<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>RD Platform Archives - Reflective Data</title>
	<atom:link href="https://reflectivedata.com/category/platform/feed/" rel="self" type="application/rss+xml" />
	<link></link>
	<description></description>
	<lastBuildDate>Thu, 08 Oct 2026 09:38:53 +0000</lastBuildDate>
	<language>en-US</language>
	<sy:updatePeriod>
	hourly	</sy:updatePeriod>
	<sy:updateFrequency>
	1	</sy:updateFrequency>
	<generator>https://wordpress.org/?v=7.1.5</generator>

<image>
	<url>https://reflectivedata.com/wp-content/themes/reflective-data/img/icon-512.png</url>
	<title>RD Platform Archives - Reflective Data</title>
	<link></link>
	<width>32</width>
	<height>32</height>
</image> 
	<item>
		<title>Case Study: How Barbora Got One Version of the Truth and Cut Marketing Data Warehouse Costs by About 75%</title>
		<link>https://reflectivedata.com/case-study-barbora-semantic-layer-ai/</link>
					<comments>https://reflectivedata.com/case-study-barbora-semantic-layer-ai/#respond</comments>
		
		<dc:creator><![CDATA[Jason Dolan]]></dc:creator>
		<pubDate>Thu, 08 Oct 2026 09:38:53 +0000</pubDate>
				<category><![CDATA[Case Study]]></category>
		<category><![CDATA[Data infrastructure]]></category>
		<category><![CDATA[RD Platform]]></category>
		<category><![CDATA[case study]]></category>
		<category><![CDATA[dbt]]></category>
		<category><![CDATA[e-commerce]]></category>
		<category><![CDATA[MCP]]></category>
		<category><![CDATA[semantic layer]]></category>
		<guid isPermaLink="false">https://reflectivedata.com/?p=44196</guid>

					<description><![CDATA[<p>Barbora, one of the largest e-commerce businesses in the Baltics, built a trusted semantic layer on our platform. Metrics finally match, marketing warehouse costs dropped by about 75%, and AI now works on the same governed data in Slack and Claude.</p>
<p>The post <a href="https://reflectivedata.com/case-study-barbora-semantic-layer-ai/">Case Study: How Barbora Got One Version of the Truth and Cut Marketing Data Warehouse Costs by About 75%</a> appeared first on <a href="https://reflectivedata.com">Reflective Data</a>.</p>
]]></description>
										<content:encoded><![CDATA[<p>Barbora is one of the largest e-commerce businesses in the Baltics. Like many fast-growing retailers, it had plenty of data and plenty of dashboards, but its marketing numbers rarely agreed with each other. We helped Barbora build one trusted data foundation on the Reflective Data platform, and then put AI on top of it.</p>
<h2>At a glance</h2>
<ul>
<li><strong>Client:</strong> Barbora, a large e-commerce business in the Baltics</li>
<li><strong>What we built:</strong> a dbt project and semantic layer on our platform, interactive dashboards in Explore and a custom MCP server</li>
<li><strong>Result:</strong> metrics that finally match across sources, about 75% lower data warehousing costs in the marketing department, and AI access to the same trusted data in Slack and Claude</li>
</ul>
<h2>The challenge</h2>
<p>For a long time, the numbers coming from different sources did not match. The same question about marketing performance could produce different answers depending on which report, tool or query you used. This is a familiar problem in e-commerce, where the data comes from many places and every platform defines its metrics a little differently.</p>
<p>The effects were practical. Marketing and leadership spent time debating which number was right instead of deciding what to do next. Analysts spent hours reconciling reports. And with every new reporting need, queries against the warehouse grew, along with the bill.</p>
<p>Barbora also wanted to use AI to get answers faster. But pointing AI at inconsistent, unmodelled data would only have produced faster confusion. The foundation had to come first.</p>
<h2>The solution</h2>
<p>We built the solution in three steps, all on the Reflective Data platform.</p>
<h3>1. A dbt project and semantic layer</h3>
<p>We built Barbora&#8217;s <a href="/dictionary/dbt/">dbt</a> project on our hosted dbt, so every metric and dimension is defined once, in code, tested and version-controlled. A <a href="/platform/semantic-layer/">semantic layer</a> sits on top, so a number means the same thing wherever it appears. The models also produce compact, aggregated reporting tables, so reports no longer need to scan large volumes of raw data.</p>
<h3>2. Interactive dashboards in Explore</h3>
<p>On top of that layer we built a set of interactive dashboards using <a href="/platform/explore/">Explore</a>. Teams can look at the same governed metrics themselves, filter and slice them, and no longer depend on a one-off report for every question.</p>
<h3>3. An MCP server for their data</h3>
<p>We then built an <a href="/platform/mcp/">MCP server</a> that gives AI tools access to the same governed data through the semantic layer. Barbora&#8217;s people now use it directly in Slack and in Claude, and it also powers their weekly stakeholder emails. The answer in Slack, the number in a dashboard and the figure in the email all come from the same definitions.</p>
<h2>The results</h2>
<h3>Metrics that finally match</h3>
<p>The biggest win was one that had been missing for years: the numbers from different sources now match. Barbora&#8217;s marketing team and leadership were delighted to see it. Discussions moved from &#8220;which number is right?&#8221; to &#8220;what should we do about it?&#8221;.</p>
<h3>About 75% lower warehousing costs in marketing</h3>
<p>Because reports and queries now read modelled, aggregated tables instead of raw data, they are far more efficient. Barbora saved about 75% on data warehousing costs in the marketing department.</p>
<h3>AI where the team already works</h3>
<p>With a trusted data layer in place, AI became useful instead of risky. People get answers in Slack and Claude, and stakeholders receive weekly emails powered by the same governed data, without anyone assembling them by hand.</p>
<h2>Why it worked</h2>
<ul>
<li><strong>Foundation first.</strong> We fixed the definitions and the models before adding dashboards or AI, so everything built on top inherited the same logic.</li>
<li><strong>One layer for every tool.</strong> Dashboards, Slack, Claude and email all read from the semantic layer, so there is only one version of each number.</li>
<li><strong>Efficient by design.</strong> Reporting tables mean queries touch far less data, which is better for speed and for cost.</li>
</ul>
<p>Want to know more about the approach? Read <a href="/semantic-layer-ai-access/">why a semantic layer is essential for giving AI access to your data</a>.</p>
<h2>Get the same for your business</h2>
<p>If your numbers do not match across tools, or you want AI you can trust on your data, we can help. <a href="/contact/">Request a demo</a> of the platform, or <a href="/contact/">talk to one of our experts</a> about our <a href="/services/data-build/">data build and semantic layer</a> and <a href="/services/ai/">AI services</a>.</p>
<p>The post <a href="https://reflectivedata.com/case-study-barbora-semantic-layer-ai/">Case Study: How Barbora Got One Version of the Truth and Cut Marketing Data Warehouse Costs by About 75%</a> appeared first on <a href="https://reflectivedata.com">Reflective Data</a>.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://reflectivedata.com/case-study-barbora-semantic-layer-ai/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Why a Semantic Layer Is Essential for Giving AI Access to Your Data</title>
		<link>https://reflectivedata.com/semantic-layer-ai-access/</link>
					<comments>https://reflectivedata.com/semantic-layer-ai-access/#respond</comments>
		
		<dc:creator><![CDATA[Jason Dolan]]></dc:creator>
		<pubDate>Thu, 08 Oct 2026 07:58:00 +0000</pubDate>
				<category><![CDATA[RD Platform]]></category>
		<category><![CDATA[Technical]]></category>
		<category><![CDATA[agents]]></category>
		<category><![CDATA[AI]]></category>
		<category><![CDATA[data modelling]]></category>
		<category><![CDATA[dbt]]></category>
		<category><![CDATA[semantic layer]]></category>
		<guid isPermaLink="false">https://reflectivedata.com/?p=44175</guid>

					<description><![CDATA[<p>AI can write SQL, but it cannot know what your numbers mean. A semantic layer defines every metric once, so people, dashboards and AI agents all get the same, trustworthy answer.</p>
<p>The post <a href="https://reflectivedata.com/semantic-layer-ai-access/">Why a Semantic Layer Is Essential for Giving AI Access to Your Data</a> appeared first on <a href="https://reflectivedata.com">Reflective Data</a>.</p>
]]></description>
										<content:encoded><![CDATA[<p><strong>In short:</strong> Ask five people at your company for last month&#8217;s revenue and you may get five different numbers. An AI agent will happily give you a sixth. A semantic layer fixes this by defining every metric once, in code, so people, dashboards and AI agents all get the same answer. If you want AI to work with your data, this is the layer to build first.</p>
<h2>The problem: AI does not know what your numbers mean</h2>
<p>Large language models are very good at writing SQL. What they cannot know is what your business means by &#8220;revenue&#8221;, &#8220;active customer&#8221; or &#8220;qualified lead&#8221;. Is revenue gross or net of refunds? Does it include tax? Which orders count as completed? Is a customer someone who signed up, or someone who paid?</p>
<p>Those answers live in people&#8217;s heads, in old dashboards and in scattered queries. When an AI agent is pointed at raw tables, it has to guess. Sometimes it guesses right. Often it produces a confident, well-formatted answer that is quietly wrong, and nobody notices until the number reaches a board slide.</p>
<h2>What a semantic layer is</h2>
<p>A <a href="/dictionary/semantic-layer/">semantic layer</a> is a governed description of your business on top of your data models. It defines:</p>
<ul>
<li><strong>Metrics:</strong> how each measure is calculated, such as revenue, conversion rate or churn</li>
<li><strong>Dimensions:</strong> the ways you slice them, such as date, channel, country or product</li>
<li><strong>Entities and relationships:</strong> how customers, orders and sessions connect</li>
<li><strong>Ownership and permissions:</strong> who is responsible for a definition and who may see the data</li>
</ul>
<p>Here is what a metric definition can look like in code (illustrative):</p>
<pre><code>metric: revenue
type: simple
measure: sum(order_total)
filter: status = 'completed'
owner: finance
used_by: [dashboards, notebooks, agents]</code></pre>
<p>It is written once, reviewed in git like any other code, and used everywhere. Common ways to build one include <a href="/dictionary/dbt/">dbt</a> with MetricFlow, LookML, SQLMesh and Omni, and we work with all of them.</p>
<h2>Why AI specifically needs it</h2>
<h3>Consistent answers</h3>
<p>Without a semantic layer, the same question phrased two ways can produce two different queries and two different numbers. With one, the agent does not write the calculation. It asks for the metric by name and the layer produces the correct query. The answer matches your dashboards because it comes from the same definition.</p>
<h3>Fewer hallucinations</h3>
<p>A model choosing from a short list of defined metrics and dimensions has far less room to invent a join, pick the wrong column or misread a field. The space of possible mistakes shrinks from &#8220;anything in the warehouse&#8221; to &#8220;something the business has already agreed on&#8221;.</p>
<h3>Trust you can verify</h3>
<p>When an executive asks &#8220;where does this number come from?&#8221;, the answer should not be &#8220;the AI said so&#8221;. With a semantic layer, every figure traces back to a named metric, an owner and a tested model. Agents can show their working, and people can check it.</p>
<h3>Security and access control</h3>
<p>Giving an agent raw warehouse access also gives it everything in the warehouse, including data it should not show. A semantic layer is a natural control point: you expose only the metrics and dimensions the user is allowed to see, and the same rules apply whether the request comes from a person, a dashboard or an agent.</p>
<h3>Speed and cost</h3>
<p>Agents that explore raw event tables scan far more data than they need, which is slow and expensive. Queries generated from a semantic layer run against modelled, optimised tables, so answers come back faster and cost less.</p>
<h3>One layer for every tool</h3>
<p>Your analysts use BI, your marketers use spreadsheets, your product team uses notebooks and now your AI assistants use chat. A semantic layer gives all of them the same definitions. Add a new tool or a new AI model later and it inherits the existing logic instead of starting from scratch. This is also why the <a href="/reflective-data-ai-era/">Reflective Data agent and MCP server</a> read from a semantic layer and not from raw sources.</p>
<h2>What goes wrong without one</h2>
<p>We see the same patterns again and again when teams connect AI to unmodelled data:</p>
<ul>
<li>Revenue, customers and conversion rate calculated differently in every dashboard, and now a different way again in the chat assistant</li>
<li>Agents that quietly include test orders, bot traffic or duplicate events</li>
<li>Answers that change from one day to the next because the model made different assumptions</li>
<li>No way to audit or explain a number once it has left the tool</li>
<li>A pilot that impressed everyone in the demo and was switched off after the first wrong answer in front of leadership</li>
</ul>
<h2>In practice: how we built it for Barbora</h2>
<p>Barbora is one of the largest e-commerce businesses in the Baltics. They wanted their teams and their AI tools to work from the same trusted numbers, so we built the foundation described above, step by step, on the Reflective Data platform:</p>
<ol>
<li><strong>A dbt setup on our platform.</strong> We built their dbt project and semantic layer on our hosted dbt, so their business logic and metric definitions are written once, tested and version-controlled.</li>
<li><strong>Interactive dashboards in Explore.</strong> On top of that layer we built a set of interactive dashboards using the platform&#8217;s Explore feature, so teams can look at the same governed metrics themselves.</li>
<li><strong>An MCP server for their data.</strong> We then built an MCP for them that gives AI tools access to the same governed data, through the semantic layer.</li>
</ol>
<p>Today the whole company uses it where they already work. People ask questions in Slack and directly in Claude, and the same layer powers their weekly stakeholder emails. Because everything reads from one set of definitions, the answer in Slack, the number in a dashboard and the figure in the email all match.</p>
<p>That is the real payoff of a semantic layer. You do not build one integration per tool. You build the foundation once, and every new way of using your data inherits it.</p>
<h2>How to build a semantic layer that AI can use</h2>
<ol>
<li><strong>Start with clean models.</strong> Raw data is staged, tested and documented in dbt, so the layer sits on a foundation you trust.</li>
<li><strong>Pick the metrics that matter.</strong> Begin with the dozen or so numbers leadership and teams use every week. Do not try to model everything on day one.</li>
<li><strong>Agree on definitions with the business.</strong> Finance, marketing and product often disagree. Settling it once, in writing, is most of the value.</li>
<li><strong>Define them in code.</strong> Metrics, dimensions and relationships live in version control, with review, tests and an owner.</li>
<li><strong>Add descriptions for humans and machines.</strong> Clear names and descriptions help AI choose the right metric. Good documentation improves the answers.</li>
<li><strong>Connect your tools.</strong> BI, notebooks and AI agents all read from the same layer, for example through an MCP server.</li>
<li><strong>Monitor and improve.</strong> Track which questions are asked, where answers are wrong, and extend the layer accordingly.</li>
</ol>
<h2>How we can help</h2>
<p>Building this layer is what our <a href="/services/data-build/">data build and semantic layer</a> service is for. We design and build the models, work out the metric definitions with your stakeholders, and connect the layer to your BI tools and AI agents. If you already have a semantic layer in one tool, we can <a href="/services/bi-reporting/semantic-layer-build/">improve it or migrate it</a>. If you would rather not run it yourself, our platform includes <a href="/platform/">hosted dbt and a semantic layer</a>, and our team can run it for you as a <a href="/services/managed-platform/">managed service</a>.</p>
<p>Once the foundation is in place, our <a href="/services/ai/">AI services</a> help you put it to work: a company brain, custom agents, bots for your website or Slack, and training for your team.</p>
<p>Want to see how this would look on your data? <a href="/contact/">Request a demo</a> or <a href="/contact/">talk to one of our experts</a>.</p>
<p>The post <a href="https://reflectivedata.com/semantic-layer-ai-access/">Why a Semantic Layer Is Essential for Giving AI Access to Your Data</a> appeared first on <a href="https://reflectivedata.com">Reflective Data</a>.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://reflectivedata.com/semantic-layer-ai-access/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>How Reflective Data Is Built for the AI Era, and How We Help You Get There</title>
		<link>https://reflectivedata.com/reflective-data-ai-era/</link>
					<comments>https://reflectivedata.com/reflective-data-ai-era/#respond</comments>
		
		<dc:creator><![CDATA[Jason Dolan]]></dc:creator>
		<pubDate>Wed, 07 Oct 2026 07:15:07 +0000</pubDate>
				<category><![CDATA[RD Platform]]></category>
		<category><![CDATA[agents]]></category>
		<category><![CDATA[AI]]></category>
		<category><![CDATA[dbt]]></category>
		<category><![CDATA[MCP]]></category>
		<category><![CDATA[semantic layer]]></category>
		<guid isPermaLink="false">https://reflectivedata.com/?p=44155</guid>

					<description><![CDATA[<p>We use AI to ship faster, our platform has a full agent mode and an MCP server, and AI reads from a governed dbt and semantic layer instead of raw data. Here is why, and how we help you get there.</p>
<p>The post <a href="https://reflectivedata.com/reflective-data-ai-era/">How Reflective Data Is Built for the AI Era, and How We Help You Get There</a> appeared first on <a href="https://reflectivedata.com">Reflective Data</a>.</p>
]]></description>
										<content:encoded><![CDATA[<p><strong>In short:</strong> AI is only as good as the data you give it. At Reflective Data we use AI to ship product features faster, we built an agent mode and an MCP server on top of our platform, and we give AI access to data through a governed dbt and semantic layer, never through raw connections. Here is how that works, why we do it this way, and how we help our clients do the same.</p>
<h2>AI changes how fast we build</h2>
<p>We are a team of data engineers, and we use AI every day. It helps us write and review code, draft and test <a href="/dictionary/dbt/">dbt</a> models, generate documentation and prototype new product features in a fraction of the time it used to take. That is why new capabilities in the <a href="/platform/">Reflective Data platform</a> reach customers faster than they used to.</p>
<p>The important part is what stays human. Our engineers design the architecture, review every change and own what ships. AI speeds up the work. It does not replace the judgement.</p>
<h2>Talk to your data: agent mode and MCP</h2>
<p>The <a href="/platform/#explore">platform</a> now has a full <a href="/dictionary/ai-agent/">agent</a> mode. Instead of writing SQL or hunting through dashboards, you ask a question in chat and the agent works with all of your data to answer it. It can:</p>
<ul>
<li>answer questions across your sources, from marketing and product to finance and operations</li>
<li>produce clear, good-looking charts and visualisations on request</li>
<li>explain how a number was calculated, so people can trust it</li>
</ul>
<p>You do not have to come to us to use it. The Reflective Data <a href="/dictionary/mcp/">MCP server</a> brings the same governed data to the tools where you already work, such as your AI assistant, your editor or your internal agents. Your data goes where your people are.</p>
<h2>Why AI should read from dbt and a semantic layer, not raw sources</h2>
<p>This is the most important design decision for any company that wants AI to work with its data. There are three tempting shortcuts, and each one causes problems.</p>
<h3>1. Raw API connections and source-platform MCPs</h3>
<p>It is easy to connect an AI assistant straight to Google Analytics, a CRM or an ads platform. But every platform has its own definitions of users, sessions, conversions and revenue, and they rarely match. Sampling, attribution rules, API limits and quotas change what you get back. An agent that queries each source separately will happily give you numbers that disagree with each other, and with finance, without telling you why.</p>
<h3>2. Raw warehouse exports</h3>
<p>Pointing AI at raw BigQuery exports looks more robust, because the data is complete. In practice raw tables are large, nested, undocumented and full of edge cases: duplicate events, test traffic, bots, schema changes and fields that only an insider can interpret. The model has to guess the business logic on every query, so answers vary from one question to the next. It is also slow and expensive, because the agent scans far more data than it needs.</p>
<h3>3. What works: dbt plus a semantic layer</h3>
<p>With <a href="/dictionary/dbt/">dbt</a>, your business logic is written once, as tested and documented code. The <a href="/dictionary/semantic-layer/">semantic layer</a> sits on top of it and defines your metrics, dimensions and relationships: what counts as revenue, an active customer or a qualified lead. When AI asks a question, it uses those definitions instead of inventing its own. That gives you:</p>
<ul>
<li><strong>Consistent answers.</strong> The same question returns the same number, whether a person or an agent asks it.</li>
<li><strong>Trust and traceability.</strong> Every figure traces back to a tested model.</li>
<li><strong>Control.</strong> You decide which data AI can see, and permissions apply the same way everywhere.</li>
<li><strong>Lower cost and better speed.</strong> The agent queries modelled, aggregated tables instead of scanning raw events.</li>
<li><strong>Less hallucination.</strong> A model that picks from defined metrics has far fewer chances to make something up.</li>
</ul>
<p>This is exactly how our platform is built. <a href="/data-pipelines/">Pipelines</a> load your data, hosted dbt with a semantic layer models it, and Explore, agent mode and the MCP server read from that governed layer.</p>
<h2>How we help you become AI-driven</h2>
<p>Most companies do not need another AI tool. They need their data in a state AI can use, and a plan for putting it to work. Our consultants take you through that step by step:</p>
<ol>
<li><strong>Assess.</strong> We <a href="/services/audits/">audit</a> your tracking, data sources and reporting to find what blocks reliable AI answers.</li>
<li><strong>Build the foundation.</strong> We set up clean data collection, pipelines into your warehouse, dbt models and a <a href="/services/bi-reporting/semantic-layer-build/">semantic layer</a>, as part of our <a href="/services/data-infrastructure/">data infrastructure</a> work. This is the part that makes everything else trustworthy.</li>
<li><strong>Give AI access safely.</strong> We expose your governed data to agents and assistants through the semantic layer and MCP, with permissions and logging.</li>
<li><strong>Build your company brain and agents.</strong> We connect your documents and knowledge into a <a href="/services/ai/company-brain/">company brain</a>, and build <a href="/services/ai/custom-harness/">custom agent harnesses</a> and <a href="/services/ai/bots/">bots</a> for your website or Slack that follow your own processes.</li>
<li><strong>Train your team.</strong> We run hands-on <a href="/services/ai/ai-training/">AI training</a> sessions based on your real workflows, so people use AI well and safely from day one.</li>
<li><strong>Run it.</strong> If you prefer, our <a href="/services/managed-platform/">managed service</a> operates the platform and keeps the data, models and agents healthy.</li>
</ol>
<p>We work with all major AI labs (see our <a href="/services/ai/">AI services</a>) and stay vendor-neutral, so you are never locked into one model. You can use the Reflective Data platform, or we can build on the stack you already have.</p>
<h2>Where to start</h2>
<p>If you are curious about what an AI-ready data setup would look like for your business, we are happy to talk it through. <a href="/contact/">Request a demo</a> of the platform to see agent mode on real data, or <a href="/contact/">talk to one of our experts</a> about the right first step for your team.</p>
<p>The post <a href="https://reflectivedata.com/reflective-data-ai-era/">How Reflective Data Is Built for the AI Era, and How We Help You Get There</a> appeared first on <a href="https://reflectivedata.com">Reflective Data</a>.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://reflectivedata.com/reflective-data-ai-era/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
	</channel>
</rss>
