403-405 Sarthak-1, Swastik Cross Road, Ahmedabad, Gujarat info@pcnetinfotech.com 079 2656 3124
Manufacturing — Industrial Rubber Products

Darshan Rubber Industries

Prompt2Publish — Serverless Generative AI on Amazon Bedrock

300+
SKUs described in month 1
30–60s
Per SKU generation time
₹1–2.5L
Agency spend avoided/year
2–3 days
Trade-fair prep (was 2–3 weeks)

01 The Challenge

Darshan Rubber Industries manufactures industrial rubber components — gaskets, seals, moulded and extruded parts — for domestic B2B buyers and export markets. Despite a rich product catalogue, their written presence lagged far behind their factory's actual range.

300+ SKUs carried one-line descriptions — datasheets existed but were never converted to buyer-facing copy.
Product copy was written by technical staff between production responsibilities — slow, inconsistent, uneven export English.
Listings across website and marketplace portals drifted out of sync, confusing buyers comparing the same SKU.
Application and industry content was almost absent — surrendering long-tail buyer-keyword search traffic.
Trade-fair collateral was rushed together in the weeks before each exhibition season.
External agency writers lacked polymer-domain accuracy — copy came back with spec errors engineers had to fix.
Business Impact
  • 300+ SKU description backlog representing an estimated 75–100 hours of technical-staff writing time.
  • Export enquiry conversion weakened by inconsistent listings against international competitors.
  • Application notes took 1–3 days each — so very few were ever produced.
  • Long-tail SEO demand for compound- and application-specific keywords went uncaptured.

02 The Solution

Pcnet Infotech deployed Prompt2Publish — a fully serverless generative AI content pipeline on Amazon Bedrock — turning Darshan Rubber's datasheet library into the raw material for their entire content estate.

SKU Description Backlog → Batch Generation

Datasheet rows exported as CSV and dropped into S3 ingest/. S3 event notifications fan each row into the Lambda generator. 300+ SKU backlog cleared in days of unattended runs at a fraction of a cent per description.

Export-Grade English → Template-Enforced Style

A tone guide for international B2B procurement engineers rides inside every prompt. A banned-superlatives list keeps copy factual and professional. One template governs structure — website and marketplace listings finally match.

Technical Accuracy → Datasheet-Grounded Prompts

Specifications are injected verbatim from SKU master data into every prompt. The system prompt forbids inventing or altering any rating, standard, or test value. Engineers verify specs — they no longer write paragraphs.

Thin Application Content → Industry Application Notes

The application_note template targets one sector at a time — automotive sealing, pharma-grade gaskets, infrastructure joints. Steady cadence builds a long-tail keyword library that listings alone cannot capture.

Seasonal Collateral Crunch → On-Demand Trade-Fair Kits

Exhibition briefs batch-generate product one-pagers and booth social posts from the same grounded templates. Trade-fair preparation compresses from 2–3 weeks of scramble to 2–3 days of review and assembly.

03 Architecture & Data Flow

Marketing User
API Gateway
IAM / SigV4
Lambda
Python 3.11
Amazon Bedrock
Claude 3 Haiku
S3
generated/

Batch path: SKU datasheet CSV → S3 ingest/ → S3 event notification → Lambda → Bedrock → S3 generated/
Search: Marketing team → Amazon Q Business → indexed content library

04 Results & Outcomes

99%
Faster time to first draft per SKU
₹60–90K
Staff time avoided in month 1
$20–30
Total AWS platform cost per month
+10–20%
Projected export enquiry uplift
"Three hundred products, and until now most of them had one-line descriptions. Prompt2Publish turned our datasheets into consistent, export-ready copy across the website and marketplace listings — our engineers now verify specifications instead of writing paragraphs." — Managing Partner, Darshan Rubber Industries

05 Technology Stack

AWS ServicePurposeConfiguration
API GatewaySecure ingressREST API; IAM SigV4; HTTPS; 10 rps burst
AWS LambdaContent generationPython 3.11; 512 MB; 60s timeout; concurrency 1
Amazon BedrockLLM inferenceClaude 3 Haiku; on-demand; temp 0.3–0.7
Amazon S3Templates & contentSSE encryption; 90-day lifecycle on drafts/
Amazon Q BusinessContent searchWeb experience; 1 Pro user (~$20/mo)
CloudWatchLogs & metrics30-day retention; latency, error, token alarms
CloudTrailAPI auditFull Bedrock, S3, Lambda, API Gateway logging
Deployment Reference
CustomerDarshan Rubber Industries
IndustryManufacturing — Rubber
AWS PartnerPCNet Infotech
ACE OpportunityO22465608
Status✓ Approved — Launched
Regionus-east-1
Primary FMClaude 3 Haiku
Run Cost$20–30/month
Content Types Deployed
product_description
Per-SKU buyer-facing copy
application_note
Sector-specific tech marketing
industry_blog
Long-tail SEO articles
export_brochure
Family-level collateral sections
trade_fair_post
Exhibition promotion posts
Key KPI Targets
Spec accuracy100%
Template adherence>95%
Human edit rate<20% minor
Generation latency30–60 sec
API availability>99.9%
Cost per SKU<$0.10
Deploy for Your Business All Case Studies

Ready to Transform Your IT Infrastructure?

Let our experts build the right solution for your business. Talk to us today.