跪拜 Guibai
← All articles
AIGC

A Bid Review in Minutes: How OCR Skills and an AI Agent Caught a 3,959-Yuan Quotation Gap

By 一只牛博 ·
Read original on juejin.cn ↗ Google Translate ↗ Alt translation

Document-heavy compliance workflows—bid reviews, contract checks, supplier onboarding—still consume hours of manual cross-referencing across PDFs. This pipeline shows that off-the-shelf OCR APIs, wired together by an agent that can self-diagnose configuration failures, can shrink that to minutes while producing a page-cited audit trail that holds up under scrutiny.

Summary

A complete bid review pipeline was assembled by combining two Tencent Cloud OCR Skills—one for document field extraction and one for borderless table recognition—inside the WorkBuddy AI agent. The process turned 8 page-by-page PDFs from a fictional printer procurement into four structured deliverables: a compliance matrix, a graded risk list, a quotation comparison CSV, and an executive summary. The agent corrected a supplier's self-assessed "non-compliant" delivery time by recalculating `12 <= 15` against the tender threshold, and it surfaced a 3,959.96-yuan mismatch between the same supplier's quotation summary and its independent pricing sheet.

The workflow was built without manually writing Skill definitions or configuring API keys from scratch. When the OCR Skills failed on first run, the agent diagnosed the missing environment variables and Python SDK dependency, then guided the user through Tencent Cloud's CAM console to create credentials and activate the OCR service. A custom business Skill, `tender-compliance-reviewer`, was generated from a natural-language description of the review rules, locking in the field schema, judgment logic, and evidence-formatting requirements so the process is repeatable.

Processing time per supplier set was tens of seconds to a few minutes, compared with an estimated hour or more of manual cross-referencing. The final output distinguishes hard failures (resolution below spec, missing warranty) from items needing clarification (conflicting amounts, missing attachments), leaving the procurement officer with a structured, page-cited audit trail rather than a black-box recommendation.

Takeaways
Eight page-by-page PDFs were processed into four review files: a 30-item compliance matrix, a graded risk list, a quotation comparison CSV, and a one-page summary.
Supplier B's quotation summary stated 178,960 yuan while its independent pricing sheet totaled 175,000.04 yuan—a 3,959.96-yuan gap flagged automatically.
Supplier B self-assessed its 12-day delivery as non-compliant; the agent recalculated `12 <= 15` against the tender's 15-day threshold and corrected the status to compliant.
Two Tencent Cloud OCR Skills handled distinct tasks: `extractdocagent` pulled structured fields from tender and response documents, while `recognizetableaccurate` reconstructed a borderless quotation table and exported it to Excel.
A custom Skill, `tender-compliance-reviewer`, was generated from a natural-language description and locked in the field schema, judgment rules, and evidence-formatting template so the workflow is repeatable.
When OCR Skills failed on first run, the agent diagnosed missing `TENCENTCLOUD_SECRET_ID/KEY` environment variables and an uninstalled `tencentcloud-sdk-python` package, then walked the user through key creation in CAM and fixed the local venv.
The agent explicitly warned that green local checks do not guarantee the cloud API is usable; OCR service activation in the console remained a manual step.
Free-tier quotas applied: Table Recognition V3 shared a 1,000-call/month pool, and the Document Extraction Agent offered 1,000 calls per user with a 1-year validity after first activation.
A final re-check on seven borderline items forced the model to output requirement, response value, comparison expression, judgment, and evidence page for each, pulling conclusions back to the source text.
Processing time per supplier set was tens of seconds to a few minutes; the estimated 10x efficiency gain is a scenario-based comparison against manual baseline work, not a controlled experiment.
Conclusions

The workflow's most practical design choice is separating "hard non-compliance" from "needs clarification" in the output. Conflating a missing attachment with a failed spec would force the procurement officer to untangle the agent's reasoning; keeping them distinct makes the matrix immediately actionable.

Generating the business Skill from a natural-language description, rather than hand-authoring SKILL.md and validation scripts, lowers the barrier to locking in domain-specific review rules. The model still calls deterministic OCR APIs underneath, so the "customization" is about fixing the schema and judgment logic, not hoping a general-purpose LLM gets the comparison right every time.

The agent's self-diagnosis loop—check Skill files, check env vars, check Python imports, report specific failures—turns the usual "it doesn't work" dead end into a sequence of fixable steps. This is a concrete pattern for making AI tooling usable by people who do not know where API keys live or how venvs load environment variables.

The deliberate decision to retain original decimals (152,000.04 yuan, not 152,000) and to avoid overwriting the independent quotation sheet with the response file's summary figure is what makes the output auditable. Rounding or merging sources would have hidden the 3,959.96-yuan gap that became the key finding.

Cache reuse on identical inputs is a double-edged feature: it saves API costs on repeated runs but can mislead performance measurements if a cached hit is mistaken for a fresh OCR call. The note about creating a new session for benchmarking is a small but important operational detail.

Concepts & terms
WorkBuddy
An AI agent desktop client that orchestrates tasks by calling Skills, reading local files, and executing local repairs. It uses Tencent Hunyuan (Hy3) as its reasoning model and can self-diagnose configuration failures.
Tencent Cloud OCR Skills
Pre-built capability packages for WorkBuddy that wrap Tencent Cloud's OCR APIs. Two were used: `extractdocagent` for structured field extraction from documents, and `recognizetableaccurate` for borderless table recognition and Excel export.
Skill (in WorkBuddy context)
A packaged capability that can be installed and called by the WorkBuddy agent. Skills wrap API calls, define input/output schemas, and handle authentication. Custom Skills can be generated from natural-language descriptions to lock in business-specific rules.
Compliance Matrix
A structured table mapping each tender requirement against each supplier's response, with a status of Compliant, Non-compliant, Missing, or To Be Confirmed, plus the source file and page number for every judgment.
CAM (Cloud Access Management)
Tencent Cloud's identity and access management service. API keys (SecretId/SecretKey) are created here; for production use, sub-user keys with scoped OCR permissions are preferred over root account keys.
Source: juejin.cn ↗ Google Translate ↗ Backup ↗