实操示例 · 扫描的 PDF 转 Excel
将 50 页扫描发票转换为已核对的应付账款分类账
一位会计收到了一份用于市议会会议的 50 页发票扫描 PDF。他们要求 将其转换为 Excel 格式的应付账款分类账,每张发票占一行,且每个金额都直接从扫描的发票中读取。
读取了扫描件,构建了带有总计以及与封面备忘录核对项的分类账,且数字精确到分。随后,在同一个 Excel 文件中,它为每张发票添加了总账类别以及一个可加回总计的汇总表。
1上传:一份 50 页的扫描 PDF、一份封面备忘录和九张发票
City of Bunker Hill Village · Agenda Request
Item XII · Payment(s) Above $50,000
A. City of Houston water bill $74,323.35
B. City of Houston water bill $69,767.50
C. HR Green, BHV WO#1 Inv. #4 $101,219.80
D. HR Green, TIRZ WO#2 Inv. #5 $93,264.32
E. HR Green, BHV WO#1 Inv. #6 $72,728.61
F. HR Green, TIRZ WO#2 Inv. #6 $77,070.36
G. Memorial Villages Water Auth. $66,669.83
H. Memorial Villages Water Auth. $64,160.86
I. Red Oak Construction $97,231.03
- 50已读取的扫描页数
- 9找到并录入的发票
- $0.00与备忘录的差额
- 1Excel 文件,针对请求 2 进行了再次编辑
打开真实文件
这些是运行中产生的实际文件。请下载并在 Excel 中打开它们。
文件来源: City of Bunker Hill Village, Texas, 2026 年 8 月 18 日市议会会议资料包,议程第 XII 项,“金额超过 50,000 美元的付款”。一份公开记录。
01 · 初始文件
一份打印的备忘录,后接 49 页扫描件
第 1 页是一份打印的备忘录,列出了九张发票(A 到 I),每张金额均超过 50,000 美元。第 2 至 50 页是扫描件:水费账单、附有审批备忘录及分包商支持文件的工程发票、带有流量记录的废水账单以及一份施工发票。
扫描的页面是没有文本的图片。在询问了它们使用的语言后, 首先通过光学字符识别(OCR)读取了这些页面。PDF 本身没有被更改。
02 · 请求 1
“构建应付账款分类账,每张发票一行”
会计输入的内容
This PDF is Item XII ('Payment(s) Above $50,000') from the Bunker Hill Village council packet for 18 August 2026. Page 1 is the typed cover memo listing nine invoices, A to I. Pages 2-50 are scans of those invoices plus backup (TGC approval memos, sub-consultant invoices, status reports, water flow records).
Please build an AP ledger in Excel with one row per memo item A-I. Do not add rows for the sub-consultant invoices (United Engineers, TEI, Cypress Environmental) - they are backup inside the HR Green packages. Columns:
1. Memo item (A-I)
2. Vendor
3. Invoice no. (for the City of Houston bills, use the account number)
4. Invoice / bill date
5. Service period or description
6. Amount due - read from the scanned invoice itself, not the memo
7. PDF page of the invoice's first page
Add a total row, and a reconciliation line showing the memo total, the ledger total and the difference.的操作
- 在扫描件中找到了九份发票中每一份的第一页,并记录了其对应的 PDF 页码。
- 按照要求,每个金额都是直接从发票本身读取的,而不是从备忘录中读取。
- 按照要求,排除了夹在工程文件包中的分包商发票。
- 制作了一个新的 Excel 文件,备忘录中的每个项目(A 到 I)各占一行。
- 添加了使用公式计算的总计行,以及三行核对数据:备忘录总计、分类账总计和差额。点击下方绿色单元格可查看其中一个。
=SUM(F3:F11)| A | B | C | D | E | F | G | |
|---|---|---|---|---|---|---|---|
| 1 | Bunker Hill Village - Council 18 Aug 2026 - Item XII Payments Above $50,000 - AP Ledger (one row per memo item A-I) | ||||||
| 2 | Memo item | Vendor | Invoice no. (City of Houston = account no.) | Invoice / bill date | Service period or description | Amount due (USD, from scanned invoice) | PDF page (invoice first page) |
| 3 | A | City of Houston | 7099-3004-6012 | 06/19/2026 | May 2026 water purchase - treated water | 74323.35 | p.2 |
| 4 | B | City of Houston | 7099-3004-6012 | 07/22/2026 | June 2026 water purchase - treated water | 69767.5 | p.4 |
| 5 | C | HR Green, Inc. (BHV Work Order #1) | 202503 | 05/19/2026 | Professional services 04/01/2026-04/30/2026, Memorial/Gessner Phase II (Proj 2502530.01/.02) | 101219.8 | p.7 |
| 6 | D | HR Green, Inc. (TIRZ Work Order #2) | 203719 | 06/18/2026 | Professional services 05/01/2026-05/31/2026, Memorial/Gessner Phase II (Proj 2502530.03/.04) | 93264.32 | p.25 |
| 7 | E | HR Green, Inc. (BHV Work Order #1) | 204899 | 07/15/2026 | Professional services 06/01/2026-06/30/2026, Memorial/Gessner Phase II (Proj 2502530.01/.02) | 72728.61 | p.33 |
| 8 | F | HR Green, Inc. (TIRZ Work Order #2) | 204901 | 07/15/2026 | Professional services 06/01/2026-06/30/2026, Memorial/Gessner Phase II (Proj 2502530.03/.04) | 77070.36 | p.41 |
| 9 | G | Memorial Villages Water Authority | 001-26-27 | 06/01/2026 (due 07/10/2026) | May 2026 wastewater treatment | 66669.83 | p.46 |
| 10 | H | Memorial Villages Water Authority | 002-26-27 | 07/01/2026 (due 08/10/2026) | June 2026 wastewater treatment | 64160.86 | p.48 |
| 11 | I | Red Oak Construction LLC | 1057 | 04/30/2026 (due 06/03/2026) | BHV Water Plant #1-2 Chloramine Upgrades | 97231.03 | p.50 |
| 12 | TOTAL (ledger) | 716435.66 | |||||
| 13 | |||||||
| 14 | Memo total (sum of memo items A-I) | 716435.66 | |||||
| 15 | Ledger total (SUM formula cell reference) | 716435.66 | |||||
| 16 | Difference (Ledger - Memo) | 0 | |||||
我们将其与扫描件进行了核对: 所有九个金额、发票号和页码均正确无误,并且分类账总计 716,435.66 美元与备忘录总计相符。
03 · 请求 2
“现在添加总账(GL)类别和汇总表”
会计输入的内容
In the same workbook, extend the ledger: 1. Add a 'GL category' column using these categories: Water purchase; Wastewater treatment; Engineering - Memorial/Gessner Ph II - BHV Work Order #1; Engineering - Memorial/Gessner Ph II - TIRZ Work Order #2; Capital - water plant. 2. Add a 'Handwritten notes' column: transcribe any handwriting on that item's pages (e.g. marks on the TGC approval memo or the invoice itself). Write 'none' if there is no handwriting, and 'illegible' if you can't read it - please don't guess. Typed 'Approved by' stamps don't count. 3. Add a 'Summary' sheet with SUMIF totals by GL category, and a check that the category totals add back to the ledger total. Don't change the turn-1 rows or amounts.
的操作
- 打开刚刚制作的分类账,并对同一个 Excel 文件进行编辑。
- 添加了“总账(GL)类别”列,仅使用会计提供的五个类别。
- 添加了新的“汇总”工作表:每个类别一个 SUMIF 公式、一个总计,以及一个最终核对差额为 0 的检查行。
- 未更改请求 1 中的任何行或金额。我们核对了每一个单元格。
Water purchase| A | B | C | D | E | F | G | H | |
|---|---|---|---|---|---|---|---|---|
| 1 | Bunker Hill Village - Council 18 Aug 2026 - Item XII Payments Above $50,000 - AP Ledger (one row per memo item A-I) | |||||||
| 2 | Memo item | Vendor | Invoice no. (City of Houston = account no.) | Invoice / bill date | Service period or description | Amount due (USD, from scanned invoice) | PDF page (invoice first page) | GL category |
| 3 | A | City of Houston | 7099-3004-6012 | 06/19/2026 | May 2026 water purchase - treated water | 74323.35 | p.2 | Water purchase |
| 4 | B | City of Houston | 7099-3004-6012 | 07/22/2026 | June 2026 water purchase - treated water | 69767.5 | p.4 | Water purchase |
| 5 | C | HR Green, Inc. (BHV Work Order #1) | 202503 | 05/19/2026 | Professional services 04/01/2026-04/30/2026, Memorial/Gessner Phase II (Proj 2502530.01/.02) | 101219.8 | p.7 | Engineering - Memorial/Gessner Ph II - BHV Work Order #1 |
| 6 | D | HR Green, Inc. (TIRZ Work Order #2) | 203719 | 06/18/2026 | Professional services 05/01/2026-05/31/2026, Memorial/Gessner Phase II (Proj 2502530.03/.04) | 93264.32 | p.25 | Engineering - Memorial/Gessner Ph II - TIRZ Work Order #2 |
| 7 | E | HR Green, Inc. (BHV Work Order #1) | 204899 | 07/15/2026 | Professional services 06/01/2026-06/30/2026, Memorial/Gessner Phase II (Proj 2502530.01/.02) | 72728.61 | p.33 | Engineering - Memorial/Gessner Ph II - BHV Work Order #1 |
| 8 | F | HR Green, Inc. (TIRZ Work Order #2) | 204901 | 07/15/2026 | Professional services 06/01/2026-06/30/2026, Memorial/Gessner Phase II (Proj 2502530.03/.04) | 77070.36 | p.41 | Engineering - Memorial/Gessner Ph II - TIRZ Work Order #2 |
| 9 | G | Memorial Villages Water Authority | 001-26-27 | 06/01/2026 (due 07/10/2026) | May 2026 wastewater treatment | 66669.83 | p.46 | Wastewater treatment |
| 10 | H | Memorial Villages Water Authority | 002-26-27 | 07/01/2026 (due 08/10/2026) | June 2026 wastewater treatment | 64160.86 | p.48 | Wastewater treatment |
| 11 | I | Red Oak Construction LLC | 1057 | 04/30/2026 (due 06/03/2026) | BHV Water Plant #1-2 Chloramine Upgrades | 97231.03 | p.50 | Capital - water plant |
| 12 | TOTAL (ledger) | 716435.66 | ||||||
| 13 | ||||||||
| 14 | Memo total (sum of memo items A-I) | 716435.66 | ||||||
| 15 | Ledger total (SUM formula cell reference) | 716435.66 | ||||||
| 16 | Difference (Ledger - Memo) | 0 | ||||||
=SUMIF('AP Ledger - Item XII'!H3:H11,A3,'AP Ledger - Item XII'!F3:F11)| A | B | |
|---|---|---|
| 1 | Summary by GL category (SUMIF from AP Ledger - Item XII) | |
| 2 | GL category | Total (USD) |
| 3 | Water purchase | 144090.85 |
| 4 | Wastewater treatment | 130830.69 |
| 5 | Engineering - Memorial/Gessner Ph II - BHV Work Order #1 | 173948.41 |
| 6 | Engineering - Memorial/Gessner Ph II - TIRZ Work Order #2 | 170334.68 |
| 7 | Capital - water plant | 97231.03 |
| 8 | Total of categories | 716435.66 |
| 9 | Ledger total (AP Ledger - Item XII!F12) | 716435.66 |
| 10 | Check (categories - ledger; must be 0) | 0 |
从一堆扫描件到可核对的数据
大多数 AI 工具只能在聊天回复中总结扫描件。而 会阅读每一页,将数据放入带有公式的真实电子表格中,标明每个数据来自哪一页,并且在你提出更多要求时,继续在同一个文件中进行编辑。
继续阅读




