<?xml version="1.0" encoding="utf-8"?><?xml-stylesheet type="text/xsl" href="atom.xsl"?>
<feed xmlns="http://www.w3.org/2005/Atom">
    <id>https://finsight-tech.com/blog</id>
    <title>Finsight Blog</title>
    <updated>2026-09-05T08:00:00.000Z</updated>
    <generator>https://github.com/jpmonette/feed</generator>
    <link rel="alternate" href="https://finsight-tech.com/blog"/>
    <subtitle>Finsight Blog</subtitle>
    <icon>https://finsight-tech.com/img/favicon-48x48.png</icon>
    <entry>
        <title type="html"><![CDATA[Introducing the Finsight Client Opportunity Assistant — A Palantir Accelerator for Capital Markets]]></title>
        <id>https://finsight-tech.com/blog/introducing-client-opportunity-assistant</id>
        <link href="https://finsight-tech.com/blog/introducing-client-opportunity-assistant"/>
        <updated>2026-09-05T08:00:00.000Z</updated>
        <summary type="html"><![CDATA[We're introducing the Finsight Client Opportunity Assistant, the first solution in Finsight's new Palantir Accelerators offering.]]></summary>
        <content type="html"><![CDATA[<p>We're introducing the <strong>Finsight Client Opportunity Assistant</strong>, the first solution in Finsight's new <strong>Palantir Accelerators</strong> offering.</p>
<p>Built on <strong>Palantir Foundry and AIP</strong>, the accelerator helps capital-markets teams identify and explain actionable client opportunities by connecting information that is often spread across multiple systems.</p>
<p>It brings together client context, portfolios and positions, market observations, firm inventory, product eligibility, preferences and restrictions, then applies governed business logic to identify opportunities such as reinvestment, hedging, concentration, liquidity, inventory match, new issues, relative value and cross-sell.</p>
<p>A salesperson can ask:</p>
<blockquote>
<p><strong>“Why should I speak with Phoenix today?”</strong></p>
</blockquote>
<p>The assistant resolves Phoenix to the appropriate client, evaluates the current opportunities and explains the supporting evidence.</p>
<p>It can also work across the broader client population:</p>
<blockquote>
<p><strong>“List the top 10 opportunities across all active clients today.”</strong></p>
</blockquote>
<p>The assistant can identify active clients, run opportunity discovery across them and produce a prioritised cross-client summary.</p>
<!-- -->
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="built-around-the-palantir-ontology">Built around the Palantir Ontology<a href="https://finsight-tech.com/blog/introducing-client-opportunity-assistant#built-around-the-palantir-ontology" class="hash-link" aria-label="Direct link to Built around the Palantir Ontology" title="Direct link to Built around the Palantir Ontology" translate="no">​</a></h2>
<p>The <strong>Palantir Ontology</strong> is central to the solution.</p>
<p>Rather than treating client, portfolio, position, instrument, market and control data as unrelated datasets, the Ontology represents them as connected business objects and relationships.</p>
<p>For example:</p>
<div class="language-text codeBlockContainer_Ckt0 theme-code-block" style="--prism-color:#393A34;--prism-background-color:#f6f8fa"><div class="codeBlockContent_QJqH"><pre tabindex="0" class="prism-code language-text codeBlock_bY9V thin-scrollbar" style="color:#393A34;background-color:#f6f8fa"><code class="codeBlockLines_e6Vv"><div class="token-line" style="color:#393A34"><span class="token plain">Client</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">  ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Portfolio</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">  ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Position</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">  ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Instrument</span><br></div></code></pre></div></div>
<p>alongside other business context such as:</p>
<div class="language-text codeBlockContainer_Ckt0 theme-code-block" style="--prism-color:#393A34;--prism-background-color:#f6f8fa"><div class="codeBlockContent_QJqH"><pre tabindex="0" class="prism-code language-text codeBlock_bY9V thin-scrollbar" style="color:#393A34;background-color:#f6f8fa"><code class="codeBlockLines_e6Vv"><div class="token-line" style="color:#393A34"><span class="token plain">Client</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> ├── Preferences</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> ├── Product Eligibility</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> ├── Restrictions</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> └── Activities</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain" style="display:inline-block"></span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Portfolio</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> ├── Exposures</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> └── Cash Flows</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain" style="display:inline-block"></span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Instrument</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> ├── Market Observations</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain"> └── Firm Inventory</span><br></div></code></pre></div></div>
<p>This connected semantic context gives the opportunity engine a consistent business view across multiple data domains.</p>
<p>The solution combines several Palantir capabilities:</p>
<div class="language-text codeBlockContainer_Ckt0 theme-code-block" style="--prism-color:#393A34;--prism-background-color:#f6f8fa"><div class="codeBlockContent_QJqH"><pre tabindex="0" class="prism-code language-text codeBlock_bY9V thin-scrollbar" style="color:#393A34;background-color:#f6f8fa"><code class="codeBlockLines_e6Vv"><div class="token-line" style="color:#393A34"><span class="token plain">Foundry data pipelines</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Palantir Ontology</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Ontology SDK</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Python Functions</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Governed opportunity logic</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">AIP Chatbot</span><br></div></code></pre></div></div>
<p>AIP can use <strong>Object Query</strong> to resolve business entities such as clients and <strong>Function tools</strong> to invoke the opportunity engine.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="a-code-driven-palantir-accelerator">A Code-Driven Palantir Accelerator<a href="https://finsight-tech.com/blog/introducing-client-opportunity-assistant#a-code-driven-palantir-accelerator" class="hash-link" aria-label="Direct link to A Code-Driven Palantir Accelerator" title="Direct link to A Code-Driven Palantir Accelerator" translate="no">​</a></h2>
<p>The Finsight Client Opportunity Assistant is not just a chatbot configuration or a standalone demonstration.</p>
<p>It is designed as a <strong>code-driven accelerator package</strong> containing reusable implementation assets for building this type of solution on Palantir.</p>
<p>The accelerator includes patterns for:</p>
<ul>
<li class="">canonical financial data;</li>
<li class="">Palantir Ontology object types and relationships;</li>
<li class="">governed Python Functions;</li>
<li class="">reusable opportunity rules;</li>
<li class="">product eligibility and client restriction controls;</li>
<li class="">AIP tool orchestration;</li>
<li class="">reference scenarios and validation data.</li>
</ul>
<p>The objective is not to provide a fixed application that assumes every financial institution works in the same way.</p>
<p>Instead, the accelerator provides a working starting point that can be adapted to an institution's own data sources, terminology, Ontology, business rules and controls.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="governed-business-logic-first-ai-second">Governed Business Logic First, AI Second<a href="https://finsight-tech.com/blog/introducing-client-opportunity-assistant#governed-business-logic-first-ai-second" class="hash-link" aria-label="Direct link to Governed Business Logic First, AI Second" title="Direct link to Governed Business Logic First, AI Second" translate="no">​</a></h2>
<p>A key design principle is that the AI does <strong>not</strong> decide whether an opportunity exists.</p>
<p>Deterministic business logic performs the qualification.</p>
<p>For example, the opportunity engine can evaluate whether:</p>
<ul>
<li class="">a position is approaching maturity;</li>
<li class="">an exposure has grown beyond a configured threshold;</li>
<li class="">a portfolio has become concentrated;</li>
<li class="">a material cash requirement is approaching;</li>
<li class="">firm inventory matches client preferences and passes controls;</li>
<li class="">a new issue is relevant and eligible;</li>
<li class="">a held instrument is showing a relative-value deviation;</li>
<li class="">a client exposure maps to an available product capability.</li>
</ul>
<p>AIP then provides the conversational layer around those governed results.</p>
<p>Its role is to:</p>
<ul>
<li class="">resolve the user's intent;</li>
<li class="">identify the relevant Ontology objects;</li>
<li class="">invoke governed Functions;</li>
<li class="">summarise the resulting evidence;</li>
<li class="">support the user with clear conversational explanations.</li>
</ul>
<p>In simple terms:</p>
<div class="language-text codeBlockContainer_Ckt0 theme-code-block" style="--prism-color:#393A34;--prism-background-color:#f6f8fa"><div class="codeBlockContent_QJqH"><pre tabindex="0" class="prism-code language-text codeBlock_bY9V thin-scrollbar" style="color:#393A34;background-color:#f6f8fa"><code class="codeBlockLines_e6Vv"><div class="token-line" style="color:#393A34"><span class="token plain">Governed business logic</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        =</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">determines WHAT opportunities exist</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain" style="display:inline-block"></span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">AI</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">        =</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">helps the user understand WHY they matter</span><br></div></code></pre></div></div>
<p>This separation makes the business logic explicit and testable while preserving the flexibility of conversational AI.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="eight-opportunity-patterns-in-the-current-reference-solution">Eight Opportunity Patterns in the Current Reference Solution<a href="https://finsight-tech.com/blog/introducing-client-opportunity-assistant#eight-opportunity-patterns-in-the-current-reference-solution" class="hash-link" aria-label="Direct link to Eight Opportunity Patterns in the Current Reference Solution" title="Direct link to Eight Opportunity Patterns in the Current Reference Solution" translate="no">​</a></h2>
<p>The current accelerator supports eight opportunity families:</p>
<p><strong>Reinvestment</strong>
Identify positions approaching maturity where a client conversation may be required.</p>
<p><strong>Hedging</strong>
Identify material exposure changes that may warrant a hedging discussion.</p>
<p><strong>Concentration</strong>
Surface portfolio exposures that exceed configured concentration thresholds.</p>
<p><strong>Liquidity</strong>
Identify significant upcoming liquidity requirements.</p>
<p><strong>Inventory Match</strong>
Match available firm inventory against client preferences while applying eligibility and restriction controls.</p>
<p><strong>New Issue</strong>
Identify relevant open new issues for eligible clients.</p>
<p><strong>Relative Value</strong>
Compare instruments already held by a client with maturity-comparable peers and identify meaningful yield deviations.</p>
<p><strong>Cross Sell</strong>
Connect material client exposures with relevant product capabilities available within the institution.</p>
<p>These are reference patterns rather than a fixed model for every institution. The rules and thresholds can be adapted to the institution's own products, business model and control framework.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="a-reusable-starting-point">A Reusable Starting Point<a href="https://finsight-tech.com/blog/introducing-client-opportunity-assistant#a-reusable-starting-point" class="hash-link" aria-label="Direct link to A Reusable Starting Point" title="Direct link to A Reusable Starting Point" translate="no">​</a></h2>
<p>The purpose of the accelerator is to shorten the path from Palantir platform capability to a working financial business solution.</p>
<p>Instead of starting from a blank implementation, a financial institution can begin with an established pattern covering:</p>
<div class="language-text codeBlockContainer_Ckt0 theme-code-block" style="--prism-color:#393A34;--prism-background-color:#f6f8fa"><div class="codeBlockContent_QJqH"><pre tabindex="0" class="prism-code language-text codeBlock_bY9V thin-scrollbar" style="color:#393A34;background-color:#f6f8fa"><code class="codeBlockLines_e6Vv"><div class="token-line" style="color:#393A34"><span class="token plain">Financial data</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">      ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Palantir Ontology</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">      ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Governed Python Functions</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">      ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">Opportunity intelligence</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">      ↓</span><br></div><div class="token-line" style="color:#393A34"><span class="token plain">AIP Assistant</span><br></div></code></pre></div></div>
<p>The data mappings, Ontology model, opportunity rules, controls and AI behaviour can then be adapted to the institution's own environment.</p>
<p>The same architecture can also be extended beyond the current Client Opportunity use case to other financial workflows that require multiple data sources, explicit business meaning, governed logic and AI-assisted interaction.</p>
<p>The <strong>Finsight Client Opportunity Assistant</strong> is the first solution in our <strong>Palantir Accelerators</strong> offering.</p>
<p><a class="" href="https://finsight-tech.com/solutions/client-investment-intelligence">Learn more about Client Investment Intelligence →</a></p>
<p><em>Finsight Technologies is an independent company. References to Palantir, Foundry and AIP describe technologies used in our solutions and do not imply endorsement or affiliation with Palantir Technologies.</em></p>]]></content>
        <author>
            <name>Finsight Technologies</name>
            <uri>https://finsight-tech.com</uri>
        </author>
        <category label="Announcement" term="Announcement"/>
        <category label="Palantir Accelerators" term="Palantir Accelerators"/>
        <category label="Governed Financial AI" term="Governed Financial AI"/>
    </entry>
    <entry>
        <title type="html"><![CDATA[Building in the Open: The AI-Native Financial Data Foundation Series]]></title>
        <id>https://finsight-tech.com/blog/ai-native-financial-data-foundation-series</id>
        <link href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series"/>
        <updated>2026-07-16T12:00:00.000Z</updated>
        <summary type="html"><![CDATA[Much of the thinking behind Finsight AI-Foundry did not start as a product. It started as a long-running research and engineering series — still ongoing, with new posts published regularly — written in public by our founder, Linxiao Ma, on Data Ninjago, under the category AI-Native Financial Data Foundation. Thirty posts and counting.]]></summary>
        <content type="html"><![CDATA[<p>Much of the thinking behind <strong>Finsight AI-Foundry</strong> did not start as a product. It started as a long-running research and engineering series — <strong>still ongoing, with new posts published regularly</strong> — written in public by our founder, Linxiao Ma, on <a href="https://dataninjago.com/" target="_blank" rel="noopener noreferrer" class="">Data Ninjago</a>, under the category <a href="https://dataninjago.com/category/ai-native-financial-data-foundation/" target="_blank" rel="noopener noreferrer" class="">AI-Native Financial Data Foundation</a>. Thirty posts and counting.</p>
<p>If you want to understand <em>why</em> AI-Foundry is designed the way it is — not just what it does — that series is the place to look, and the place to keep looking as the architecture evolves.</p>
<!-- -->
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="what-the-series-covers">What the series covers<a href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series#what-the-series-covers" class="hash-link" aria-label="Direct link to What the series covers" title="Direct link to What the series covers" translate="no">​</a></h2>
<p>The series works through the problem from first principles, in roughly the order the ideas matured:</p>
<h3 class="anchor anchorTargetStickyNavbar_Vzrq" id="1-modelling-financial-products-and-their-lifecycle">1. Modelling financial products and their lifecycle<a href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series#1-modelling-financial-products-and-their-lifecycle" class="hash-link" aria-label="Direct link to 1. Modelling financial products and their lifecycle" title="Direct link to 1. Modelling financial products and their lifecycle" translate="no">​</a></h3>
<p>The early posts dig into the <strong>ISDA Common Domain Model (CDM)</strong>: how a financial product is defined — its nature, economic terms, payouts and schedules — and how a trade then <em>lives</em>: <code>Trade</code>, <code>TradeState</code>, the primitive operators that change them, and the event model that captures the full lifecycle. One post walks an end-to-end novation to show how "before state plus instruction produces after state" gives you lineage and two distinct kinds of change for free.</p>
<h3 class="anchor anchorTargetStickyNavbar_Vzrq" id="2-semantics-as-a-graph">2. Semantics as a graph<a href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series#2-semantics-as-a-graph" class="hash-link" aria-label="Direct link to 2. Semantics as a graph" title="Direct link to 2. Semantics as a graph" translate="no">​</a></h3>
<p>The middle of the series asks how financial meaning should be <strong>modelled, stored and exposed</strong> so that machines can use it — modelling semantic relationships in Neo4j, querying a financial semantic graph with Cypher, and handling the messy reality that a source field <em>may</em> map to a semantic concept rather than cleanly matching one.</p>
<h3 class="anchor anchorTargetStickyNavbar_Vzrq" id="3-ai-over-governed-semantics--not-instead-of-them">3. AI over governed semantics — not instead of them<a href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series#3-ai-over-governed-semantics--not-instead-of-them" class="hash-link" aria-label="Direct link to 3. AI over governed semantics — not instead of them" title="Direct link to 3. AI over governed semantics — not instead of them" translate="no">​</a></h3>
<p>The most recent posts develop the case for an <strong>agentic semantic layer</strong>: the LLM as an <em>orchestrator over governed semantic capabilities</em>, rather than a standalone generator of answers. That includes reading notes on tool-using LLMs, GraphRAG and graph-grounded reasoning with evidence checking, plus a retrieval pattern for connecting imprecise user language ("fixed coupon") to governed field names (<code>FixedRate</code>).</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="from-blog-series-to-ai-foundry">From blog series to AI-Foundry<a href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series#from-blog-series-to-ai-foundry" class="hash-link" aria-label="Direct link to From blog series to AI-Foundry" title="Direct link to From blog series to AI-Foundry" translate="no">​</a></h2>
<p>The through-line of the series is exactly the architecture behind <a class="" href="https://finsight-tech.com/solutions/ai-foundry">Finsight AI-Foundry</a>:</p>
<ul>
<li class="">industry-aligned semantics (CDM, ontologies) as the starting foundation;</li>
<li class="">an institution-specific semantic layer connected to real systems and fields;</li>
<li class="">a graph runtime that makes meaning queryable;</li>
<li class="">evidence and lineage attached to every answer;</li>
<li class="">and AI used only where interpretation adds value — bounded, validated and governed.</li>
</ul>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="an-ongoing-series--follow-along">An ongoing series — follow along<a href="https://finsight-tech.com/blog/ai-native-financial-data-foundation-series#an-ongoing-series--follow-along" class="hash-link" aria-label="Direct link to An ongoing series — follow along" title="Direct link to An ongoing series — follow along" translate="no">​</a></h2>
<p>This is a live, evolving body of work, not an archive: the series continues as the AI-Foundry work continues, and new posts land as new parts of the architecture are designed, prototyped and tested. If you are exploring these problems in your own institution, start with the <a href="https://dataninjago.com/category/ai-native-financial-data-foundation/" target="_blank" rel="noopener noreferrer" class="">category page</a> and read in either direction — or <a class="" href="https://finsight-tech.com/contact">start a conversation</a> with us directly.</p>]]></content>
        <author>
            <name>Finsight Technologies</name>
            <uri>https://finsight-tech.com</uri>
        </author>
        <category label="AI-Native Financial Data" term="AI-Native Financial Data"/>
        <category label="Financial Semantics and Ontology" term="Financial Semantics and Ontology"/>
        <category label="Architecture and Implementation" term="Architecture and Implementation"/>
    </entry>
</feed>