the analysis game
Posts
Utilising Generative AI at Project Discovery.

Utilising Generative AI at Project Discovery.

Prompt Engineering: The Essential New Skill for Business Analysts.

Gabriel Botsie
September 20, 2024 • Estimated Reading Time: 9 minutes

Modern office, image depicting a project discovery between client and project teams

Project Discovery Meeting by the Author

In the book Business Analysis Techniques, 72 Essential Tools for Success, the authors divide Requirements Management¹ into four categories:

Requirements Elicitation: Techniques used to investigate, identify requirements from stakeholders
Requirements Analysis: The organisation, grouping and management of requirements
Requirements Development: Techniques to ensure that requirements are created to specific standards
Requirements Modelling: Conceptual modelling techniques such as use case diagrams, Entity Relationship Diagrams that support textual requirements

All four categories are evident in the project discovery phase.

Illustration showing the aims of project discovery - vision, goals and scope

Project Discovery by the Author.

Project Discovery typically occurs over a 2–3 week period, with multiple activities between project and client teams; the purpose - to align on vision, goals and scope.

Business Analyst Responsibilities

For Business Analysts (BA), non-linear activities during this period include:

Requirements Elicitation	Requirements Analysis	Requirements Development	Requirements Modelling
Market & Project Analysis	Draft Requirements	Review Requirements with Project Team	Prototype Development
Stakeholder Meeting Prep	Organise Requirements	Acceptance Criteria	Process Mapping
Identify Key Requirements / Processes	Requirements Prioritisation	Data Analysis	Diagramming
Document Objectives / Goals	Initial Documentation Review	Requirements Sign off	Conceptual Modelling

At the end of the 2- 3 week period, client and project teams must be aligned on vision, goals and scope.

Generative AI Capabilities

Generative AI² capabilities include the generation of images, videos, code all using natural language via prompt engineering³ .

Generative AI: Transforming Project Discovery for Business Analysts, by the Author

It supports all the above tasks undertaken by BA’s within project discovery.

Tools

The list of Generative AI tools grows by the hour!

I’ve chosen three tools for BA use across all requirements management categories.

I recently came across BA Co-Pilot - a BA specific tool.

Developed by EXL, BA Co-Pilot "integrates Generative AI into the Software Development Life Cycle enhancing efficiency and reducing defects during User Acceptance Testing (UAT)."

Later in the video, they outline EXL’s performance metrics:

50% effort reduction in creating user stories, test cases and business requirement documents
Reduce UAT issues by 50%
70% reduction in re-work.

I haven't tried EXL so I can't validate the above.

However, impressive time savings!

Meet ClaudeClaude is AI for all of us. Whether you're brainstorming alone or building with a team of thousands, Claude is here to help.www.anthropic.com/claude

UI Design Made Easy, Powered By AI | UizardUI design for apps, websites, and desktop software in minutes. If you can use Google Slides or PowerPoint, you can use Uizard!uizard.io

As a general tool, I prefer Anthropic's Claude.

The regular models as well as Projects and Artifacts assist with all the capabilities shown in the Generative AI diagram.

For prototyping, there's Uizard allowing the generation of wireframes and prototypes using natural language prompts.

All these tools cut the time to produce documentation (as much as 50% according to EXL), leaving more time to focus on project success.

Prompts

Below are two examples of prompts I’ve used to help with meeting preparation and document summarisation.

Content Generation	Summarisation
Act as a Senior Business Analyst I want you to provide and describe 5 key considerations for starting a new client project. The client wants to develop a new mobile app for their fashion conscious millennial consumers. Each consideration should have it's own headings. Describe each consideration and how it helps project success. Format each heading in Bold. Points should flow logically. Write in a 1 on 1 conversational style. Use sub headings where relevant, Write in a professional yet informal manner. Client goals and needs are paramount. As the Analyst you must ensure that their needs and met as well as ensuring the project team - that you're part of - delivers what the client requires. Outline key questions and actions that must be undertaken / considered by the client and project teams to ensure project success. Write from the perspective of an experienced Analyst, with software development and stakeholder management experience. The goal is to ensure the Analyst is fully prepared for a new client project. Source: Using Large Language Models for New Project Preparation and Planning.	You will generate increasingly concise, entity-dense summaries of the attached text. Repeat the following 2 steps 5 times. Step 1. Identify 1-3 informative Entities (" ; " delimited) from the Article which are missing from the previously generated summary. Step 2. Write a new, denser summary of identical length which covers every entity and detail from the previous summary plus the Missing Entities. A Missing Entity is: - Relevant:to the main story. - Specific:descriptive yet concise (5 words or fewer). - Novel:not in the previous summary. - Faithful:present in the Article. - Anywhere:located anywhere in the Article. Guidelines: - The first summary should be long (4- 5 sentences, ~80 words) yet highly non-specific, containing little information beyond the entities marked as missing. Use overly verbose language and fillers (e.g., "this article discusses") to reach ~80 words. - Make every word count: re- write the previous summary to improve flow and make space for additional entities. - Make space with fusion, compression, and removal of uninformative phrases like " the article discusses" . - The summaries should become highly dense and concise yet self-contained, e. g., easily understood without the Article. - Missing entities can appear anywhere in the new summary. - Never drop entities from the previous summary. If space cannot be made, add fewer new entities. Remember, use the exact same number of words for each summary. Source: Chain of Density Prompting: A New Way to Generate Better Summaries with Generative AI.

Content Generation

Summarisation

Act as a Senior Business Analyst I want you to provide and describe 5 key
considerations for starting a new client project.
The client wants to develop a new mobile app for their fashion
conscious millennial consumers.
Each consideration should have it's own headings.
Describe each consideration and how it helps project success.
Format each heading in Bold. Points should flow logically.
Write in a 1 on 1 conversational style.
Use sub headings where relevant, Write in a professional yet informal manner.
Client goals and needs are paramount. As the Analyst you must ensure that their
needs and met as well as ensuring the project team - that you're part of -
delivers what the client requires.
Outline key questions and actions that must be undertaken / considered
by the client and project teams to ensure project success.
Write from the perspective of an experienced Analyst, with software development
and stakeholder management experience.
The goal is to ensure the Analyst is fully prepared for a new client project.

Source: Using Large Language Models for New Project Preparation and Planning.

You will generate increasingly concise, entity-dense summaries of the attached text.
Repeat the following 2 steps 5 times.

Step 1. Identify 1-3 informative Entities (" ; " delimited) from the Article which are missing from the previously generated summary.
Step 2. Write a new, denser summary of identical length which covers every entity and detail from the previous summary plus the Missing Entities.

A Missing Entity is:
- Relevant:to the main story.
- Specific:descriptive yet concise (5 words or fewer).
- Novel:not in the previous summary.
- Faithful:present in the Article.
- Anywhere:located anywhere in the Article.

Guidelines:
- The first summary should be long (4- 5 sentences, ~80 words) yet highly non-specific, containing little information beyond the
entities marked as missing. Use overly verbose language and fillers
(e.g., "this article discusses") to reach ~80 words.
- Make every word count: re- write the previous summary to improve flow and make space for additional entities.
- Make space with fusion, compression, and removal of uninformative phrases like " the article discusses" .
- The summaries should become highly dense and concise yet self-contained, e. g., easily understood without the Article.
- Missing entities can appear anywhere in the new summary.
- Never drop entities from the previous summary. If space cannot be made, add fewer new entities. Remember, use the exact same number of words for each summary.

Source: Chain of Density Prompting: A New Way to Generate Better Summaries with Generative AI.

Benefits and Risks

I've saved significant amounts of time utilising both prompts.

With the summarisation prompt, Claude:

reads and summarises documents
I check for hallucinations and inaccuracies

Once reviewed and verified, on to the next task.

Organisations are rightly concerned about data privacy and security - will LLM's learn / train off key organisational data? Valid concerns. However, there are now Enterprise versions of Claude and Open AI's ChatGPT designed with these concerns in mind.

In Conclusion

Generative AI augments BA’s by combining domain skills with Generative AI capabilities. We can automate elicitation, analysis, development and modelling activities, ensuring their outputs are at the required quality.

For example, set up prompt repositories to undertake:

Market / Project Analysis
Automatically write backlogs to agreed standards
Prototype design prompts for eCommerce desktop and mobile sites

At the start of a new project, visit the repository, use and modify the prompts as needed.

Time Savings

I’ve spoken about time savings, but how much time is actually saved?

Figures vary:

Databox⁴ put the figure between 25-74% for marketers
A Forbes article⁵ states that Generative AI could save 40% of our time.
Pearson⁶ looking at hours saved by UK workers on routine and repetitive tasks, highlights a saving of 19 million hours per week!

It’s confusing, unclear. However, 40% time savings ‘feels’ about right.

In a 3 week project discovery that’s 6 days back!

Time enough to solidify the vision, goals and scope as well as mitigating against the common causes of project failure.

Call to Action

Clearly, Generative AI isn’t just for project discovery. It’s universally used by BA’s and others for a range of tasks. There’s the following calls to action for BA’s:

Automate: Review, evaluate project and client processes. Automate where it makes sense
Educate: Dive into courses to learn more about Generative AI applications
Be Informed: Keep up to date with the latest developments, courses and tools
Review: Prompt repositories, what worked, what could be improved

Some resources to get you started.