The Client
David is a consultant for a real estate management company in Europe. They need to replace outdated meters with new digital ones equipped with radio transmitters to comply with European laws mandating meter replacement every 5-10 years. This upgrade facilitates remote meter readings.
Accurate data collection requires managing approximately 90 parameters per meter, including location, apartment details, ownership, and account settlement information. This ensures correct data collection and association with the proper billing accounts.
The company currently uses a combination of PDFs and existing databases to map buildings for account settlement.
David's client manages complexes with over 10,000 units, each requiring individual electric, gas, and water meters.
The Challenge
- The existing process involves significant manual data entry, leading to errors and inefficiencies.
- David couldn't find any existing tools specifically addressing their image recognition needs for gas meters.
- High labor costs arise from manually checking serial numbers and meter readings.
- The client's organization struggles to meet demand without significantly increasing the employee pool.
- Inaccurate readings and lack of data transparency lead to disputes with tenants, increasing compliance risks and additional administrative work.
The Solution
David proposed a system using Smartsheet for data collection.
Connex introduced their automation and AI expertise.
The Results
- Automation with OpenAI has achieved over 80% accuracy in recognizing meter types, models, serial numbers, and readings during preliminary phases.
- Production accuracy is expected to reach 99%.
- Automation is estimated to increase efficiency by a factor of 3 to 4 times compared to the current number of staff managing these operations manually.
Tech Used
- Smartsheet as the database. Airtable could be used as well.
- Pipedream to orchestrate workflows
- OpenAI / ChatGPT Vision for image recognition
Alternative Tech
- Airtable as the database
