Kick Off Your AI Project

Learning Objectives

After completing this unit, you’ll be able to:

Identify stakeholders, goals, and a technical solution for your AI project.
Explain the stages of an AI project.
Schedule the project timeline.

Before You Start

Tackling an AI project is a huge milestone for your company, but before you get started, make sure you complete the AI Strategy module. In that module, you identify your organization’s use cases for AI, build an AI roadmap, and choose the right project.

Prepare for Your AI Project

Navigating the complexities of AI projects can seem intimidating, especially if you’re new to the technology. If you’re responsible for implementing an AI project in your organization, what do you need to think about when you’re planning the project? What can you do to set your project up for success?

In this module, you follow along with Becca Cloudier, the Salesforce admin at Coral Cloud Resorts, as she plans her first project that combines data with AI. Coral Cloud is a vacation resort dedicated to providing each of its guests with a relaxing and fun stay. The organization wants to streamline the check-in process while still keeping it personalized. That means resort staff can give more attention to their customers, who in turn spend less time standing at a check-in desk.

So let’s dive in and see how Becca kicks off an AI project. She starts by identifying stakeholders, defining project goals, and choosing a technical solution.

Identify Project Stakeholders

First, identify your project stakeholders. These are people who have an interest, influence, or impact on the project. Knowing who these people are and looping them in early on helps ensure that your project is aligned with everyone’s goals and doesn’t neglect a specific area or audience.

Keep in mind that your project stakeholders are typically different from strategic, company-wide stakeholders for AI, although there might be a little overlap. Project stakeholders are dedicated to the implementation of that specific AI project, while strategic stakeholders manage the company-wide AI strategy. Even though strategic stakeholders provide oversight, the individual project team leads are accountable for the project.

Becca needs to get some people on board before she can start implementing her project.

People manager of the end users: This person manages the end users affected by the AI project. For Becca, this is Josef Rose, the customer success manager at Coral Cloud.
Executive sponsor: This person allocates resources and prioritizes the AI project. For Becca, this is the VP of Customer Experience at Coral Cloud.
Security and Legal: These stakeholders make sure Becca’s AI project is ethical, secure, and uses customer data legally.
Technical team: They build the AI project. In this example, Becca’s planning to build the project herself.

Throughout the project’s development, communicate regularly with your stakeholders. Understand their needs, collect feedback, and make changes to your project.

Visualize Success

Next, visualize success for your project by defining goals. Keep the use case you identified in the AI Strategy module in mind. Aim to define a SMART goal: specific, measurable, achievable, relevant, and time-bound.

Becca works with leadership to define these goals.

Reduce check-in time by 50%.
Maintain customer satisfaction at the same level as or higher than before the project.

Don’t skip this step. Without the proper metrics and key performance indicators (KPIs), you can’t gauge the success of your project. Becca knows that reducing the time spent at check-in increases customer satisfaction and leads to increased customer spend at the resort. By directly tying AI to increased revenues, you can show a high return on investment, which helps make your case for future AI projects.

Consider the Technical Requirements

When you’re getting started on an AI initiative, it’s essential to assess the technical requirements of the project. Here are some of the questions to ask.

What type of AI does the project require? Predictive, generative, or both?
Does this solution need to integrate with other systems?
Are there any out-of-the-box solutions you can use and customize, or do you need to build it yourself?
If you choose to build it yourself, does your organization have the right in-house skills?
What models, programming languages, frameworks, libraries, and tools will you use?
How will you balance trade-offs between accuracy and speed, complexity and simplicity, and innovation and cost?

If you’re building your project in Salesforce, you need to have Data Cloud and Einstein Generative AI enabled in your org.

In addition to the project’s technical requirements, you also need to consider the data requirements. You learn about those in the next unit.

Choose an AI Solution

Coral Cloud Resorts is already using Salesforce for guest check-ins, so Becca looks at the current process to identify how AI can optimize the experience.

Whenever a guest arrives at the hotel, a Coral Cloud staff member manually creates a check-in record. On top of that, a staff member manually sends a welcome email that includes personalized recommendations for fun excursions. This improves the guest experience and increases cross-sell conversion rates. This high-touch approach is time-consuming, so Becca decides to use AI to speed up the process with automatic AI-powered check-ins that trigger an AI-generated welcome email. To shorten her implementation timeline, keep costs low, and make sure the AI solution integrates easily with Salesforce, she chooses Einstein Generative AI as the right tool for the job.

Next, she takes some time to peruse the Einstein Generative AI Help portal to see if any out-of-the-box Salesforce AI features will work for her use case. She starts visualizing how she can use the features together with Data Cloud. She realizes that she can use a flow to trigger an action in Einstein Copilot, and she finds out that Prompt Builder can generate personalized emails based on a template.

Becca decides on a three-part project involving different Data Cloud and AI features. Here’s her plan.

Use a flow to create a Guest Event record based on the latest reservation data in Data Cloud.
Teach Einstein Copilot how to launch the flow through conversational language. So for example, when guest Sofia Rodriguez arrives to begin her stay, the staff can simply ask Einstein to “Check in Sofia Rodriguez” and Einstein does the rest!
Use Prompt Builder to generate a personalized welcome email that suggests excursions based on previous excursions the guest purchased.

Stages of an AI Project

Let’s look ahead at what’s in store for Becca’s AI project.

Plan: This is the phase you learn about in this module.

Define the problem to solve with AI and how to measure success.
Assess the project’s technical and data requirements.
Identify the features and customization to solve the problem.
Prepare your data.
Build a trust strategy.
Share your plan with your project stakeholders.

Build: Build your solution, test it, and refine it.

Set up, customize, or build the solution.
Conduct a pilot and collect feedback.
Refine your solution.

Launch: Deliver your project to your end users.

Announce the change to the organization.
Deliver training.
Take a baseline measurement of your metrics.
Roll out to all end users.
Collect feedback.
Evaluate the project’s success.

After your project is launched, don’t dust your hands off completely. Your project requires active maintenance to stay effective. Continue to get qualitative and quantitative feedback on your project and update your solution according to the feedback.

Determine the Project Timeline

With some important details ironed out, Becca can finally plan the timeline for her project. Keep in mind that your project might take more or less time depending on the complexity of the solution. This sample timeline is hypothetical, and it doesn’t include the amount of time it takes to reach data readiness, which can vary greatly based on the quality, availability, and accessibility of your data.

Stage	3-month plan (12 weeks)	6-month plan (24 weeks)
Plan	2 weeks	4 weeks
Build	9 weeks (1 week testing, 2 weeks pilot)	18 weeks (2 weeks testing, 4 weeks pilot)
Launch	1 week	2 weeks

Summary

Now you know how to kick off your AI project by identifying project stakeholders, goals, and a technical solution. You also understand the stages and timeline for your project. In the next unit, learn how to prepare your data and share criteria for assessing data quality.

Besoin d'aide ?

En savoir plus