The hardest problems to solve are the ones we have not truly understood yet.
We often rush to the “solution space” – eager to automate, predict, or personalize – without first slowing down to decode what the real problem actually is. In the world of Salesforce and AI, this step makes all the difference between a shiny demo and a system that genuinely transforms how people work.
When “We Need AI” Is Not the Problem
You have likely heard statements such as: “We need AI to improve our sales,” “Let’s use Copilot to summarize customer calls,” or “Can we make Service Cloud predictive?” These ideas often contain potential, yet they rarely define a real problem.
As business analysts, the instinct is to step back and ask the essential question: “Why?” A request for AI is not a problem statement – it is a desire for improvement that lacks structure. Our responsibility is to convert that desire into clarity, measurable needs, and actionable direction.
True impact comes from identifying friction – the operational bottlenecks, data blind spots, and experience gaps where intelligence can produce meaningful change – rather than pursuing technology for its own sake
Sales reps can easily spend 30% of their day updating Salesforce instead of selling. Service agents lose precious minutes digging through case histories before they can respond. Managers second-guess the forecast because the data underneath doesn’t feel reliable.
These aren’t technology gaps – they’re productivity leaks. And they signal a deeper truth: when systems aren’t designed intentionally, people pay the price in time, clarity, and confidence. Understanding these patterns is the first step – redesigning them with purpose is where transformation begins.
The Art (and Discipline) of a Good Problem Statement
A good problem statement, in my experience, isn’t about fancy phrasing. It’s about focus. It should tell a story of pain, people, and purpose. Here’s the structure I keep coming back to:
- What’s broken or inefficient?
Sales reps spend too much time on manual data entry. - Who is impacted, and how?
This reduces active selling time and lowers conversion rates. - What’s the desired outcome?
Free up time for client engagement and improve deal velocity.
Once Salesforce and AI enter the conversation, the work shifts into guided design. Agentforce can automate data entry that drains selling time and can produce clear meeting summaries that reduce administrative overhead. Einstein Discovery can highlight next-best actions that improve decisions.
The distinction is significant: you are addressing a clearly defined problem, not wandering through a technology playground hoping to find one.
AI Does Not Fix Ambiguity, It Helps Expose It
Here is the tricky truth: AI thrives on clarity, not assumptions. If the problem statement is fuzzy, the AI output will mirror that fuzziness – beautifully, confidently, and completely wrong. Anyone who’s seen a well-written nonsense summary from a chatbot knows exactly what I mean.
Take this example:
- “We need better lead quality.”
- “We need to identify leads with high conversion potential based on engagement and purchase signals.”
See the difference? The first invites guesswork; the second directs both your data and design efforts. AI does not replace analysis – it magnifies the quality of it.
The Salesforce-AI Sweet Spot
Salesforce has given us tools that make AI accessible – from Agentforce to Data 360 (formerly Data Cloud) to Prompt Builder.
However, accessibility does not mean simplicity. A BA in this space needs to operate in two modes simultaneously:
- Strategic storyteller: understanding business pain, mapping journeys, and connecting outcomes.
- System translator: knowing how data, models, and automation can bring those outcomes to life.
Let’s say you hear: “Customer service feels reactive and inconsistent.”
Instead of jumping to dashboards or chatbots, reframe it: “We need to proactively predict service issues and empower agents with contextual insights.”
Now, you have got a direction:
- Data 360 unifies customer data.
- Einstein Discovery predicts churn or issue patterns.
- Copilot helps agents with context summaries.
That is not just automation – it’s intelligent enablement.
Ask Better, Solve Smarter
AI can generate outputs in seconds, but a BA’s questions shape the value of those outputs.
So when you are asked, “Can AI fix this?”, try asking back:
- What behavior do we want to change or improve?
- What decision do we want to make faster or smarter?
- What will success look like if AI worked perfectly here?
Every question sharpens the lens on the real issue. Because when you decode the problem right, the right solution almost reveals itself.
Seeing Through the Lens of Value
A well-crafted problem statement is more than documentation – it’s an alignment tool. It ensures that the technology we build, the dashboards we configure, and the AI we deploy actually create business value and human impact.
In Salesforce and AI projects, that’s our north star. We’re not just gathering requirements – we’re uncovering meaning.
Final Thoughts
The tech is only as smart as the problems we feed it. As business analysts, we are the translators – from business pain to machine logic, from ambition to architecture.
So next time someone says, “We need AI,” smile and ask: “What problem are we solving again?” because clarity is still the most intelligent tool we have. Once that is in place, even Einstein would approve.
