Introducing Rule Builder: a new feature designed by – and for – compliance and risk experts

Over the past few years we’ve built considerable in-house compliance and engineering expertise in how to automate complex financial advice assessments.

And now we’re previewing a new feature called Rule Builder that makes it easy for non-engineers to create, test, and run custom audit rules that run on the TIQK platform. Audit rules automate compliance and risk assessments on the information written in financial advice documents.

Rule Builder will also enable our clients to selectively extract useful data points from each file audit, saving them into the TIQK data analytics platform. This will allow financial services licensees to perform slice-and-dice trend analysis on advice quality and financial adviser behaviour – going well beyond an individual file audit outcome.

Our compliance analysts are already using it internally with impressive results. Once battle-hardened, we plan to make this feature available to clients and consultants to build their own rules on-demand.

Until then here’s a sneak peek showing just how easy it is to build, test, and deploy a common financial advice test using Rule Builder.

Rule Builder on desktop computerWhy Rule Builder?

TIQK is a unique combination of natural language processing algorithms; an expert (“rules-based”) system; and artificial intelligence.

The platform features a library of more than 650 pre-built audit rules, built by our compliance analysts and engineers. They automate a wide range of regulatory and Best Interest tests and quickly flag risky advice and behaviours in written financial advice.

Clients and consultants regularly ask us to add custom rules to match their internal policies and methodologies. And we’ve heard loud and clear that they’re losing patience with software vendors that take weeks or months to customise their platforms.

Using the expertise we’ve gained over the past few years, we designed Rule Builder to make it easy for our internal teams, clients, and consultants to create, test, and run automated tests on written financial advice – in just minutes, not weeks or months.

Doing an Asset Allocation Variance Test manually

An Asset Allocation Variance Test checks that the split of investments recommended by a financial planner in a Statement of Advice (SoA) file is in their client’s best interest. If there’s a variance to an accepted standard split, the rationale should be explained in the file.

An auditor might take the following approach when manually reviewing an SoA:

  • If the SoA lists a Domestic Equity Assets ‘Target vs. Recommended’ value of less than 10%: mark it as compliant
  • If the value is more than 10% but a justification has been included by their financial planner: it might be compliant, but further review is warranted
  • If the value is more than 10% and there’s no justification: mark it as non-compliant

The auditor will look through the file for data in tables …

Asset Allocation Variance test - tabular data in SoA

… and for information in natural-language commentaries (potentially edited by the financial advisor):

Asset Allocation Variance test - commentary data in SoA

While straightforward, manual reviews like this are time consuming. Worse, it can be challenging or impossible to manually review the hundreds or thousands of Statement of Advice files generated by an organisation’s financial advisers.

Automating it with Rule Builder

Using Rule Builder we can quickly build an Asset Allocation Variance Test rule that runs automatically across all the Statement of Advice files audited by an organisation using TIQK.

Start by adding information like a name and description for the Rule, including how it will appear in the on-screen audit report and downloadable PDF.

You can also elect to run this rule for advice documents created in a specific date range. This is because TIQK automatically identifies the date an advice file was prepared by an Adviser:Rule Builder - options

Optional: Add relevant obligations

You can list legislation, regulations, regulatory guides, or internal business policies that relate to your rule. This helps the person reviewing an audit report understand why the rule is important, and the consequences of non-compliance. It’s easy to add internal business policies to the system, and TIQK can automatically create links to online references for more information:

Rule Builder - obligations

Describe the Possible Outcomes

Each Audit Rule outputs a result during a file audit (like “compliant”, “non-compliant”, “advice”, “high risk”, “low risk”) based on the logic tests that you create and the contents of the file being audited. We call the range of potential results “Possible Outcomes”.

In this case and following the logic that a human auditor would follow for the Asset Allocation Variance Test, our Possible Outcomes are:

  • Compliant: if the client’s “target vs. recommended” value for “domestic equity” < 10%
  • Advice: if the client’s “target vs. recommended” value for “domestic equity” >= 10% and there’s an explanation related to “overweight domestic equities”
  • Non-compliant: if the client’s “target vs. recommended” value for “domestic equity” >= 10% and there’s NO explanation related to “overweight domestic equities”

Creating these tests is done visually using point-and-click. Behind the scenes Rule Builder automatically creates highly efficient and safe code that stays perfectly synchronised with your visual design:

Click to view full-sizeRule Builder - outcomes

 

Chips: powerful access to document data during an audit

The items inside the logic of each Possible Outcome are called “Chips”:

Rule Builder - Chip - Domestic Asset Allocation

Chips are powerful components that let you tap into the data extraction and natural language capabilities of TIQK. They enable the rule to obtain and work with data from the file being audited. Working with a Chip is easy: just click it to reveal its capabilities and modify its settings for your needs:Rule Builder - find in table

We’ve built chips for the most common functionality, and the list is growing daily:

  • Extract data from tables with pinpoint accuracy
  • Find text in natural-language commentary
  • Custom “variables” – values that you can use in multiple rules e.g. an Asset Allocation Variance threshold = 10%
  • Custom “functions” – logic (created using the same point-and-click interface) that you can embed into other rules, making it possible to create complex, multi-step compliance and risk tests
  • Mathematical operators – like add, subtract, multiply, divide, average, min, and max
  • Comparison operators – like less/greater than, less/greater than or equal to, equal to, not equal to
  • Logical (boolean) operators – like AND, OR, and NOT
Test and deploy

Finally, we know how important it is to test before you deploy, so we’ve made it simple.

Just upload a sample document and tap the Test button. Rule Builder shows you which Outcome is returned based on your logic tests and the contents of the test document:Rule Builder - test

It will even flag any logic problems:Rule Builder - syntax error

Once you’re happy with the rule, click the ‘Enabled’ option to deploy it instantly. (Don’t worry: your custom rules remain private to your organisation.)

The next time you run a document audit in TIQK, the newly-enabled rule will be run automatically alongside TIQK’s pre-built rules. Its results will also be merged into the on-screen and PDF audit report.

What’s next?

Today Rule Builder gives TIQK’s internal team an efficient way to build, test, and deploy audit rules in minutes, rather than weeks or months.

Over time we’ll be exposing more of TIQK’s advanced natural language and artificial intelligence capabilities inside the Rule Builder tool, powering even more complex risk assessments:

  • Information about the clients, and their goals and objectives;
  • Information about the financial adviser, like their qualifications and authorisations;
  • Information about the financial strategies and products recommended;
  • Information from the client’s Fact Find records; and
  • Aggregated organisation audit data analytics and trends.

We’ll also be offering the ability to selectively save extracted data points and information collected during a rule’s operation into the TIQK data analytics platform. This will enable our analysts and clients to conduct slice-and-dice trend and behavioural analysis across their organisation’s entire written advice output.

Like to know more?

Rule Builder is currently in use as an internal tool by TIQK compliance analysts. We’d love to hear your feedback.

If you’d like to know more, reach out to our Customer Success team.