From Frustration to Flow - How to Use AI to Reduce Customer Friction
Many AI-driven projects start with the wrong questions.
Questions like:
What should we do with AI voice (or agents or reasoning models)?
How could we use Microsoft Azure's newest feature?
What are our competitors doing?
Questions like these start from the point-of-view of technology and what it enables.
In this article, I propose a more strategic and revenue-driving starting point.
What are the most significant points of friction for your customers?
Picture a line. Let's call it a customer friction journey.
On one end, you have a customer with a need - they want to go somewhere, eat healthier, or accomplish their Q3 sales goals. On the opposite end, they have successfully fulfilled their need with your solution.
What are the steps between these endpoints?
Often, they are points of friction — steps we need to take with companies to fulfill our needs. We're typically not that eager to take these steps.
Friction for customers can take many forms:
Research - Gathering and analyzing information to make an informed decision or solve a problem
Communication - Engaging in back-and-forth exchanges with a company to express needs, get answers, or complete necessary processes
Choices - Navigating complex decision-making with multiple options, variables, and constraints to find the optimal solution
If you can use empathy, data, and intuition to identify your most pressing points of friction, you'll have a useful starting point for exploring how AI can help your customers.
Next, let's unpack each of these categories of friction (Research, Communication, Choices) along with cases from other companies using AI to reduce their customers' friction.
1. Enrich Research
Buying a used car can be stressful.
Electric, gas, or hybrid? What is the estimated range? Is the trunk large enough for our strollers?
CarMax uses AI to give customers glanceable overviews of vehicle reviews. Their system synthesizes thousands of customer reviews into concise summaries highlighting key takeaways like comfort and reliability. This helps buyers quickly understand other owners' experiences without reading dozens of individual reviews – transforming what would have taken years of manual work into a streamlined car shopping experience.
I worked with AI safety startup Saidot that makes it easier for decision-makers to understand the different AI models and AI governance and risks involved.
Here's how other forward-thinking companies are already tackling research friction with AI:
Notion's Q&A transforms how employees find information in their company knowledge base. Using Claude, Notion created a universal search tool that answers questions about any content in a user's workspace. As Simon Last, co-founder at Notion explains, "We've integrated our retrieval system throughout the platform... This allows users to instantly access relevant information from their entire workspace, whether drafting a document or asking a specific question." The impact is significant – Osaka Gas reduced time spent searching for information by 35%, while other companies save an estimated 10 minutes per search across hundreds of daily queries.
Tome's AI Sales Assistant helps sales teams quickly research target accounts by synthesizing vast amounts of data from company websites, financial filings, and news. "Time savings is huge," notes Sarin Devraj, Tome executive. The assistant identifies key strategic initiatives and recommends how to position products accordingly. What used to take hours of manual research now happens automatically, allowing sales reps to engage with more relevance and speed
Amazon's Review Summarizer uses generative AI to condense hundreds of customer reviews into concise, theme-based summaries. Shoppers can instantly grasp key product attributes and common opinions instead of reading dozens of individual reviews.
Tome's AI Sales Assistant analyzing a target account, showing strategic initiatives and recommended positioning for a sales rep.
⠀Think about your customer journey and ask:
Could we summarize and synthesize several information sources to make our customers' research process easier?
Could we use hundreds of customer reviews to give a glanceable overview of a product?
Could we make it simpler for customers to compare our product offerings based on their unique circumstances?
What information overload are our customers experiencing that AI could help filter and prioritize?
Where in our customer journey do people spend excessive time searching or comparing options?
Which complex product specifications could be translated into simple, personalized recommendations?
Unless it involves a beloved hobby or an upcoming holiday trip, we rarely enjoy doing the research required for high-stakes complex purchase decisions. By making it easier, companies can become valuable partners early in the buying journey and take mind-share from AI tools like Claude or Perplexity.
2. Streamline Communication
With rare exceptions, customers don't particularly enjoy communicating with companies. I'm referring specifically to the back-and-forth dialogues customers are forced to have – those support calls, email chains, form completions, and chat sessions where customers must explain their needs and wait for responses.
As described earlier, we'd rather fulfill our needs than talk to customer service or our account manager.
How could you use AI to streamline the communication required between you and your customers?
Here's how innovative companies are minimizing communication friction with AI:
AppFolio's Realm-X Messages uses Claude to help property managers handle complex communications with residents, owners, and vendors. "Property managers spend up to half their workday on communications," explains Teddy Ho, Principal Product Manager at AppFolio. Their AI assistant incorporates resident history, property-specific policies, and maintenance records to generate contextually appropriate responses. The system has processed over half a million AI-suggested responses, reducing message response time by an average of 26 seconds while helping property managers save about 11 hours per week on communications alone.
Newfront's Benefits Assistant provides employees with instant answers about their insurance coverage, eliminating the need to wait for HR responses. "Whether it's questions about prescription sunglasses coverage or help during a divorce, the assistant handles sensitive inquiries privately and accurately," says Gordon Wintrob, Co-founder and CTO at Newfront. This AI-powered tool saves HR teams a full month of productivity annually by handling thousands of routine questions that previously required manual responses.
Mercari (a Japanese online marketplace) transformed product listing from a 30-minute manual process to a 2-minute flow. Sellers now upload a single image, and AI generates a complete description, categorization, and pricing recommendations, dramatically simplifying the communication of product details.
Mercari's AI-powered listing tool transforms a single product image into a complete marketplace listing with automated description, categorization, and pricing.
⠀Potent questions to ask include:
Where do our customers face the longest wait times or most frustrating communication hurdles?
What repetitive questions are consuming our support team's time that AI could handle?
How might we use AI to prepare both sides of a conversation (customer and employee) to make interactions more efficient?
Could we translate customer messages into actionable internal requests automatically?
What communication gaps exist between technical teams and customers that AI could bridge?
How might we use AI to capture and organize information exchanged during customer conversations?
3. Simplify Choices
Modern consumer society often brings about analysis paralysis.
We have to choose between 50 different athletic shirts, 20 email marketing providers, or 80 brands of morning cereal.
How could you use AI to simplify choices for your customers? To boil them down to a handful of smart, personal recommendations provided alongside credible reasoning?
Here's how leading companies are using AI to cut through choice overload:
Spotify's AI DJ transforms music discovery by curating personalized playlists with contextual commentary, eliminating the need for users to spend time browsing through millions of songs. This reduces decision fatigue and has led to 25% of listening time being spent with the feature among users who try it.
Newfront's Contract Review Tool transforms days of back-and-forth emails into instant insights about insurance coverage gaps. CFOs and legal teams get immediate recommendations on policy adjustments rather than spending days reviewing complex documents. The result is that contract reviews now take minutes instead of days, eliminating the need for time-consuming calls and email chains.
Upwork's Job Post Generator has reduced the time to create effective job listings by 80%. What typically took 15 minutes of careful writing now takes 2-3 minutes with AI assistance, simplifying a crucial decision point for clients seeking freelancers.
⠀Relevant questions to ask about simplifying choices with AI:
Which decision points in our customer journey cause the most abandonment or hesitation?
How might we use AI to provide personalized recommendations based on a customer's specific needs and past behavior?
Could we create a conversational interface that guides customers through complex decisions step by step?
What data points would help our AI make truly relevant recommendations instead of generic ones?
How might we use AI to explain the reasoning behind recommendations to build customer trust?
Could AI help customers articulate their needs more clearly to make better choices?
Friction as a Starting Point
It's easy to forget about the friction our customers face. As professionals, we know our products inside out. Personally, we might be very different from our customers.
That's why it's crucial to return to a fundamental question in design thinking: What are the biggest points of friction or frustration for our customers?
I've worked with companies like KONE to identify customer challenges to discover potential areas for AI-enabled innovation.
The key is to focus on the biggest points of friction and solve them first with the most obvious and easy ways. AI can help in new and surprising ways, but it shouldn't be the main focus or starting point.
Often, it’s not about replacing human ability but augmenting it with AI to reduce friction where possible.
As Newfront's Gordon Wintrob puts it, "We see it differently—the insurance industry needs both human expertise and advanced technology. Our platform uses AI to handle routine tasks so our brokers can focus on what matters: providing strategic guidance and solving complex problems for clients."
Quick Self-Assessment
Before you jump into your next strategic design meeting, take five minutes to ask yourself these questions:
What's the most time-consuming research task our customers face before they can use our product effectively?
Where in our customer journey do people struggle to express their needs or understand our information?
Which decision points cause the most hesitation, abandonment, or post-purchase regret?
If you were to shadow a customer for a day, which interaction with your company would likely frustrate them most?
The friction points that immediately come to mind are likely your best opportunities for meaningful AI investment—the places where you'll see real impact rather than just technological novelty.
