Date: January 2023 - May 2023

Role: Product Designer

Team: Andrew Chen, An Ho, Ariya Zheng, Angela Wei, Karina Avalos, Gayatri Kalra

I had a difficult, yet rewarding journey with this product. It was the first project that Tufts students had to pair up with a client. In this project, the client was LeanIX and by the end of the class and 4-month semester, we had to solve one of their problem statements and deliver a product to them that they could later implement within their platform for their respective clients.

LeanIX is an enterprise architecture company that provides IT infrastructure/software to help manage and organize different employees and teams (e.g. IT employees, business leaders). LeanIX does this by consolidating data collection of all business and IT needs into one place. At a high level, the company provides software as a service (SaaS), where clients of the company use their software to get data about their company and make more informed decisions based on this data. 

Dr. Matthew Grant is an experienced software content strategist and writer on B2B technology. This is defined as the exchange of services and information between businesses. Dr. Grant is well versed in software and IT management for mid-sized companies, and has written articles on everything from cloud management and security to UX design.

Matthew Grant is now working full time as the director of content and marketing at LeanIX and was our main point of contact for this project. Throughout the project, we work closely with Dr. Grant to understand the project requirement and expectation, define the target user, and determine measures of success.

LeanIX currently has difficulty navigating their marketing tactics in reaching a greater number of  clients with greater efficacy. They would like to explore methods to create more compelling and engaging content for existing and new customers. One of which, would potentially be using ChatGPT to better skew content on their blog and resource pages based on the background information of LeanIX’s clients.

Additionally, another possible solution would be to implement a tool/feature that could take in specific information from the user and generate a personalised outline as to why the user should take on LeanIX as a solution to their IT and software needs. 

Our main objective with this project is to to create a personalized website navigation experience for LeanIX’s clients that is curated to each user type that we have identified through user research and competitive analysis. The users are the following: Enterprise Architect, CTO, and an Application Owner.

Each user is guided through a series of specific questions as part of a survey related to their role, company details, EAM structure, and IT landscape. The survey then takes in these inputs and using ChatGPT this tool will generate a tech summary and recommended blog posts/resources. The user’s personalised tech summary will include a comparison of the client’s own IT needs and pain points with LeanIX features and capabilities that can be used to solve these challenges. 

Our second goal is to use and exploit any information available on these users such as their role, background, or where they are on the website so that AI can be used to refresh and re-frame content for them. ChatGPT will be used to make the blog and marketing content more engaging and compelling to incentivize new clients to take on an application like LeanIX.

We will be understanding how to better achieve this goal through desk research on ChatGPT and the user personas we will form from competitive analysis on companies with the Enterprise Architecture Management (EAM) industry. 

To better understand LeanIX products and gain more insight about their targeted user, we conducted two interviews with LeansIX’s Anna Konieczny (Head of Knowledge) and Lennart Hennings (Head of UX). From the conversation, we determine the three main user groups of LeanIX products: CTO/CIO, Enterprise Architect, and Application Owner.

CTO/CIO are the people who manage the IT resources at the company. They are primarily focused on business impact, and are in charge of how well an enterprise architecture management service assists them in building strategy for the IT department.

Enterprise Architect is both the primary user and the buyer of LeanXI products. They have a good technological background and when it comes to adopting a new enterprise architecture management service, they focus on the product's ease of implementation, usability and integration of it with their existing tech landscape.

They often come to the table knowing exactly what their problem is, not necessarily ideas about product features. They face difficulties in translating data and information into digestible non-tech concepts to convince the internal stakeholder.

Last but not least is Application Owner. Although they are not involved in the buying decision, they take up 80% of LeanIX product users. Their main task on the LeanIX platform is to complete a survey about their application.

Although this user group uses the application regularly, there is a lack of focus on them (both in terms of product and marketing) because they are not the buyer. As LeanIX value does not apply to them, they are often demotivated to adopt a new platform which might lead to low adoption rate.

Currently, IBM and Oracle are two of LeanIX’s biggest competitors within the EAM (enterprise asset management) industry. IBM's EAM platform, called IBM Maximo, offers comprehensive asset management and maintenance solutions using advanced analytics and AI capabilities.

IBM also provides a hands-on approach to market and sell their products, which they do by offering demos, free trials, and proof-of-concepts and thereby creating a stronger sense of personalisation and thoughtfulness in their product branding.

Similarly, Oracle also provides outstanding flexibility and customisation in marketing their EAM solutions. The company's sales team offers tailor-made proposals and unique pricing packages that cater to the specific needs of each potential client.

For instance, when working with a company to implement new EAM technology, Oracle's sales team would provide a personalised proposal that includes a proof-of-concept and a comprehensive cost-benefit analysis.

What we can see from these industry leading companies is that a general direction for designing a strong EAM product is flexible customisation and personalisation of EAM solutions through the use of AI-integration. Since all client companies have different needs and issues that they want solved through EAM products, the ability to provide specialized features to perfectly fit client goals will lead to obvious increases in company efficiency.

To better understand our end user, we have conducted two user interviews with Enterprise Architects – Harrison Coldwell (Senior Enterprise Architect at Adobe) and Jason Stehle (Solutions Architect at Amazon - previously an Enterprise Architect), who we recruited from LinkedIn. Our goals during the interview sessions are to understand their current workflow of adopting a new Enterprise Architecture Management Service, needs, goals, and pain points.

In the semi-structured interviews, we asked the interviewees open-ended questions about their steps for finding a third-party enterprise architecture management software, the criteria they look for, how they interact with stakeholders and clients during the process, what they value, and other follow-up questions. From the interview result, we determined what content they are most interested in learning about LeanIX to present them with personalised content accordingly. The interview guide and other notes can be found in Appendix A. 

Our final solution, after understanding the key takeaways from the interviews with an EA, was to incorporate a feature on LeanIX that serves as a chatbot. Instead of having an actual chatbot, we will cultivate this concept of interactive AI, where the user can customize what information they are looking for based on a survey they have the option to complete upon navigating the website.

These questions broadly cover the user's role, company size and description, the company’s current challenges, and what they value in managing their software and IT landscape. The main benefit of this feature is that it enables users to derive very specific information as to why they should take on LeanIX as one of their IT management tools, and allows them to further customize LeanIX to their company’s particular needs. 

We started our design process by laying out the positions of each frame and blocking out where wanted things to fit on the screen, such as having a progress bar at the top, a banner/header to get started, and multiple choice questions as options for the user to choose from and select their answer.

Each of us drafted our own designs and then pooled together what we liked best from each to consolidate our ideas and concepts. After several iterations, we decided on several components and variants for different buttons as well as finalized the color scheme and aesthetic of LeanIX branding. 

Along with a full tech summary, our solution also includes ChatGPT-generated suggestions for LeanIX recommended blog posts and white paper resources that may be useful for that particular user to read and learn about. The user can click on each of the resources personally customised to their needs and also have the option to view more and access LeanIX’s overall resource library directly. 

In addition to having an interactive AI feature that users can use to personalize their LeanIX navigation experience, we are also using ChatGPT to skew the overall content on the blog posts based on our three user personas: CTO, Enterprise Architect, and an Application Owner.

We crafted our prompt for this experimentation by explaining to ChatGPT the target user and then inputting the link to the original blog post that we want to tailor/change, and instructing the AI to generate a customised blog post to this specific user while still keeping the response in a blog post style, including headers and numbered points. 

After we completed the prototype, we conducted usability testing sessions with 2 EAs. During the remote testing sessions, we asked the participants to use the Concurrent Think-aloud method as they use our interactive prototype. Their task was to fill out the survey and get the customized report. The goal of the usability test is to evaluate the design heuristic and how effective and relevant the survey questions are. 

The LeanIX project presented a number of challenges for our team, including finding users to talk to and establish credibility with them. Due to the users' distance from our social and academic circles, it was difficult to recruit them for interviews and user tests. Consequently, we relied heavily on secondary research to collect information about LeanIX's users. To gain the trust of users who were willing to participate, we simplified and generalised the questions we asked, removing any that might be considered too sensitive.

Another challenge we faced was incorporating ChatGPT into the project effectively. We needed to ensure that its responses aligned with LeanIX's goals and messaging while remaining accurate and relevant. This required iterative testing and refinement to fine-tune the model and train it on the specific context of LeanIX's industry and offerings.

Client communication was also difficult at times. Since the client's vision for the project was not clearly defined at the outset, we had to work with them to create a problem statement, define the project scope, and determine relevant deliverables. Additionally, both our team and the client had very tight schedules, which made finding a mutually convenient time to communicate challenging. However, we eventually established our direction and updated our client and professor regularly about our progress.

Effective collaboration and communication require commitment from all team members involved in a project. Regular team meetings and client interactions foster collaboration, communication, and feedback exchange.

Although not all feedback may be positive, maintaining open communication channels and continuing to engage stakeholders throughout the design process can lead to more cohesive and successful design outcomes. A lack of commitment to communication and collaboration can lead to misunderstandings, miscommunications, and ultimately failed projects.

Desk research is an important tool in the design process, and it can reveal valuable insights about users and the industry. While many designers believe that user research requires direct engagement with users, desk research can provide a wealth of information. By leveraging industry reports and online literature, designers can enhance their understanding of industry challenges and standards.

Through desk research, designers can also uncover preliminary pain points and desired features for enterprise architecture management software. By using both user research and desk research, designers can build a more complete understanding of users and their needs.

Continuous iteration and improvement is still the golden rule for design. While it may be tempting to stick with the first or second idea, embracing an iterative design approach allows for continuous improvement and refinement of the solution.

Incorporating user feedback, testing, and iteration cycles throughout the design process can lead to surprising insights and critical improvements. Through continuous iteration and improvement, designers can build more polished and user-centric design solutions. Ultimately, this leads to more successful projects and satisfied users.

Creating a business case generator requires large-scale research and training an AI model with more data. The model should include information that is most crucial to the decision-making process of the target users. This will enable the generator to produce a customized product overview, cost analysis, integrations compatibility checker, and relevant case studies that help users make a compelling business case to clients or internal stakeholders.

To help users pitch LeanIX to their clients and internal stakeholders, a rationalization chatbot can be trained with large language models. The chatbot should allow users to customize their "dossier" and specific concerns of the personas, and help users build the most appealing argument to these personas with the strongest examples and statistics.

Personalized recommendations can be enhanced by incorporating machine learning algorithms that analyze user inputs and behaviours. This will enable the system to provide more accurate and targeted recommendations for specific LeanIX features, resources, and case studies that align with the user's needs and challenges.