LinkedIn

Twitter

GitHub

Resource Library

<h1>How to drive analytics adoption with generative AI</h1>
<p>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.</p>
<center>
<p><img class="aligncenter size-full wp-image-744474" src="https://sisense.wpengine.com/wp-content/uploads/2023/11/Chat-AI-Gif-3-1.gif" alt="" width="600" height="465" /></p>
</center>
<p>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.</p>
<p>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:</p>
<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>What are the keys to ensuring strong compliance and security?</li>
	<li>And one more: How can we deliver it quickly while circumventing LLM lock-in risk?</li>
</ul>
<p>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. </p>
<p>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.</p>
<p>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. </p>
<p>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.</p>
<h2>The evolution to conversational analytics</h2>
<p>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.</p>
<p>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.</p>
<p>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.</p>
<center>
<p><img class="aligncenter wp-image-744463" src="https://sisense.wpengine.com/wp-content/uploads/2023/11/1699279568767-1-300x172.png" alt="" width="600" height="344" /></p>
</center>
<p style="text-align: center;"><span style="font-size: 8pt;">Source: <a href="https://www.linkedin.com/posts/kevinpetrietech_bi-businessintelligence-generativeai-activity-7127295096743432192-I4Jb">Eckerson Group</a></span></p>
<p>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 Gen AI data analytics to your apps.</p>
<h2>A path to embedding generative AI analytics</h2>
<p>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:</p>
<h3>1. Conversational and contextual</h3>
<p>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.</p>
<h3>2. Composable</h3>
<p>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. </p>
<h3>3. Compliant, responsible, and trusted</h3>
<p>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 so that when they engage in conversation, there is no room for misunderstanding, hallucination, or error. It’s also critical to ensure that strong controls are in place; that data, prompts, and outputs are fully isolated from the LLM and provider, and not used to improve and train the model, or shared in any way.</p>
<h3>4. Cloud and LLM agnostic</h3>
<p>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 compliance, 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.</p>
<h3>5. Cost</h3>
<p>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.</p>
<h2>Sisense Compose SDK now with GenAI-powered analytics built-in</h2>
<p>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.</p>
<p>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.</p>
<h3>Introducing the Analytics Chatbot</h3>
<p>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. </p>
<p>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.</p>
<center>
<p><img class="aligncenter wp-image-744464" src="https://sisense.wpengine.com/wp-content/uploads/2023/11/AI-Chat-300x233.png" alt="" width="600" height="467" /></p>
</center>
<p>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. </p>
<p>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.</p>
<center>
<p><img class="aligncenter wp-image-744466" src="https://sisense.wpengine.com/wp-content/uploads/2023/11/GenAI-Hero-Img-1024x775.png" alt="" width="600" height="454" /></p>
</center>
<p>&nbsp;</p>
<h3>Enable your users to interpret answers easily with GenAI-powered Narratives</h3>
<p>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. </p>
<p>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.</p>
<center>
<p><img class="aligncenter wp-image-744465" src="https://sisense.wpengine.com/wp-content/uploads/2023/11/Insights-1024x542.png" alt="" width="600" height="317" /></p>
</center>
<h3>Lego-like composability for conversational control</h3>
<p>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.</p>
<h3>Designed for trusted, responsible, and compliant development</h3>
<p>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 that those answers are right and precisely answer the question. But it’s a crucial factor as you build out your roadmap around conversational analytics. </p>
<p>It’s 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 control to ensure that questions are turned into the right queries and from analytics answers to the 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.</p>
<p>Sisense uses Azure OpenAI Services, GPT-family models that deliver the data privacy and security of Azure. For further compliance and security, if your organization is running its own OpenAI instance, you can provide your OpenAI key to ensure that conversations are private to your instance.</p>
<h3>GenAI for analytics: create a roadmap that flows with change</h3>
<p>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 compliance needs emerge. </p>
<p>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. </p>
<p>We aim to enable you to deliver the best GenAI-powered analytics experience while ensuring complete back-end and architecture flexibility, from database to LLM. It means you can build for today but embrace what tomorrow will bring.</p>
<h2>Turn your generative AI analytics ideas into reality</h2>
<p>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>.</p>
<p>Joined by Paul Turner, Founder &amp; CEOSkyview Consulting, we’ll discuss the emerging conversational analytics landscape, highlight practical strategies to seamlessly incorporate these analytics into applications, and explore how to ensure ethical use and maintain trust in AI-driven analytics.</p>

How to drive analytics adoption with generative AI

How to Meet the Demand for Analytics Across the Organization

<h2>Introduction</h2>
<p>Launched by OpenAPI at the end of November, ChatGPT has changed the perception of what’s possible using Artificial Intelligence. A wide range of industries and use cases are adopting ChatGPT for its ability to reduce manual effort and accelerate operations. Not only can ChatGPT summarize information, synthesize new content, and assist development, it can also supercharge your analytics. In this white paper, you will learn how to use it to save time and effort in your operational analytics processes and gain more insight.</p>
<p>Built on top of GPT–3.5 as a Generative Pre-training Transformer language model, ChatGPT responds to a broad range of user inputs. As a result, ChatGPT can support a massive array of use cases, from automating coding, creating customer service responses, providing language translation, and more. It’s even poised to change how we regularly get answers to simple questions that we currently use Google for today. </p>
<p>ChatGPT is starting to change the way we work, but it is already clear that AI models can reduce costly operational overhead. For example, organizations routinely collect a wide range of ‘write-once, read-rarely’ profile information, such as company size, industry segmentation, or funding details. This type of information is generally available from multiple sources and only valuable when it’s needed. Plus, It usually requires significant curation and reconciliation.</p>
<p>Automating enrichment and augmentation of analytics with new fields, segmentations, and classifications is a challenge that analytics professionals and operations face daily. They must choose between spending costly hours grooming information that will likely never be used or waiting until it is required, when the lead time to gather it from external sources is prohibitive. </p>
<p>By operating as an on-demand source for augmented and ancillary information, ChatGPT has the potential to eliminate much of the pain and effort associated with traditional data preparation and, in the process, enable teams to segment and slice data in new ways, on-the-fly.</p>
<p>However, applying an on-demand model is only feasible if ChatGPT is tightly integrated with your data and analytics tooling. Without it, the ease of self-service analytics is quickly lost. Seamlessly integrating ChatGPT to supercharge your analytics is now possible. It’s easy to do, and we think you’ll find it’s a game-changer.</p>
<h2>Why you need to integrate ChatGPT with your analytics tooling</h2>
<p>Teams often spend way too much time in manual data prep and enrichment today. Estimates range as high as business and data analysts spending 70 to 80% of their time getting data ready for analytics, a large portion of that spent enriching and labeling data and creating new attributes to aid analysis. Adding new fields to the data sets for analysts to slice and dice by is often painful, onerous work. For example, consider a seemingly simple query:</p>
<blockquote>
<p style="text-align: center;">What’s the breakout of my customers by industry?</p>
</blockquote>
<p>It’s the kind of question critical to making decisions about entering new markets and launching new products. But what if the thousands of customers in the data set were not segmented by industry beforehand? What if Industry isn’t a field in the data or data model? As an analyst, it almost certainly usually means you’re out of luck without investing significant time and effort.</p>
<p>It typically 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. </p>
<p>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.</p>
<h2>Gains from a tight integration of ChatGPT with your analytics tooling</h2>
<p>ChatGPT 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 ChatGPT-sourced information into your analytics experience is critical to make it truly useful. For example, collect this information during ETL, and both storage and query costs will stay high; collect it once requested and you may struggle to insert it into the data model to enable the user's expected BI functions.</p>
<p>The optimal approach collects the requested data on-demand and uses ChatGPT to recommend the ideal insertion point in a model, providing a highly-responsive experience for the user. With this approach, users receive their own AI-powered always-on personal data prep and data analyst bot as part of their analytics experience. They can extend and augment their data and data model by simply asking ChatGPT. </p>
<p>Achieving it creates a massive opportunity for teams, freeing them from manual work prepping and augmenting data while ChatGPT does the lifting, 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 more fluid in their analytical queries, segmenting data in new ways, with ChatGPT doing the work behind the scenes. </p>
<h2>Sisense + GPT-3: A new way to supercharge your analytics</h2>
<p>A reference implementation of this strategy is available for Sisense Fusion, and is a perfect example of how to incorporate GPT models into analytics processing. This implementation leverages the uniquely flexible capabilities of Sisense Fusion to inject dynamic queries to GPT-3 as part of the query process, automatically generating new database tables and relationships in the existing data model while leveraging guidance from GPT-3. </p>
<p>This reference architecture also demonstrates low-code app experience development, creating an interactive Q&amp;A app via Sisense BloX so users can tell GPT-3 how they want to augment the data set and model, and do so directly in their dashboards. Sisense and GPT-3 immediately handle the rest. </p>
<p>How it all works can be demonstrated by answering the question posed earlier: “What’s the breakout of my customers by industry?”—when the customer attribute industry doesn’t appear in the original data model or data set. Sisense and GPT-3 will be used to fill the gap.</p>
<p>Below is the example data with a list of accounts. Ultimately, we’re looking for the percentage split of customers by industry even though the attribute doesn’t exist in the data.</p>
<p><img class="aligncenter wp-image-741530 size-large" src="https://sisense.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-4-1024x920.jpg" alt="" width="1024" height="920" /></p>
<p>To make it easy for the user to ask questions of GPT-3, Sisense BloX is used to build a simple interactive widget on a Sisense dashboard. This new UI assists in capturing what the user wants to request from GPT-3:</p>
<p><img class="aligncenter size-large wp-image-741528" src="https://sisense.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-2-1024x352.jpg" alt="" width="1024" height="352" /></p>
<p>In this example, GPT-3 is told the user wants the industry segmentation for a list of companies in natural language. GPT-3 is given the dashboard, widget, and column name, which ultimately contains the list of companies that need industry segment labels. </p>
<p>The UI executes a Python script containing the question and the dynamically populated list of customer accounts formatted as JSON and submits it to GPT-3, which retrieves the industry for each customer.  Alternatively, if we want to add GPT-3 enrichment to the scheduled periodic data model refresh, simply executing Python rather than using the Q&amp;A UI, we can do that too. </p>
<p>The result set is passed back from GPT-3 to Sisense Fusion as a JSON document. The Python script dynamically creates new tables in the origin database based on the GPT-3 result set and adds new relationships between tables in the model. Ultimately, it transforms hours or days of data enrichment effort into seconds.</p>
<p>In the example, GPT-3 returns each customer and their associated industry label in this example 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. </p>
<p><img class="aligncenter size-large wp-image-741529" src="https://sisense.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-3-1024x833.jpg" alt="" width="1024" height="833" /></p>
<p>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.</p>
<p><img class="aligncenter size-full wp-image-741527" src="https://sisense.wpengine.com/wp-content/uploads/2023/04/chatgpt-whitepaper-1.png" alt="" width="859" height="353" /></p>
<p>The whole integration enables the analysis to be completed in a few seconds rather than waiting days or weeks for an operations team to prep and publish the new information. Using the interactive BloX widget, augmenting and enriching any other data, all using natural language, is simple when powered by GPT-3.</p>
<p>As an exciting and related use case, a software product can easily add GPT-3 supercharged analytics by embedding this BloX widget directly into the user interface. As a result, users gain flexibility that extends the value of the product and its data by enabling augmented or ancillary data to be added without any involvement of the product vendor.</p>
<p>This approach to on-demand data isn’t necessarily easy to implement in a business intelligence product. 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 user experience. However, just about any ETL engine can at least augment data at ingestion using this technique. Sisense’s reference architecture for GPT-3 gives you a good starting point to try these techniques out for your own use cases. Here’s a <a href="/blog/how-to-make-bi-and-analytics-adoption-simple/">link</a> to an embedded example of the exact analysis we shared. </p>
<p>&nbsp;</p>

Resource Library

‌

‌

‌

‌

‌

‌

‌

‌

‌

‌

‌

‌