Overview:
In November 2023, we announced the release of Playbooks for Auto Insights, a groundbreaking new feature powered by Alteryx AiDIN. The initial release of Playbooks saved time and effort across the organization by automating the manual process of identifying, selecting, and building new analytics use cases for end users. This initial release reimagined the analytics development process by enabling quicker iterative development for business teams by removing technical dependencies in the development process.
Playbooks “Bring Your Own Data” Update
Now, with the latest release of Playbooks, customers can start from a dataset they have uploaded into Auto Insights, and Playbooks will recommend use cases tailored to that dataset within Auto Insights.
The steps are simple; the user starts by selecting a dataset available to them in Auto Insights from a drop down.
Once selected, Playbooks will randomly sample data, using the structure and topics of the dataset to recommend use cases for the user. From there, the user simply selects the use case they’d like to explore, and Auto Insights will build the report in seconds.
With this latest enhancement, Playbooks users have the ability to move from data to data story in less than 5 minutes, without building a single report or dashboard. This is a perfect starting point for non-technical users who want to explore their data but may need some help completing the initial build out in Auto insights.
Introduction:
Successful analytics initiatives, like any other technology initiative, require two key elements to achieve results: establishing business requirements and assessing technical feasibility.
At the onset of the project or initiative, the business requirements need to be established. For decision-making analytics projects, this often revolves around questions such as “What is driving X up or down?” or “How can we better understand Y?”.
Business requirements are typically informed by a business issue or problem that needs to be examined with the help of metrics & KPIs. At the end of the initiative, the business stakeholder evaluates the impact – was this project useful? Did we learn something important about our business? Can we use this moving forward to inform decision-making?
At the same time, technical requirements and technical feasibility need to be established and executed. Common questions that may arise during this time often pertain to availability of data, access to data, and shape/structure of data. For example, an analyst may ask themselves, “What does the data need to look like to arrive at this result?” or “What data elements are required to examine this metric?”
Herein lies one of the challenges of analytics development – until technical requirements are met, the viability or usefulness of an analytics solution is merely hypothetical. In other words, the business user cannot validate the value of a solution until it is built and ready to be used.
The Challenge
Business requirements and technical requirements are interconnected, and dependencies exist that can often slow down delivery times. Add the complication that to be successful, neither business nor technical requirements can be completed in isolation. Moreover, the steps in the analytics process must be executed in a predetermined order, which can be frustrating for those involved.
This frustration is rooted in misunderstandings or misaligned expectations between business and technical stakeholders, and frequently results in friction between the two camps. The speed of development is usually the friction point – development is too slow or inflexible because of dependencies.
Formulating the business challenge is the beginning of the process. But once the problem is identified, a slew of technical steps, such as data understanding, data preparation, and exploratory analysis must be performed just to build a prototype solution. Common steps here include things like:
- Locating where the data lives (e.g., source systems, cloud vs. on-prem, etc.)
- Gaining access to the data
- Preparing, blending, and transforming the data into a useful shape and structure
- Examining different features or variables that could impact the business objective
Because of these technical dependencies, it can take weeks or even months to arrive at an initial draft or “Version 1” analytics deliverable – a dashboard, data story, insight, etc. At this point, the business can re-engage to validate, interpret, and determine whether the solution is valuable. Often, this results in feedback from the business, which means all of those technical steps must be revisited and updated to deliver a new version.
What ensues is a rather rigid waterfall style of development, where the completion of new versions is dependent on hand-offs between the business stakeholder and the analytics professional building the solution.
The Challenge Illustrated: Waterfall Development & Dependencies
The Solution
One solution to these challenges would be to decouple the different sections of the process. This has historically been extremely difficult to do, until now. But with the power of generative AI, it is possible to decouple these steps and remove what once were strict dependencies in the development lifecycle.
If the business-led efforts can be iterated on without technical dependencies, then final requirements can be handed off after the business has proven the value. As a result, the business can rapidly iterate on new analytics projects, getting to valuable insights much faster than before.
Technical teams can reduce their workloads as well, as they can be removed from tasks during the prototyping phase. Technical stakeholders can be engaged for design work later in the development cycle when requirements are closer to being finalized.
The overall effect is that the entire analytics process has been re-engineered to remove technical dependencies during initial development stages. This results in operational efficiency benefits for both business and technical stakeholders, as both parties experience fewer dependencies and more autonomy during development.
Conclusion:
Analytics development is ripe with automation opportunities at every stage of the analytics process – from initial business problem formulation and ideation to data preparation & exploration to insight generation and analysis. Playbooks, powered by Alteryx AiDIN, unleashes the power of generative AI on the initial stages of the process to accelerate development cycles and remove friction & dependencies between business and technical stakeholders. We are excited to reimagine the analytics development process with this feature and, in doing so, make analytics more broadly available and accessible to everyone in the organization.