We’re very excited about the rollout of the DeepSee platform in the bank. While large scale trading has historically faced limitations using interconnected systems prone to latency and human error, DeepSee now allows us to analyze and process functions in minutes where it used to take hours.

— Managing Director, Tier 1 Bank

Opportunity

In the capital markets industry, data has never been so valuable. Financial services is one of the most data-intensive sectors in the economy, with companies generating and storing structured and unstructured data in massive volumes. At the same time, an array of artificial intelligence (AI) and machine learning models are being researched, enhanced, developed, and trained at incredible rates.

This wealth of data and the data science to harness it presents a prime opportunity for financial services companies to radically transform their operations—optimizing efficiency and efficacy, while reducing risk and maximizing revenue.

All the data and AI power in the world, however, offers little business value unless the data science can be “operationalized” to reduce cost, mitigate risk, and and increase customer satisfaction to address today’s complex business challenges.

Challenge

One of the world’s largest banks was faced with high error rates in reviews, unacceptably slow processing times, and large percentages of manual controls for regulatory compliance when reconciling securitized and OTC derivatives trades.

Analysis

The process for post-trade reconciliation for derivatives has four key stages:

  1. Workflow Identification: Identify which trades should be reviewed using a 10-step, manual process performed by agents on a T+30 basis.
  2. Transaction Assignments: Transactions are assigned to agents by template training for review.
  3. Document Sourcing: Agents source the corresponding pricing supplement, term sheet, XML and any other documentation required to review.
  4. Review Transactions: Agents conduct manual review of assigned transactions using multiple tracking tools and raise any escalations via email.

Each of these stages offered opportunities for improvements through DeepSee’s end-to-end automation platform, which is a SaaS solution currently delivered on AWS.

AI-Powered Transformation with DeepSee

After implementing the DeepSee platform, the bank was able to reduce its post-trade review time by 30x, automate a significant portion of its regulatory obligations, and increase its Straight-Through-Processing (STP) times by 80%. Additionally, after it was fully deployed, the platform increased the bank’s capacity to handle 10x more volume with the same personnel.

Average Review Time
Per Transaction:

BEFORE: 60-90 MIN

WITH DEEPSEE: 2-3 MIN

30xTime Reduction

Average Daily Transaction Throughput:

BEFORE: 6 TRANSACTIONS

WITH DEEPSEE: 60 TRANSACTIONS

10xMore Volume

Average Review
Time Period:

BEFORE: T+30 DAYS

WITH DEEPSEE: T+5 DAYS

80%Faster Turn

Results

Increased Throughput

Reduced Review Time

Better Accuracy

Mitigated Risk