Integrate Generative AI into Salesforce Marketing Cloud to increase SMB adoption
Timeline
August 2024 - Dec 2024
Problem
SMBs struggle with planning and executing high ROI marketing campaigns using
Salesforce Marketing Cloud due to its complexity.
Result
Made capturing inspiration and converting them to high performing campaigns
intuitve and accessible for SMB's and startups
Process
User research and market research to inform feature prioritization and roadmap
planning. Evaluative testings and a lot of iteration.
Tools
Figma, Axure, Miro
Defining KVP's
Lower Barrier to Entry →
Simplified marketing workflows for SMBs, making Salesforce accessible to non-marketers
AI Differentiation →
Strengthens Salesforce’s AI capabilities (Einstein AI), positioning it ahead of competitors
Improved Customer Retention →
AI-powered automation keeps SMBs engaged, reducing churn
1
Business goals
Long term enagagement
Expanding market reach to SMB's
Stay ahead of the competitors
2
User goals
Plan, create and execute high ROI campaigns without expertise on the go using SFMC
How we did this
OVERVIEW
What does Salesforce Marketing Cloud do?
Currently optimized for big enterprises, SFMC helps them strengthen relationships with their customers through multi-channel (Email, SMS, WhatsApp) communication.
Goal? Get the right message, to the right customer, at the right time.
10
iterations
and concept testings
78
%
boost in idea to execution rate for SMB's
What does the current ecosystem look like?
SFMC is a very complex ecosystem with multiple tools interacting to do ONE thing - execute marketing campaign. To target our efforts, we mapped out how SFMC currently executes campaigns, how the data flows and what dependencies occur within these workflows
What I learned about how current users use SFMC
This is how different enterprise teams collaborate to execute large scale marketing campaigns
What I learned about why SMB's struggle with this ecosystem?
Based on our research we found out that SMB's struggle to use SFMC because - It takes time and expertise to learn the tool, people executing marketing campaigns are not specialized in marketing and they lack the required data on customer segmentation, marketing trends and key analytics insights.
We reimagined the flow to look something like this
We wanted a single SMB marketer to seamlessly manage campaigns across all studios. Instead of navigating fragmented, manual workflows, AI now would help with reducing cognitive load and eliminating technical barriers. It acts as a co-pilot, intelligently routing tasks to empower SMB's to execute large scale campaigns.
Current user workflow
From our research, we identified that our users follow a linear, multi-step process to create and manage marketing campaigns. This process is often manual, resource-intensive, and requires specialized expertise at each stage
Planning
Set goals, define the audience, and create a strategy
Developing
Create content and set up channels for the campaign
Post-Launch Analysis
Evaluate results and gather insights for improvement
Launch
Execute the campaign and monitor its progress in real time
Key insights about the mental model of our target users (SMB's)?
Marketing planning doesn’t happen at a desk, it happens on the go, when they see an ad, hear customer feedback, or get a new idea. Their mental model is reactive rather than strategic Unlike enterprise teams that work within structured office environments, SMB owners and small teams are constantly multitasking, handling operations, sales, and customer engagement while planning their marketing efforts.
No strategic direction
Many SMBs don’t know where to start when crafting a marketing campaign.
No data to plan
Without dedicated marketing analysts, SMBs lack data-backed insights to guide campaign planning.
Ad-hoc inspiration
SMBs rely on ad-hoc inspiration (competitor ads, customer feedback), but these ideas often get lost.
Defining mental models
Guesswork & gut instinct →
data-driven & confident decisions
SMBs often lack the data expertise to make informed marketing choices. We wanted to shift their mental model from trial-and-error marketing to strategic decision making.
Where do I start? →
confidently create campaigns
AI proactively suggests relevant campaign ideas, audience segments, and content strategies based on past behaviors and industry
Syntheized Research Insights
We found out
Finding inspiration is the starting point for all our target users. Without past data they don't know what will work
But
Inspirations can be drawn from anywhere and there's too many ideas to go after
And
The current ecosystem doesnt allow converting inspirations -> successful campaigns.
So let's
Plan, create and execute high ROI campaigns without expertise on the go
How might we incorporate generative AI to help SMB's capture inspiration and convert it to a successful campaign on the go?
BRAINSTORMING
Can and SHOULD Gen-AI be implemented to automate this flow?
Yes and No. Here's how AI should be implemented.
AI should support and augment human decision making capabilities, not replace human agency
The capabilities and limitations of AI should be clearly communicated
User should comprehend and trust the output created by AI
Multilayer Personalisation and
Memory Scaffolding
As the users engaged with SF, we wanted the AI to understand their tasks and preferences to generate more specific interfaces.
AI can act as memory scaffolding, helping SMBs collect, refine, and structure their ideas into campaigns.
This goal was also created to differentiate it from it's competitors
Salesforce's Einstein AI provides predictive analytics, NLP, and content personalization, allowing AI-powered inspiration capture, campaign recommendations, and automated content generation to work within the system.
AI-generated marketing materials can be stored, modified, and deployed through Content Builder.
Salesforce’s API-driven architecture allows AI-enhanced capabilities to integrate into existing marketing workflows, third-party data sources, and automation tools
Can the current ecosystem support this?
EVALUATIVE RESEARCH
How we carried out our concept testings
To shape and evaluate our concepts, we conducted testing with Salesforce designers and early startup owners.
Feedback from design team
Regular sessions critiqued for relevance and usability
Iterative Process
For fine tuning features, making the app more user focused
Result
Clearer workflows, smarter AI suggestions that meet startup needs
Key insights that drove our design
decisions
We discarded and iterated on a lot of concepts closely collaborating with our internal stakeholders to understand the technical feasibility and relevance of our implementations. We made sure to keep our fidelity low in the early concept testing stages to make sure we were evaluating the concept and feasibility and not usability.
How did we prioritize features through testings?
Our feature prioritization framework was driven by three key principles: reducing friction in marketing ideation, enabling data-driven personalization, and ensuring seamless AI-human collaboration. Through iterative user testing and stakeholder alignment, we identified features that balanced AI automation with user control.
Natural Language
Processing
This feature lets the user type their task in their own words and the AI recognizes the actionable parameters
Empowering, not controlling
This feature lets the user draw inspiration from a campaign and tailors it for their next campaign
Task flow generation
Once the ideas are finalised, Salesforce synthesises it and suggests a real-time task flow from start to execution
Let's see them in action
Learning and growth
Finding the best solution for my project involved learning a lot of new concepts and skills, getting guidance where needed, taking on tasks and going out of my comfort zone. The first phase of the project focused on understanding the main problem I was tackling. This meant going out of your way to learn more about the product itself and have a sense of their pain points and goals.
A large portion of my time was spent wireframing and iterating on the product feature. I had a lot of ideas which I committed to pixels on the screen, and created as much as time allowed. I learned how to be a better storyteller when presenting my work in team design critique sessions, in order to get as much valuable feedback as I could. After creating numerous iterations and an interactive prototype for the features, I validated and tested my assumptions. I wrote a script and conducted several concept testing sessions which provided me valuable insight that I then applied to the next versions of my work.
Throughout the project, we followed an agile framework, cycling through research, design, testing, and iteration. Our biggest revelation was that agile is not just a process; it's a mindset. Rather than treating research, design, and testing as sequential steps, we learned to blend them fluidly, letting our findings continuously shape our decisions. From managing stakeholders and synthesizing research to presenting to leadership and receiving peer feedback, this project was pivotal in my growth as a designer.
Copyright 2024 by Priyanvada Darshankar