Steps Ahead

Do more with less: Harnessing the power of AI and automation to create value in your FICC business

Current market conditions have created a turbulent period for banks across the globe. For those that have remained profitable, a laser-sharp focus on efficiency has become central to their businesses. Sales and trading teams must be equipped with tools to improve productivity and efficiency in a cost-effective manner against a backdrop of cost cutting and headcount slimming.

Specifically, client facing sales teams are challenged with bringing a higher and improving participation rate from their clients and trading teams are responsible for executing that flow. But this is no small feat when resources are scarce. Thankfully, AI and automated data analytics technology can change the ways in which FICC-trading investment banks can engage with their clients and help them do more with less.

Automation = profit

As McKinsey stated in a recent report, to capitalise on the opportunities of AI, investment banks “can make targeted investments that will drive productivity tailored to their unique client and product franchises. They will need to pay particular attention to choosing the right gen AI use cases, building capabilities that can scale, and managing the associated risks.”

With the right AI-driven analytics platform in place, delivered by a specialised partner, banks can understand and measure the variables driving their objectives and create value across their FICC desks by seeing more business, winning more of the business they see, retaining more of the profit and reducing cost and risk.

Taking interest rate swaps as an example, how can an automated, AI-driven data analytics solution turn raw, unstructured data into an actionable insight and help FICC traders and salespeople do more with less?

The following diagram shows the automated process of a trade being booked on Mosaic, through to being transformed into a piece of actionable insight that can lead to client outreach:

Example use case: automating IRS flow


Client Input


  • trade_date: 2021-09-10
  • trade_date_time: 2021-09-10T11:01:46.000Z
  • packageId: 751
  • trade_venue: STW
  • total_no_leg: 3
  • order_status: Traded
  • leg_id: 751_1
  • side: Receive
  • price_type: SWAP RATE
  • price: 1.755
  • cover_price: 1.754
  • nominal_amount: 47000000


  • sales_person_id: ABCD
  • sales_person_name: ABC ABCD
  • trader_id: EDCBA
  • trader_name: EDCBA
  • portfolio: EURSWAPS
  • desk_name: APAC DESK


  • cpty_name: ABCDE HK FUND
  • cpty_parent_name: ABCDE
  • cpty_country_code: HK
  • cpty_sector: HEDGE FUND
  • cpty_lei: 99
  • cpty_tier: TIER 1
  • cpty_trader_name: EDC CBA


  • isin: USD30YSWAPBB
  • FloatLegRef: Libor 3m
  • PaymentFrequency: Semi-annual
  • Daycount: 30/360
  • settlement_date: 2023-01-14
  • maturity_date: 2053-01-12
  • swap_type: FIXED VS FLOAT
  • currency: USD


Aggregation and Normalisation

  • The disparate data sources from the client are joined together and normalised using a proprietary MSX data model.

Step 3

Data Quality

This layer is for data validation. Making sure to filter out any:

  • fat fingers
  • trades with missing key fields
  • non-client specific validation logic compliant trades e.g., a client wants to filter out all
  • trades from a specific trade desk/platform



We use the normalised client input, with sometimes including market data, to calculate some additional risk metrics e.g.,:

  • negotiationStrategy: SWAP BUTTERFLY
  • negotiationStrategyType: 3 LEG
  • negotiationDV01: 141676
  • legDistanceToCoverBps: 0.234
  • legSpreadCapturedBps: 1.065


Storage / Presentation

The enriched data is stored in our low latency interactive database which can handle queries over large multiple terabyte datasets (also used by Netflix). On top of the database sits our high performance pivot query / search engine that handles both the real-time and historic queries to and from the UI.


Reporting and Alerts

  • Rates Briefing Reports (GBP, SCANDI, JGB, USD, EUR)
  • Numerical Anomaly
  • Relative Value
  • Franchise Trends
  • Franchise Anomaly
  • Seasonality
  • Categorical Anomaly
  • Custom Threshold

With analytics automated and running 24/7, value-added functionality such as real-time alerts and automated reporting can be implemented, giving the bank a reason to contact its client with a targeted IRS opportunity – an action that is beneficial for both bank and client. And this is just one example of an automated transaction – imagine the benefits once automation is applied across the entire FICC business.

Value: created

In summary, Mosaic’s powerful automated technology creates value in the following ways:

To find out more about Mosaic’s ability to enable your FICC business to do more with less, contact us today to arrange a demo.