Transforming Work With an AI Analyst For Every Team

"We've been able to give everyone in our company  their own personal analyst. I asked questions yesterday and Iris gave me a response in under a minute in a very detailed, very easy to understand manner. With Iris, Twiddy has eliminated the analytics bottleneck while transforming every employee into their own business analyst—from guest services answering customer calls to executives making strategic decisions."  

Blake Stockslager, CIO at Twiddy

Challenge

Data Expertise Was a Bottleneck

Twiddy was experiencing the classic scaling challenge that breaks most data-driven companies. As their business grew and data culture matured, analysis requests were taking days to fulfill.  Cole Steigman, Twiddy's data lead and one of  the company's most valuable—and overwhelmed—employees,  had become the bottleneck that influenced how data driven the organization could actually be relative to its aspirations. There were several noticeable challenges that Cole and the the CTO of Twiddy & Company, Blake Stockslager recognized

  • Response time constraints: "My turnaround time for questions that I can't answer off the top of my head is like a few days," Cole explains. "So if somebody wants an answer while they're on the phone with a customer or home owner, I'm probably not able to get them that."
  • Meeting inefficiency: A good portion of the weekly strategic meetings were consumed in gathering the metrics about the business when it would have been more valuable to focus on what needed to get done. All of these meetings required Cole to be present as the data sherpa in the team
  • Cultural constraints: Despite investment in Google BigQuery and BI tools like Looker, business users "were still not able to get to insights needed fast enough" Blake notes. These tools needed Cole to connect the dots between a business situation and the data for many of the teams.

Solution

Everyone Gets an AI Analyst That Speaks "Twiddy"

Twiddy  & Company launched Iris systematically from department to department to address this bottleneck. There were several attribute of iris that made it a strong fit for Cole and Blake to embrace it as a solution

  • Eliminated data-literacy as a pre-requisite to gain insights: While traditional BI tools force business users to think in data terms, Iris understands the language of Twiddy like any employee. In their industry, several important functions require employees that excel at relationship building with empathy and outstanding communication, as opposed to technical skills to analyze data. Iris is purpose built for such teams, providing business insights as opposed to data tables and charts as the primary output. Marketing teams could ask "What should I write a blog about?" and homeowner relationship managers could ask "Should I advise the homeowner of E069 to make their home pet friendly?" and get recommendations based on comprehensive analyses that Iris does on their behalf.
  • No custom modeling overhead for the data team: The system doesn't require the heavy lift of custom model development or building out a semantic layer. It simply works with whatever is in the warehouse. "The product does not require any custom modeling. It is able to infer business interpretation from the data in the tables," Cole explains. Iris learned the nuances of Twiddy's vacation rental business through natural conversations and training that was like getting a really smart, experienced analyst familiar with Twiddy.
  • Strategic business problem solving, not just data retrieval: Iris provides business-level insights with the same analytical depth as Cole himself. It engages in an intelligent conversation to understand what the employee is trying to answer, brings in the full contextual understanding of the business and comes back with a calibrated, actionable insight. "It feels more like you're talking to an actual analyst about the information... It's going to provide them actual insights and information based on what the data entails," Cole notes.

Blake describes the transformation: "We've been able to give everyone in the company their own personal analyst."

Results

Data Activated Actions Up Across Twiddy!

  • It's working - The number of business outcomes activated with data has increased noticeably - Twiddy is now observing a 500% increase in the number of questions that business teams previously used to ask their data analysts. With the growth in volume, they are noticing questions from users who have never engaged with data before and contexts that are concretely work outcome driven. From marketing content optimization ("What should I write a blog about?") to guest services support ("Which week is most available?") to owner services strategy ("Will this home make more rent if it were pet-friendly?"), Iris is transforming how work gets done across the company. 
  • It's preserving resources. Cole and the data team are freed from constant ad-hoc requests, allowing focus on strategic initiatives rather than repetitive question-answering. Cole spends less time at weekly team meetings as a data interpreter.