LinkedIn

Twitter

GitHub

Resource Library

Your customers and users expect generative AI analytics capabilities in the apps you’re building and delivering. If you don’t deliver it, your competitors will.
<center>
<img class="aligncenter size-full wp-image-744474" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/11/Chat-AI-Gif-3-1.gif" alt="" width="600" height="465" />
</center>
According to a just-published Battery Ventures <a href="https://batteryventures.substack.com/p/5-key-takeaways-on-the-current-state">survey</a>, enterprise buyers are adopting AI and machine learning, finding that 79% of respondents plan to consume generative AI/LLMs within the next 12 months.
But simply adding GenAI to your app for the sake of it isn’t a solution. After all, it has to deliver real value. When I speak to product leaders and development teams who are on the path to building GenAI into their roadmaps, they are consistently asking the same questions:
<ul>
	<li>Where can we innovate and add real value using AI-first experiences??</li>
	<li>How can we deliver it responsibly and avoid <a href="https://www.ibm.com/topics/ai-hallucinations#:~:text=AI%20hallucination%20is%20a%20phenomenon,are%20nonsensical%20or%20altogether%20inaccurate.">hallucinations</a> that compromise user trust?</li>
	<li>How can we deliver it quickly while circumventing LLM lock-in risk?</li>
</ul>
The answer to the first question is immediate for many teams. Even before the emergence of GenAI and LLMs, a surefire way to engage users has been to embed analytics as a natural part of your product’s experience. Applications with thoughtful analytics experience have a significant increase in user engagement compared to those without. 
Analytics can enhance the user experience by providing personalized and interactive data visualizations in-app, which drives higher engagement levels. And it’s an area ripe for GenAI reinvention to <a href="https://sisense.com/platform/conversational-analytics/">shift from analytics clicks to conversations</a>—conversational analytics.
In this post, you’ll learn why the new era of conversational analytics, which is at the intersection of analytics and GenAI, will have profound implications for your roadmap.
I’ll share the five Cs of conversational analytics―consistent emerging themes for success that product and AI teams are thinking about as they look to adopt. Finally, I’ll delve into how Sisense is expanding <a href="https://sisense.dev/guides/sdk/">Compose SDK</a> with GenAI and share how you can start on the path today.
If you'd like to get a jumpstart, <a href="https://www.sisense.com/get/gen-ai-partner/">apply to become a beta partner of Sisense's GenAI components</a>—including conversation analytics capabilities, natural language generation narratives, and our Analytics Chatbot.
<h2>The evolution to conversational analytics</h2>
We know conversational AI systems and LLMs excel at understanding and contextualizing user inquiries. They can leverage previous interactions and user history to provide more personalized and relevant insights. This contextual understanding can now add a layer of intelligence to analytics, enabling more meaningful and tailored insights for each user. It provides a substantial leap over old-school first-gen search-based analytics that lack context and are unable to facilitate an ongoing conversation.
Conversational AI also provides a more intuitive and user-friendly interface for interacting with analytics. Instead of navigating through complex dashboards or writing specific queries, users can have an ongoing conversation with the AI to access the desired information, from asking questions to explaining answers. This improves the user experience, reduces the learning curve, and facilitates the broader adoption of analytics.
In fact, Eckerson Group sees conversational business intelligence as the next step in analytics to make it easier for users to engage with, beyond previous-generation visual analytics and search-based BI tools.
<center>
<img class="aligncenter wp-image-744463" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/11/1699279568767-1-300x172.png" alt="" width="600" height="344" />
</center>
Source: <a href="https://www.linkedin.com/posts/kevinpetrietech_bi-businessintelligence-generativeai-activity-7127295096743432192-I4Jb">Eckerson Group</a>
So whether you’re embarking on your analytics journey or looking to evolve the analytics you’re already providing, it’s time to consider how you will add GenAI data analytics to your apps.
<h2>A path to embedding generative AI analytics</h2>
When it comes to embracing this new era of analytics and being successful, there’s a fundamental path that product and engineering teams can follow. We call it The Five Cs of Conversational Analytics:
<h3>1. Conversational and contextual</h3>
Enable your users to quickly identify insights that are relevant to them. Start them off by suggesting natural language analytical starter prompts based on their data model like “What is my revenue by country?” Let them follow up by asking questions about their scope of analytics such as “What fields can I ask about?” Let the questions flow, such as enabling contextual follow-on prompts later in the conversation, such as asking for narrative guidance, like “Explain these results.” It’s about quickly delivering an end-to-end and ongoing conversational experience around data.
<h3>2. Composable</h3>
If conversational analytics looks and feels different from the rest of your app, that creates a roadblock to adoption. This is where composable analytics SDKs come in, providing development teams with a range of React Components (or other frameworks)—from visualizations to an AI analytics chatbot, as well as strong TypeScript/JavaScript language support for complete control. A composable SDK enables your product team to build the exact conversational experience and workflow for users and makes natural language more intuitive. 
<h3>3. Capable, responsible, and trusted</h3>
Trust is paramount regardless of how and where we engage with an LLM. For conversational analytics, a foundational way to ensure trust is to build it on a semantic layer or data model. It enables analytics delivery teams to define the scope, meaning, fields, and relationships that encompass the end-user analytics experience. Another important consideration is what controls are in place over data, prompts, and outputs—for example, are they fully isolated? Will they be used to improve and train the LLM model or shared in any other way?
<h3>4. Cloud and LLM agnostic</h3>
In AI, accelerating change is a constant. When you embed conversational analytics, it’s important that it doesn’t lock you into any one LLM—which may not be the best LLM to use for your use case in the future, or perhaps for your compliance needs, your organization may need to shift to use its own private LLM.  So, whether it is OpenAI GPT 3.5 to GPT-4, using Anthropic/Claude, Bard, LLaMa, or using your own private LLM (or just your own private OpenAI instance), it pays to be cloud- and LLM-agnostic.
<h3>5. Cost</h3>
Finally, with any discussion around LLMs, it’s also critical to consider cost. Ultimately, it’s about flexibility, because managing cost requires the flexibility to plug and play different LLMs from public to in-house models; in this way, you remain in control as LLM costs change and evolve across different providers. It also means you get the flexibility to exact-fit the most cost-effective LLM for balancing, scalability, affordability, and output quality.
<h2>Sisense Compose SDK now with GenAI-powered analytics built-in</h2>
We launched Compose SDK earlier this year to enable software engineers and developers to integrate and embed analytics capabilities powered by Sisense Fusion into any web application, utilizing modern frameworks such as React and TypeScript or their preferred alternatives. Its composable and modular design ensures developers can add analytics in a flexible and versatile manner to suit their specific needs.
Now, we are expanding Compose SDK to make it incredibly easy for product teams to responsibly incorporate generative AI analytics into their products, their user experience, and workflows.
<h3>Introducing the Analytics Chatbot</h3>
Bring the power of Generative AI analytics directly into your product with ease. Think of the <a href="https://sisense.com/platform/conversational-analytics/">Analytics Chatbot</a> as an embeddable LLM-powered sidekick, a whole chatbot experience you can rapidly add to your app with complete control. It’s provided as a React Component and enables your users to begin their conversation with pre-generated QuickStart natural language questions, or if they prefer, begin with their own specific questions. 
The Analytics Chatbot makes it simple for your users to ask questions around the data model, such as fields and metrics they can query, ask analytics questions, get visual answers, or generate detailed narratives for data storytelling based on results.
<center>
<img class="aligncenter wp-image-744464" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/11/AI-Chat-300x233.png" alt="" width="600" height="467" />
</center>
Better yet, the Analytics Chatbot is completely composable. Let’s say you want to control how it works with your broader user experience. Your development team can extend and customize it, using it in tandem with the rest of your app. 
For example, you might want to use Analytics Chatbot as the natural language starting point for your users, extending it with a “pin-to-palette” so they can establish their own personalized, NLQ-driven analytics home page in your app. The possibilities are endless.
<center>
<img class="aligncenter wp-image-744466" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/11/GenAI-Hero-Img-1024x775.png" alt="" width="600" height="454" />
</center>
&nbsp;
<h3>Enable your users to interpret answers easily with GenAI-powered Narratives</h3>
What if you could easily add explanations for any of the analytics in your app? Now you can, with Narratives, another GenAI-powered Compose SDK React Component. With Narratives, it’s easy for your users to get textual explanations that provide insight into key findings within any data set. 
These explanations are generated using natural language generation techniques and supported by advanced algorithms to produce clear and concise narratives that highlight key insights, trends, and anomalies. Narratives is also available as an API within Compose SDK for headless use.
<center>
<img class="aligncenter wp-image-744465" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/11/Insights-1024x542.png" alt="" width="600" height="317" />
</center>
<h3>Lego-like composability for conversational control</h3>
In fact, because Compose SDK is composable from the ground up, every conversational aspect is available as a discrete React Component, including QuickStart questions, Narratives, and Data Topics (that enable users to select a data model to converse with). And, of course, all of the functionality is available as an API if you’re looking to weave GenAI-powered analytics deep into your application.
<h3>Designed for trusted and responsible development</h3>
It may seem common sense that for your users to confidently act on the GenAI-powered analytics answers they get from your application, they must trust those answers. But it’s a crucial factor as you build out your roadmap around conversational analytics. 
This is the reason why all of the GenAI capabilities we’ve built into Compose SDK are created on top of the Sisense data model, our semantic layer. It gives your analytics team the power to ensure that your data is optimally designed for the right queries and enabled to go from analytics answers to in-context narratives. The Sisense data model includes fields, metrics, formulas, and relationships and is all governed by centralized row and column-level security, so everyone is chatting with the same, managed single version of the truth.
<h3>GenAI for analytics: create a roadmap that flows with change</h3>
Whether you start with OpenAI today, are looking to perhaps bring your LLM in-house in the future, or even considering switching gears to LLaMa 2, it’s important to be able to plug and play as new LLMs or needs emerge. 
In just the same way that Sisense is cloud and database agnostic, whether you want to run Sisense on AWS, Azure, or GCP or use RedShift, Google BigQuery, or another database for analytics, we’re designing Compose SDK’s GenAI capabilities to be LLM fluid too. 
We aim to enable you to deliver the best GenAI-powered analytics experience while enjoying complete back-end and architecture flexibility, from database to LLM. It means you can build for today but embrace what tomorrow will bring.
<h2>Turn your generative AI analytics ideas into reality</h2>
It’s an incredibly exciting time to be a Sisense customer. Transformation in analytics is at hand. If you’re as excited about what this could mean for your product roadmap, I invite you to join our webinar: <a href="/webinars/gen-ai-in-context-analytics-data-monetization/">GenAI, in-context analytics, and the future of data monetization</a>.
Joined by Paul Turner, Founder &amp; CEO of Skyview Consulting, we’ll discuss the emerging conversational analytics landscape, highlight practical strategies to seamlessly incorporate these analytics into applications, and explore how to maintain trust in AI-driven analytics.

How to drive analytics adoption with generative AI

How to Meet the Demand for Analytics Across the Organization

<h2>Understanding AI-powered analytics with LLMs</h2>
Since the arrival of OpenAI's ChatGPT, there has been an explosion of Large Language Models (LLMs) that have transformed perceptions of what’s possible with Artificial Intelligence. A wide range of industries and use cases are adopting LLMs for their ability to drive productivity and deliver innovation across businesses. The pace of innovation has been profound, with models such as OpenAI’s GPT-4 and GPT-5, Google Gemini, Cohere, LLaMa, PaLM, Claude, and a vast array of emerging domain-specific models.
The first wave of LLM adoption focused on chat-based interactive conversational experiences, helping to summarize information, synthesize new content, and assist development. Now, there is a new opportunity: using LLMs to supercharge analytics data processes. This can enable data engineering teams to provide richer data and new ways for analytics end-users to slice, dice, and visualize information, all while reducing effort in the data pipeline. In this whitepaper, you’ll learn more about this opportunity, and how to reimagine your analytics approach by infusing LLMs to unlock the next level of value.
LLMs offer support for a vast array of use cases, including automating coding, generating customer service responses, providing language translation, and much more. An article in MIT Sloan Management Review recognizes the complementary nature of machine learning-based advanced analytics and the natural language capabilities of LLMs. It highlights opportunities for generative AI to address challenges in both the <a href="https://sloanreview.mit.edu/article/how-generative-ai-can-support-advanced-analytics-practice/">development and deployment phases of advanced analytics</a>, encompassing predictive and prescriptive applications.
Automating enrichment and augmentation of analytics with new fields, segmentations, and classifications is a challenge that analytics professionals and operations teams face daily. Often, they find themselves grappling with hours of data wrangling or caught in a last-minute rush to enhance and supplement data when analytics end-users demand it urgently.
LLMs can play a pivotal role here. Acting as an on-demand resource for augmented information, they have the potential to alleviate much of the pain and effort traditionally involved in data preparation. In doing so, they empower teams to segment and dissect data in innovative ways, in real-time.
<table style="width: 39.1486%; border-collapse: collapse; border-style: solid; border-color: #000000; height: 177px;">
<tbody>
<tr>
<td style="width: 1194px; text-align: left;">Estimates suggest that 70 to 80% of analysts’ time is spent on preparing data for analytics. LLMs can act as an on-demand resource for augmented information, having the potential to alleviate much of the pain and effort traditionally involved in data preparation.</td>
</tr>
</tbody>
</table>
However, applying an on-demand model is only feasible if the LLM you are using is tightly integrated with your data and analytics tools. Without this integration, the ease of self-service analytics diminishes rapidly. The good news is that it’s now possible to seamlessly integrate LLMs like OpenAI GPT to significantly boost your analytics. It’s easy to do, and we think you’ll find it’s a game-changer.
<h2>Leveraging AI-powered analytics with LLMs for efficient data preparation </h2>
Today, teams often spend way too much time in manual data prep and enrichment tasks. Estimates suggest that business and data analysts may spend as much as <a href="https://www.nytimes.com/2014/08/18/technology/for-big-data-scientists-hurdle-to-insights-is-janitor-work.html">70 to 80% of their time preparing data for analytics</a>, with a significant portion dedicated to enriching and labeling data and generating new attributes to facilitate analysis. Introducing new fields to datasets for analysts to manipulate is often a laborious and challenging process. For example, consider a seemingly simple query:
What’s the breakout of my customers by industry?
It’s the kind of question that’s critical to making decisions regarding entering new markets and launching new products. However, what if the thousands of customers in the dataset weren't segmented by industry in advance? What if "Industry" isn't a field in the data or data model? For an analyst, this usually implies that you're out of luck unless you can invest considerable time and effort.
It means going back to the data, adding external classification information to the data set, and reloading it in your data warehouse, and that’s incredibly time-consuming. In many cases, it requires manually matching with and reconciling third-party data sets, repetitive Google searches to tag customer records manually, and often painful, convoluted IF () and VLOOKUP () formulae in Excel or custom Python script. Not to mention updating your data model. 
Worse, if teams subsequently need to classify and label data in a new way–such as by zip code or county, then it’s back to manually prepping the data again.
<h2>Advantages of integrating AI-powered analytics</h2>
An LLM is exceptionally well-suited for this kind of work, and it can do this kind of search, extraction, and reconciliation automatically and quickly. However, selecting the right moment to weave LLM-sourced information into your analytics experience is critical to making it truly useful. For example, if you collect this information during ETL, both storage and query costs will be high. Collect it too late, and you may struggle to insert it into the data model to enable the end-user's expected BI functions.
The optimal approach involves collecting the requested data on demand and utilizing the LLM to suggest the ideal insertion point in a model, offering a highly responsive experience for the end-user. With this approach, end-users gain access to their own AI-powered, always-on personal data prep and data analyst bot as part of their analytics experience. They can easily extend and augment their data and data model by simply engaging with ChatGPT.
Using AI to assist in data preparation creates a massive opportunity for teams. While the LLM does the manual work of prepping and augmenting data, they are free to spend time on more value-added and strategic work. But more importantly, beyond freeing up time, there’s transformational value to be had―enabling teams to be much<a href="https://www.sisense.com/platform/conversational-analytics/"> more fluid in their analytical queries</a>, segmenting data in new ways, with the LLM doing the work behind the scenes. 
<h2>Augmented analytics integration: Leveraging AI-powered analytics with Sisense and LLMs</h2>
A reference implementation of this strategy is available for Sisense Fusion, and is a perfect example of how to incorporate an LLM like OpenAI GPT into analytics processing. This implementation utilizes the flexible capabilities of Sisense Fusion to inject dynamic queries to OpenAI GPT as part of the query process. It automatically generates new database tables and relationships within the existing data model while leveraging guidance from GPT.
This reference architecture also demonstrates low-code app experience development, creating an interactive Q&amp;A app via Sisense BloX. End-users can directly instruct GPT on how they want to augment the data set and model within their dashboards. Sisense and GPT then handle the rest.
To demonstrate how it works, consider the earlier question: “What’s the breakout of my customers by industry?”—when the "industry" attribute isn't present in the original data model or data set. Sisense and GPT can be used to fill this gap.
Below is the example data with a list of accounts. Ultimately, we’re looking for the percentage split of customers by industry even though that attribute doesn’t exist in the data.

[caption id="attachment_741530" align="aligncenter" width="1024"]<img class="wp-image-741530 size-large" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-4-1024x920.jpg" alt="A table entitled “Big Accounts” with 4 columns. The source data for enhanced LLM-driven data." width="1024" height="920" /> Sample data list of accounts, excluding the attribute of “industry.” End-users can directly instruct the GPT on how they want to augment the data set and model within their dashboards, and then Sisense and GPT handle the rest.[/caption]

To make it easy for the end-user to ask questions of GPT, Sisense BloX is used to build a simple interactive widget on a Sisense dashboard. This new UI assists in capturing what the end-user wants to request from GPT:

[caption id="attachment_741528" align="aligncenter" width="1024"]<img class="wp-image-741528 size-large" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-2-1024x352.jpg" alt="A user chat interface that asks the end-user to “please be specific what you request,” and has a field for the user to type in their data question. " width="1024" height="352" /> This Sisense interface enables end-users to specify their requests for GPT, naturally and interactively.[/caption]

In this example, GPT is told the end-user wants the industry segmentation for a list of companies in natural language. GPT is given the dashboard, widget, and column name, which ultimately contains the list of companies that need industry segment labels. 
The UI executes a Python script that contains the question and a dynamically populated list of customer accounts formatted as JSON. This script is submitted to GPT, which then retrieves the industry for each customer. Alternatively, if we want to add GPT enrichment to the scheduled periodic data model refresh, simply executing Python rather than using the Q&amp;A UI, we can do that too. 
The result set is passed back from GPT to Sisense Fusion as a JSON document. The Python script creates new tables in the origin database based on the GPT result set and adds new relationships between tables in the model. Ultimately, it reduces data enrichment efforts from hours or days to mere seconds.
<table style="width: 37.0618%; border-collapse: collapse; border-style: solid; border-color: #000000; height: 123px;">
<tbody>
<tr>
<td style="width: 1194px; text-align: left;">Ultimately, leveraging AI-powered analytics with Sisense reduces data enrichment efforts from hours or days to mere seconds.</td>
</tr>
</tbody>
</table>
In the example, GPT returns each customer and their associated industry label in JSON. Then, the script appends this as a new dimension to our data model and automatically joins our fact table of wins and losses with a natural join based on the account name. 

[caption id="attachment_741529" align="aligncenter" width="1024"]<img class="wp-image-741529 size-large" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-3-1024x833.jpg" alt="A table with 2 columns. The output, or LLM-driven data, is now categorized by “industry.”" width="1024" height="833" /> LLM-enhanced data. Using a GPT reduces data enrichment efforts from hours or days to mere seconds.[/caption]

With the augmented data added to our data model, it’s simple to continue our analysis. For example, we can quickly create two new visualizations using Sisense Fusion that show us the distribution of our customers by industry.

[caption id="attachment_741527" align="aligncenter" width="859"]<img class="wp-image-741527 size-full" src="https://sisense2dev.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-1.png" alt="A pie chart of LLM-driven data, showing the distribution of customers by industry." width="859" height="353" /> Created using Sisense Fusion and a GPT, this pie chart shows the distribution of our customers by industry.[/caption]

he whole integration allows analyses to be completed in seconds, rather than waiting days or weeks for an operations team to prepare and publish new information. With the <a href="https://www.sisense.com/blog/sisense-blox-go-beyond-dashboards/">interactive BloX widget</a>, powered by GPT, augmenting and enriching any data using natural language is simple and efficient.
A software product can easily add GPT-supercharged analytics by embedding this BloX widget directly into the user interface. This integration provides end-users with greater flexibility by allowing the addition of augmented or ancillary data without requiring any involvement from the product vendor. As a result, this extends the product's value and enhances its data capabilities. 
This approach to on-demand data isn’t necessarily easy to implement in a business intelligence product. However, the Python compute kernel and the extensive API surface in Sisense Fusion allow direct query creation and dynamic updates to data models to get this kind of on-demand end-user experience. That said, just about any ETL engine can at least augment data at ingestion using this technique. Sisense’s reference architecture for GPT gives you a good starting point to try these techniques out for your own use cases. Here’s a <a href="https://www.sisense.com/blog/how-to-make-bi-and-analytics-adoption-simple/">link </a>to an embedded example of the exact analysis we shared. 
<table style="width: 38.7312%; border-collapse: collapse; border-style: solid; border-color: #000000; height: 134px;">
<tbody>
<tr>
<td style="width: 100%;">With the <a href="https://www.sisense.com/blog/sisense-blox-go-beyond-dashboards/">interactive BloX widget</a>, powered by GPT, augmenting and enriching any data using natural language is simple and efficient.</td>
</tr>
</tbody>
</table>
<h2>Next steps in AI-powered analytics: Implementing strategies with LLMs</h2>
The potential of LLMs has now expanded to supercharging analytics data processes, enabling data engineering teams to provide richer data and empowering analytics end-users to explore data in new and dynamic ways. Here are the steps to take to infuse AI into your data processes. 
<ol>
	<li style="font-weight: 400;" aria-level="1">Identify pain points. Evaluate your current data preparation and enrichment processes to identify pain points and areas where LLMs can be leveraged for automation and augmentation.</li>
	<li style="font-weight: 400;" aria-level="1">Draft a plan. <a href="https://www.sisense.com/get/gen-ai-partner/">Develop a plan</a> for incorporating LLM-powered augmented analytics into your existing processes and BI tools. Consider potential use cases where LLMs can significantly reduce manual effort and add strategic value to your analytics approach.</li>
	<li style="font-weight: 400;" aria-level="1">Careful LLM selection. Research and select LLMs that closely align with your use case and integrate seamlessly with your data and analytics tools. Explore reference implementations, such as the one provided for Sisense Fusion.</li>
	<li style="font-weight: 400;" aria-level="1">Test for integration. Test the integration of an LLM with your analytics tool, using the Sisense reference architecture for GPT as a starting point. Ensure that the LLM can provide on-demand augmented data and seamlessly integrate with your data model.</li>
</ol>
At Sisense, we’re working with companies every day that are at the forefront of AI and analytics, combining technologies to deliver powerful value. To learn how you can get started, <a href="https://www.sisense.com/get/book-demo/">schedule a conversation and demo</a> with one of our AI and analytics experts today.

Resource Library

‌

‌

‌

‌

‌

‌

‌

‌

‌

‌

‌

‌