Automate personal finance management (e.g. loan repayments by personal agents). This might be a consumer-downloadable app that syncs with their banking products to support achievement of personal financial goals. This might include automating savings for consumer by forecasting wallet availability based on customer spending habits and preferences. A typical use case deployment is a "rounding up" function on a digital payment - so a $3.63 payment triggers a round up to $4.00, automatically depositing 37 cents in a specified savings account.

Evaluate customer credit risk using application and other relevant data to improve accuracy and reduce unconscious bias in real-time underwriting decisions. Signals may vary - for example signs of hesitation whilst form filling may indicate a higher propensity for fraud.

Evaluate and advise the most appropriate product and supplier to meet consumer need (e.g. mortgages). This should improve consumer outcomes.

Conduct "digital human" conversational interface with customers in physical locations (e.g. banks such as Natwest, soul machines). The aim is to provide effectively online functionality via some form of automated, robotic interface.

Scan social media to discover references to product and competitors for product management purposes. This may include customer sentiment analysis on key product or service attributes.

Optimise retail network locations based on multiple signals of demand (e.g., social data, footfall, transactions). This would - for example - help a retailer to plan their expansion in to a new market. Alternatuvely this might enable cost savings across a retail banking operation where it would likely cover both branches and ATMs - at the risk of medium to long term revenue loss and potential negative customer and press reaction.

Assess asset valuations (e.g. mark-to-market, regulatory capital assessment etc) for both internal risk and market control but also to ensure compliance with regulatory testing regime.

Use market and external indicators to predict risk of corporate distress and potential bankruptcy. Data sources will vary but may include personal data trails left by key directors.

Using cameras in, for example, retail sites to categorise potential customer's demographic data - typically sex and age. This potentially raises customer response, ethical and legal issues, unless carefully managed.

Identify customer emotional sentiment in branch using facial expression analysis. This may be misleading as cause and effect are not always clear cut and the risk of false positives is high.

Determining threshold level at which to flag compliance risks with the aim to minimise both 'false positives' and 'false negatives' both of which carry a cost

Monitor and analyse interactions with customers to flag compliance risks. This could include phone, messaging and email interactions, as well potentially as on 3rd party platforms (such as social media).

Optimise resource allocation to support branch and ATM network. Factors that may be open to alternative deployment and focus include staffing, cash delivery trucks, opening hours or mobile assets

Analysing payment transactions to flag potential fraudulent activity - which is typically automatically blocked and usually requires human intervention to then unblock.

Invoice redirection fraud takes place when criminals contact a business claiming to be from one of their suppliers, saying that they have changed bank and requesting that an invoice is paid into a different account.

Predict risk of loan delinquency and recommend proactive maintenance strategies. Machine learning can respond swiftly to dynamic information sources.

Detect potentially fraudulent or nefarious users (e.g. individuals under sanctions, investigation etc) through pattern matching of structured and unstructured data on transactions from different sources, such as phone numbers, addresses, company directors and news reports (HSBC, Quantexa, Silent Eight)

Evaluate customer credit risk using application and other relevant data for faster and more efficient decisions - frequently with a turnaround time measured in seconds. This offers the potential for traditionally-underserved customer groups to be offered products.

Manage brand reputation by scanning social media to discover references to the company or products or key individuals. This may be supported with sentiment analysis - although this can potentially be misleading (e.g. 'humorous' memes can be misinterpreted).

Identify and categorise known customers when they enter physical retail site to ensure prioritised service (e.g. alerting relationship management team to premier customers). This has to be carefully (and legally) managed to ensure customer sensitivities around product deployment,.

Automate collection of banking data from non-standard documentation (e.g. in shipping paperwork) - this will typically be turned in to standardised banking documentation

Automate credit risk profiling to support fundraising by small businesses through crowdfunding platform (Funding Circle). This enables reduction in borrowing costs for eligible SMEs.

Automate and simplify data identification, aggregation, extraction, verification and capture involved in KYC (Know Your Customer) process for banks.

Automate reconciliation of financial statements of related legal entities - typically used for sophisticated and complex businesses operating across multiple legal authorities. This matters for providing an audit trail.

Analyse credit worthiness of under-banked individuals to provide banking services. Typically this involves understanding and monitoring behavioural patterns.

Analyse legal documents and extract important data points and clauses - potentially to inform further workflows or simply to focus legal experts on the most relevant areas of work.

Recognise documents (eg handwriting) to extract information to streamline functions like account creation, loan and insurance origination and documentation. This is especially useful with high volume consumer products.

Determine whether there are patterns of activity that suggest money laundering and compliance risks. Investment in Anti Money Laundering (AML) increasingly a regulatory focus (and cost) for banks operating in international systems.

Tailor debt collection processes by identifying which practices are most effective for different segments of customers. This is a sensitive process with occasional risks.

Enhance risk modelling and stress testing by discovering interactions in the financial system. This increases model robustness and strengthens it. Transparency may be an issue as the interlocked drivers become harder and harder to explain -although this has potential regulatory issues.