A submittal log is the running list of every shop drawing, product data sheet, sample and test report a contractor must send for review, with its spec section, who reviews it, when it is due and what the reviewer decided. Below is a log layout you can copy into Excel or Google Sheets, then the one failure that costs money, and how to feed the log from your email so nobody has to retype it.
This page is written for subcontractors and small to mid-size general contractors who keep the log in a spreadsheet and receive reviews by email. If your project already runs on a platform with a submittals tool, read the section on when to just use it.
A submittal log template you can copy
The columns below follow the fields that construction software and guides use for the same job. Procore's support page for its submittals log lists default columns including Spec Section, Number and Revision, Title, Type, Status, Responsible Contractor, Submit By, Ball In Court, Approvers, Sent Date, Returned Date and Due Date (Procore support, submittals log columns). A Datagrid guide describes the same structure with manufacturer, status code and revision tracking added (Datagrid, submittal log guide). Copy this table into a sheet as your header row and example rows.
| Column | What goes in it | Example |
|---|---|---|
| Submittal no. | Spec section plus a running number, with a revision suffix | 08 71 00-001 R1 |
| Spec section | The specification section the item is called for in | 08 71 00 Door hardware |
| Description | One line saying what the item is | Hinges and closers, stair 2 |
| Type | Shop drawing, product data, sample, test report, certificate | Product data |
| Responsible company | The subcontractor or supplier who owes the package | Your door hardware supplier |
| Submit by | The date the package must go out to hit the schedule | 2026-03-02 |
| Date sent | The day you actually sent it for review | 2026-03-04 |
| Reviewer | The architect, engineer or consultant who reviews it | Architect of record |
| Review due | Date sent plus the review period in the contract | 2026-03-18 |
| Status | A fixed list: Not sent, Under review, Approved, Approved as noted, Revise and resubmit, Rejected | Revise and resubmit |
| Ball in court | The one person who must act next | Project engineer |
| Date returned | The day the stamped copy came back | 2026-03-20 |
| Resubmittal round | 0, 1, 2 and so on, with a new review due date each round | 1 |
| Needed on site | The date the material must arrive | 2026-05-12 |
| Lead time (weeks) | Manufacturer lead time for ordering after approval | 10 |
| Order by | Needed on site minus lead time minus buffer: the date that really matters | 2026-03-03 |
The last three columns are the ones most templates leave out, and they are where the money is. The status values (Approved, Approved as Noted, Revise and Resubmit, Rejected) are the set the Datagrid guide lists, but your contract may use different wording, so use its list.
The formulas that make the log work
A log earns its keep when the dates calculate themselves. Three formulas do most of the work, and they work the same in Excel and Google Sheets. This is an illustration with invented dates to show the logic, not a client example.
| Column | Formula idea | Why |
|---|---|---|
| Review due | Date sent plus the review period in days from your contract | The due date follows the send date, so a late send is visible |
| Days overdue | Today minus review due, only where status is Under review | Sorts the log by who is late, not by number |
| Order by | Needed on site minus lead time in weeks times 7 minus a buffer in days | Shows the last day approval can arrive and still make delivery |
| Float | Order by minus today | A negative number means the material is already late if approval is not in hand |
Worked through with made-up numbers: a door hardware package is sent on 4 March with a 14 day review period, so review is due 18 March. The material is needed on site on 12 May and the manufacturer lead time is 10 weeks, so ordering must happen by 3 March with no buffer. The package is already behind before anyone is late. A log with only a status column would show "Under review" and nothing else. This is exactly why the order-by column matters more than the status column.
Add conditional formatting so a row turns amber when float is under 10 days and red when it is negative. Then filter by Ball in court every Monday and send each person only their own rows. That alone fixes a large part of the problem for teams under about 30 open items.
Where the spreadsheet fails: late approvals delay ordering
A late review is not just a late stamp. The Datagrid guide says contractual review periods are "obligations, not suggestions", that each rejection adds a full review cycle, and that a late review on a long-lead item can be hard to recover without acceleration costs (Datagrid). It also says a static spreadsheet typically will not flag overdue reviews and quotes CMAA guidance that a system should alert staff when a submittal has not been answered in time.
Here is how that plays out in a normal subcontractor week:
- The review comes back by email, nobody logs it. A stamped PDF lands in one person's inbox. The log still says "Under review". The supplier waits for approval it already has.
- A resubmittal restarts the clock, and the log does not. The due date stays at the first round, so the new overdue review never shows up.
- The order-by date is not tracked. Approval arrives, but the ten-week lead time now puts delivery after the date the crew needs the material.
- Nobody owns the chase. The reviewer is two weeks late, and the project engineer assumes the project manager already called.
Each of these is a data-entry gap, not a hard problem. That is why it can be fixed without replacing your tools.
How to update the log from email, step by step
We design the following flow. It is a design, not a delivered result: we do not claim a construction client for it. It reads the email you already receive and keeps the sheet current, with a person approving anything that goes out.
- Watch one shared mailbox. Submittal transmittals and reviewer replies go to, or are copied to, one address. A workflow tool such as n8n or Make picks up each new message. n8n's documentation describes an Email Trigger (IMAP) node that receives emails from an IMAP server, with an option to download attachments (n8n Email Trigger (IMAP)).
- Classify the message. A language model such as Claude reads the subject, body and attachment text and decides: new transmittal, review response, resubmittal or something else. Anything it is unsure about goes to a human queue instead of the log.
- Extract the fields. For a review response it pulls the submittal number, the stamp wording (for example "Revise and resubmit"), the date and the reviewer, and any comments.
- Match and propose. The tool finds the row by submittal number and proposes the update: status, date returned, new ball in court. The match is checked against your log, not guessed.
- A person confirms status changes. The project engineer sees the proposed change next to the source email and approves it with one click. Changing a status to Approved is never automatic.
- Recalculate dates. On a resubmittal the tool opens the next round, sets the new review due date from your contract period and recomputes the order-by date from lead time.
- Flag overdue reviewers and draft the chase. Each morning the tool lists reviews past due or due in two days. It drafts a polite email to the reviewer, and a person approves it before it is sent.
- Keep an audit trail. Every change stores the source email link and who approved it, so you can show the sequence if a delay claim arises.
When not to automate, and when to use your platform
Automating this job has a cost, and sometimes the cost is higher than the pain. Do not automate if:
- You run fewer than about 30 submittals a quarter. A disciplined sheet and a weekly review meeting are enough.
- Your log has no fixed status list or numbering. Fix the data first: automation on a messy log makes mistakes faster.
- Reviewers answer inside a platform that already tracks the review. Then the platform is the log.
- Nobody will own the approval step. The human confirmation is what keeps the log trustworthy.
Use an existing platform when the whole project team is on it. Procore's submittals log, for example, has the columns above, including Ball In Court and Due Date, and only a user with Admin-level permission to the Submittals tool can change a submittal's status (Procore support). If the general contractor runs the project there, you will be asked to work inside it anyway. Our automation is for the gap around that: the subcontractor's own internal log, the email-only reviewers, and the link into your purchasing and scheduling. Related jobs have the same shape: RFI logs, certificate of insurance tracking, lien waiver tracking and the subcontractor compliance tracker. The overview is on our AI automation for construction companies page.
Keeping an automated log trustworthy
Three rules keep an automated log safe to rely on:
- The source of truth stays visible. Every row links to the email or PDF that changed it.
- Low confidence goes to a person. If the stamp is unreadable or the number does not match a row, the tool asks.
- Nothing leaves the building without approval. Chasing emails and any message to the architect are drafts until someone sends them.
Review and approval decisions remain the reviewer's. The tool records them. It does not make them.
What shapes the cost and how to start
The right build is small: a mailbox watcher, a language model step, a spreadsheet or database, and one approval screen. How it is built depends on your tools, so the cost depends on:
- where the log lives today (Excel, Google Sheets, Airtable, an ERP or a platform),
- how many people send and review submittals by email, and how consistent the stamps are,
- whether the order-by dates must feed purchasing or scheduling,
- who must approve changes, and how they want to see the proposals.
Describe the job and the budget range you have in mind in the contact form. Devvista quotes in writing after a short call. For how we build this kind of work, see our AI agents and workflow automation services, and our case studies for delivered projects in other fields. .