For now, it doesn’t. We’ll be transitioning the TKE database to the SavvyWise platform over the coming months. This will begin with a pilot program, including subscriber testing to ensure the new platform is working as expected.
We’ll keep you in the loop and let you know when changes are coming.
Subscribers who only require traditional keyword, topic and faceted search to browse the library or filter articles by author and date will receive a modernised, high-speed search platform at zero additional cost or change to their current subscription. You can continue renewing your TKE subscription through The Tax Institute as usual.
For subscribers seeking deterministic AI reasoning, this capability is being introduced as an optional add-on via the SavvyWise platform. Members who choose to utilise these advanced features can access them by subscribing to SavvyWise directly to fund the underlying reasoning and compute infrastructure, ensuring practices only pay for the specific level of research capability they require.
This additional functionality is not available just yet, but watch this space!
The SavvyWise system establishes trust by always showing its working, ensuring practitioners never have to rely on an unverified output. Every legal conclusion decomposes into the specific rules that produced it, tracing directly back to the exact legislative provisions, ATO rulings, expert commentary or TKE articles that form its foundation. This transparent architecture allows a user to follow the chain of reasoning back to the source and verify the logic independently, just as they would when checking a colleague's analysis.
Because a practitioner relying on research to advise a client carries the ultimate professional responsibility for that advice, this verifiable traceability is critical. Research that explicitly shows its logic supports the high standards members are held to and enables firms to build efficient review processes, as a senior adviser can simply check the system's reasoning rather than re-performing the research from scratch.
No, the platform does not replace professional judgement and is built specifically so that it cannot be mistaken for doing so. The Tax Knowledge Exchange and SavvyWise assist users with finding authoritative-backed answers but explicitly leaves the final judgement to the practitioner.
You should always review any results or insights and never rely on AI to provide client advice without reviewing carefully.
No, SavvyWise does not acquire copyright or ownership of any papers.
IP ownership will not change under the agreement. The Institute retains its rights and control over the resources. The copyright remains with the content’s original authors and the platform is built specifically to preserve author attribution.
The Tax Institute remains the sole owner and custodian of the Tax Knowledge Exchange database as a product.
SavvyWise operates as a specialised technology vendor and data processor, analogous to the Institute engaging an enterprise cloud provider or a third-party database host to store and index member files.
The Institute is simply exercising its existing right to publish and communicate papers to TKE subscribers through a modern technology infrastructure partner rather than its legacy software vendor.
The platform protects author integrity by maintaining a strict separation between the system's legal reasoning and the author's original commentary.
When a paper is ingested, it retains its precise metadata, including the author's name, publication date and statutory context. Rather than feeding professional content into a large language model and returning an artificial intelligence generated summary that strips away nuance, the deterministic system directs a member straight to the specific section of the article that answers their question. The author's reasoning arrives entirely intact, complete with any qualifications and framing they originally chose to place around it.
Furthermore, because tax commentary naturally ages as legislation and rulings evolve, historical papers are never surfaced as absolute statutory truth in the present. Instead, the formal model makes the relationship to current law visible, explicitly contextualising the paper alongside the current legislation so that an author's professional standing is never exposed by stale analysis being mistakenly presented as active law.
Our contributors help power the work we do at the Institute and help power the profession – this alliance will ensure your contribution creates even greater value for our community.
A contributor’s paper remains an integral part of The Tax Institute’s educational library, fully accessible to Tax Knowledge Exchange subscribers through an overhauled, high-performance interface. This allows our authors to continue sharing their expertise with our community.
Through advanced technological infrastructure and search capability, the SavvyWise platform also allows contributors’ work to be discovered by a broader range of people, growing author’s profile and enriching the profession as a whole.
Search results link directly back to the original paper, ensuring the author's name remains explicitly attached to the value delivered.
Yes, contributors are entitled to decline having their work included in the SavvyWise platform.
Our contributors have the right to understand and control how their contribution is used. If you are a contributor and would like to opt out or would like further information, please contact us with a written request.
Subscribers rely on the TKE database as their first stop for robust, reliable tax technical resources that have been produced by fellow tax practitioners. We want all TKE subscribers to have the best experience possible.
Working with SavvyWise to power a new, advanced TKE search allows our subscribers to get more value from the TKE database.
Partnering with SavvyWise allows the Institute to leverage world-leading AI technology and a platform purpose-built for Australian tax practitioners to bring these benefits to our subscribers on a quicker timeline, with more support, than would otherwise be possible.
The current SavvyWise platform is a purpose-built research tool designed specifically for Australian tax practitioners.
It operates on a sophisticated Retrieval-Augmented Generation system that uses legal-domain specific embeddings alongside mainstream large language models (LLMs).
Unlike standard, general-purpose LLMs that predict text sequences and risk confident but incorrect answers, the SavvyWise system grounds its responses in real Australian tax legislation, ATO rulings and exclusive expert commentary from highly regarded tax lawyer, Adrian Cartland. This ensures practitioners have a highly capable search tool that significantly outperforms unassisted generative artificial intelligence.
While the current system is highly effective, SavvyWise is actively developing a major upgrade for release later this year to address the intrinsic limitations of LLMs, such as hallucinations and unverified reasoning.
The platform is transitioning to a world-first deterministic reasoning engine. With this new architecture, every legal conclusion decomposes into the specific rules that produced it, allowing practitioners to trace the exact legislative provision, ruling or expert source behind an answer.
The phased rollout will begin with the small business CGT concessions. Over the next twelve months, this coverage will expand to include the full breadth of Australian federal taxes relevant to SMEs.
This transition toward verifiable, structured legal logic allows TKE subscribers to conduct comprehensive research in the TKE database, which they can then verify for accuracy and appropriateness before putting it into practice for clients.
Unlike standard artificial intelligence tools that utilise statistical text prediction algorithms to try to answer questions, deterministic search relies on a structured, verified map of legal rules and relationships. This formal model is built by actual tax practitioners who work through the legislation to identify what each section requires and express that structure so it can be applied consistently. The system effectively records the logical analysis a senior practitioner performs when reading through an unfamiliar provision, capturing statutory definitions, conditional tests, the exact order of application and the priority between conflicting rules.
Because the system evaluates queries against this established framework rather than predicting what an answer might look like, asking the same question a thousand times will yield the identical answer a thousand times. This delivers the absolute consistency and transparent logic required for professional reliance and accurate research.
Standard large language models do not maintain a database of tax law; they generate responses by predicting text sequences that appear statistically plausible based on patterns in their training data. Because plausibility and legal accuracy frequently diverge in tax matters, standard models can generate confident, fluent answers that contain flawed reasoning or completely fabricated legislative citations.
The SavvyWise deterministic reasoning engine eliminates this risk by removing statistical text prediction from the process of determining legal conclusions. Instead of generating prose, the system evaluates the practitioner's query directly against a verified, structured map of statutory provisions, definitions and rule interactions. Every conclusion decomposes back to the exact section or ruling that establishes it, meaning legislative references are verified sources rather than predicted text. Outside the specific areas of law that have been formally represented within the structured model, the platform behaves as a precise search and retrieval tool rather than attempting to manufacture an unverified answer.
Tax Knowledge Exchange content will be used in the ongoing development of the deterministic reasoning engine by providing verified legal analysis to help structure the formal map of tax rules. In the future, it will also train highly specialised small language models tailored specifically to assist with tax practitioner workflows to correctly interpret the precise professional vocabulary and structural nuances of Australian tax law.
However, it’s important to note that while TKE content is used to teach these engines the nuances of tax language and law, the models will be structured such that they cannot memorise, retain or embed any TKE content. That means that in order to draw on TKE insights, the SavvyWise platform will always need to draw from the original TKE database.
