Curative letters: by hand or software?

Published 17 October 2026. Updated 17 October 2026.

Every title agency clears curative items somehow. Many do it by hand: someone reads Schedule B, makes a list in a notebook, spreadsheet or the order system, and types letters from a Word template. Software that reads the commitment and drafts the letters is the alternative. Neither is right for every agency. This guide sets out what the job takes, where hand work tends to break, where software helps, and when the simple route is the better one.

What does the job actually take?

Clearing curative items is more than writing letters. Each file needs the same chain of work:

1. Read every requirement and exception on the commitment.
2. Decide which ones need action and which stay as exceptions.
3. Match requirements to what the search package shows: recorded satisfactions, assignments, partial releases.
4. Work out who to write to, which may be a lender, an assignee, a judgment creditor, an association or the seller.
5. Draft each request with the right reference to the recorded instrument.
6. Track what went out, what came back and what is still open.

The reading and matching steps are where mistakes start. The letters are only as good as the list they come from. For the detail of those steps, see how to review a title commitment.

What does doing it by hand look like?

Hand work means the examiner or closer reads the commitment, keeps a list, and types or edits letters from templates. It has real strengths.

• No new tool. Nothing to learn, buy or connect.
• Full control. The writer chooses every word, which matters for odd files.
• Judgment stays close to the page. Someone who reads every line tends to notice what does not fit.
• Cheap at low volume. If a few files a month have a couple of items each, a template is hard to beat.

It also has known weak points, and they show up in many offices.

• Retyping errors. A book and page, an amount or a name is copied from the commitment into a letter. A transposed digit goes out unless someone catches it.
• Missed releases. A requirement lists a mortgage that was released years ago. If nobody matches it to the recorded satisfaction, a payoff request goes to a lender who has nothing to pay off. The guide on released mortgages covers that case.
• Citations live in notes. The list says "mortgage, payoff" but not which page of the commitment it came from, so the next person re-reads.
• Version confusion. Two copies of a task list, or a letter edited in two places, and nobody is sure which is current.
• Nobody checks the second time. The writer who made the list also checks it, and tired eyes skip the same line twice.
• Dependence on one person. The file waits when the one examiner who understands it is away.

None of these is a flaw in the people. They come from copying facts by hand between documents.

What should software do for this job?

Software here means a tool that reads the commitment, lists each item with its source, and drafts letters from fields it has checked. What matters is how it does that, so compare tools on these points.

• Does every item point to a page? Without that, you must re-read to trust it.
• Are letters filled in by fixed templates, or written freely each time? Freely written letters can change a figure or a name.
• Does it notice released and assigned mortgages? Those are easy to miss by hand.
• Does it say what it could not settle? A tool that never reports a gap leaves you to find the gaps yourself.
• What happens to scanned pages? Ask whether it reads them, and what you will see when a page is not read.

Software also has limits. You still read, edit, sign and send. It does not replace knowing the trade, and you still need to read the result with the commitment beside it.

How do the approaches compare?

When doing it by hand is enough

By hand is a sound choice when:

• you clear curative items on only a few files a month;
• most files are routine, with one mortgage, clean vesting and no old liens;
• one experienced person owns the whole job and keeps good notes;
• most of your search packages are scans. ClearCurative does not read scanned pages yet, so it would add little there.

In those cases a good template and a disciplined checklist will do. Do not buy software to fix a problem you do not have.

When software starts to pay off

Hand work gets expensive when files are messy or volume is steady. Look at software if:

• letters go out with wrong references, or you find errors after the fact;
• files with assigned mortgages, partial releases or name variances take far longer than the rest;
• different people keep different versions of the same task list;
• work waits for the one person who understands the file.

Some agencies keep a mix: hand work for simple files, a tool for the heavy ones.

A checklist that works either way

Whichever route you pick, these habits catch most errors.

• Put the commitment page on every task, not just the item name.
• Check every book, page, amount and name in a letter against the recorded source before it leaves.
• Match each payoff requirement to the search package before you ask anyone for a payoff.
• Name the addressee on the list, and say why that party.
• Keep one current copy of the task list.
• Mark items you could not settle, rather than leaving them out.
• Have the letter read once by someone other than the writer when the amount at stake is large.
• Log what went out and when.

How ClearCurative does this step

You upload a title commitment as a PDF with selectable text, a Word file or a text file, and you can add the prior policy and the recorded instruments of the search package, up to 8 files of 25 MB each. A model reads each file and reports every requirement and exception with its page and a quote. Code then checks each quote on its page, links requirements to recorded satisfactions, assignments and partial releases, and fills in the letters from checked fields. A field that fails its check is left as a placeholder for you to complete.

You get a cited task list as a PDF, request letters and affidavits as editable Word drafts, and the full result as JSON. Whatever could not be settled is listed at the top of the task list and in the full result. A second, stronger AI model reviews every result and lists what is left for you to decide; the drafting model corrects the serious problems the review finds, and code checks every correction. Results are not verified by a person. Letters are drafts that you edit, sign and send. Nothing is sent for you.

It does not read scanned pages yet; they are listed as not read. Photos cannot be uploaded. It does not connect to SoftPro, and it does not track replies to your letters. One commitment is 22 credits, with no subscription. To try it on a real file, sign up.

Questions

Is a Word template good enough for curative letters?

For a handful of files a month, often yes. A template saves the boilerplate. The risk is in the fields you retype each time: names, book and page, amounts and addressees. Check those against the source on every letter.

At what volume does doing it by hand stop working?

There is no fixed number. The signal is when letters go out with the wrong recording reference, when two people keep separate versions of the same list, or when a file waits because the one person who knows it is out.

Does software remove the need to read the commitment?

No. Whoever signs and sends a letter still needs to understand the item. Software can speed up finding, citing and drafting, but the commitment and the recorded instruments always govern.