracts
Week 2-3: Process
- Upload legacy contracts in batches by type
- Run AI on existing Repository records
- Verify high-priority properties first
Week 4: Refine
- Review accuracy patterns
- Adjust custom properties based on results
- Re-run AI on contracts where you improved detection
Ongoing: Maintain
- Check “Needs verification” view regularly
- Update custom properties as contract language evolves
- Refresh predictions when adding new subsidiaries
Measuring success
Track these indicators to know your data quality is improving:
- Percentage of contracts with verified key properties (target: 95%+)
- Time spent on manual data entry (should decrease significantly)
- Completeness of expiration dates for renewal pipeline
- Accuracy of contract value data for budget forecasting
Remember: perfect AI detection isn’t the goal. The goal is getting your contract data accurate enough to generate reliable metrics that prove legal’s value to the business.
Questions about implementing these strategies? Connect with other Ironclad users in the Community or reach out to your CSM for a data quality review.
Ironclad is not a law firm, and this post does not constitute or contain legal advice. To evaluate the accuracy, sufficiency, or reliability of the ideas and guidance reflected here, or the applicability of these materials to your business, you should consult with a licensed attorney. Use of and access to any of the resources contained within Ironclad’s site do not create an attorney-client relationship between the user and Ironclad.
