Energy

Financial Reconciliation

The client’s finance team was bogged down with manual efforts around reviewing financial information, which required reconciling and validating data from their ERP with the reporting platform. Monthly losing took between 20 – 25 days.

Solution

The client was using an ERP and Hyperion. However due to the amount of moving parts and transactions, they had difficulty getting the systems match at any one point. This was further complicated because much of the reconciliation was done by the accounting team. We developed an AI bot that, on a fiscal-period schedule, logs into both the ERP and reporting environments, performs comparisons and reconciliations, and reports any mismatching information to the accounting team.

Result

The bot could run and close the systems within 5 minutes
The time to close decrease form 20 days to 8 days, a 250% decrease.
Adjustments made due to timing errors was decreased 90%
Reduce the cost to conduct audits by 20%

Indicative AI Models & Methods

Robotic Process Automation (RPA)
Machine Learning for Anomaly Detection
Natural Language Processing (NLP)

Implementation ease:

Moderate to High.

Indicative Effort:

180 – 230 days

Hyperion MTD Reconciliation

The company needs to query a report of accounts using a connection through SQL cube and needs to generate a report every day during non-working hours, in addition to this, the report must be approved with a different standard

Solution

We create a robot capable of navigating this particular application to obtain and process the correct information, generating the correct report every day.

Result

50% reduction in process hours
2-4 hours daily

Indicative AI Models & Methods

Robotic Process Automation (RPA)
Natural Language Processing (NLP)
SQL Query Automation
Report Generation Automation

Implementation ease:

Easy

Indicative Effort:

2 months

Data Ingestion and Analysis

Our clients rapid growth has led to accelerated adoption of technology solutions for the monitoring and tracking of operational performance in the different oil fields spread throughout the US territory. The client engaged to solve the challenges inherent to ingesting, processing and leveraging the new big volumes of incoming data into the decision-making process.

Solution

We built a data ingestion platform for the field IoT devices to feed data to. We trained and deployed an AI engine that facilitates the dynamic estimation of nominal and abnormal ranges in the streaming data and via an API reports back to the monitoring solutions the optimal alarm thresholds, previously defined manually. We built a classification system that can be further improved into a predictive maintenance solution leveraging AI models.

Result

80% reduction in manual hours
2/3 cost reduction
More coverage of low yielding pumps
Faster recognition of anomalies or spills

Indicative AI Models & Methods

Machine Learning for Dynamic Data Analysis
Data Ingestion and Processing Techniques
Classification Systems for Operational Data

Implementation ease:

High.

Indicative Effort:

200 – 250 days

Financial Reconciliation

The client’s finance team was bogged down with manual efforts around reviewing financial information, which required reconciling and validating data from their ERP with the reporting platform. Monthly losing took between 20 – 25 days.

Solution

The client was using an ERP and Hyperion. However due to the amount of moving parts and transactions, they had difficulty getting the systems match at any one point. This was further complicated because much of the reconciliation was done by the accounting team. We developed an AI bot that, on a fiscal-period schedule, logs into both the ERP and reporting environments, performs comparisons and reconciliations, and reports any mismatching information to the accounting team.

Result

The bot could run and close the systems within 5 minutes
The time to close decrease form 20 days to 8 days, a 250% decrease.
Adjustments made due to timing errors was decreased 90%
Reduce the cost to conduct audits by 20%

Indicative AI Models & Methods

Indicative AI Models & Methods

Robotic Process Automation (RPA)
Machine Learning for Anomaly Detection
Natural Language Processing (NLP)

Implementation:

Moderate to High.

Indicative Effort:

180 – 230 days

Hyperion MTD Reconciliation

The company needs to query a report of accounts using a connection through SQL cube and needs to generate a report every day during non-working hours, in addition to this, the report must be approved with a different standard

Solution

We create a robot capable of navigating this particular application to obtain and process the correct information, generating the correct report every day.

Result

50% reduction in process hours
2-4 hours daily

Indicative AI Models & Methods

Indicative AI Models & Methods

Robotic Process Automation (RPA)
Natural Language Processing (NLP)
SQL Query Automation
Report Generation Automation

Implementation:

Easy

Indicative Effort:

2 months

Data Ingestion and Analysis

Our clients rapid growth has led to accelerated adoption of technology solutions for the monitoring and tracking of operational performance in the different oil fields spread throughout the US territory. The client engaged to solve the challenges inherent to ingesting, processing and leveraging the new big volumes of incoming data into the decision-making process.

Solution

We built a data ingestion platform for the field IoT devices to feed data to. We trained and deployed an AI engine that facilitates the dynamic estimation of nominal and abnormal ranges in the streaming data and via an API reports back to the monitoring solutions the optimal alarm thresholds, previously defined manually. We built a classification system that can be further improved into a predictive maintenance solution leveraging AI models.

Result

80% reduction in manual hours
2/3 cost reduction
More coverage of low yielding pumps
Faster recognition of anomalies or spills

Indicative AI Models & Methods

Indicative AI Models & Methods

Machine Learning for Dynamic Data Analysis
Data Ingestion and Processing Techniques
Classification Systems for Operational Data

Implementation:

High.

Indicative Effort:

200 – 250 days