Making Slack less overwhelming
The basics at a glance
Otter.AI
Adobe Suite
Google Workspace
x1 Sr. Slack engineer
x1 ML engineer
• No access to Slack's internal company data or resources
I've felt the weight of information overload
• 100% (5/5) test users successfully created a channel summary in under 60 seconds
• 80% (4/5) of participants said it would help reduce overload in their workflows
• Slack released a similar AI summaries feature 9 months later (early 2024)
Focus restored: turning channel noise into actionable signal
Before: Endless threads
Hours spent scrolling through chatter and notifications to find project moving info.
After: Concise summaries
Key updates, action items and files distilled and delivered in seconds so users can get back to work.
Validated by the market: Slack launched a similar solution as a core feature
Slack users were drowning in noise...
The avg. knowledge worker spends one day a week hunting for updates or the right person to help.
High-volume channels bury decisions in chatter. Power users send over 1000 messages daily.
The problem: user POV
• Constant context-switching kills focus.
• Anxiety rises over missing key updates.
• Catching up feels like a second job.
The problem: business POV
• The brand's productivity promise is at risk.
• It devalues the platform by creating more work.
• It creates a major opening for competitors.
Meet Tanner, a busy product manager
STEP 1: He fires up Slack: introducing AI channel summaries!
STEP 2: He can choose up to x5 channels to summarize
STEP 3: General summary + action items come preselected
STEP 4: He selects the timeframe
STEP 5: He pops over to the summary inbox
STEP 6: He shares targeted info (not adding to the noise)
Want to see this feature set in action?
View Prototype >
Slack overload hits job roles differently
& understand what was going on with AI (early 2023).
Interviews: x5 diverse perspectives
FREELANCER:
ML Engineer
JUNIOR:
Project Manager
MID-LEVEL:
Designer
SENIOR:
Software Engineer
EXECUTIVE:
COO
My pragmatic research approach
Without access to Slack’s internal data, I relied on interviews, public research, and expert consults.
My 10+ years of scrappy, "figure it out as you go" freelance industrial design experience really helped me shine here.
I'm very comfortable working in low context environments finding the info I need to get to the next step and iterating my way forward.
INTERVIEWS (x5)
• How do different roles experience overload?
SECONDARY RESEARCH
SWOT - Competitor Analysis - Industry Trends
• Where's the business opportunity?
EXPERT TECHNICAL CONSULTS (x2)
• Is AI viable and what are the limitations?
Key discoveries
Constant context-switching was killing focus time and mental clarity.
Users hacked together personal notes and screenshots because Slack search wasn’t cutting it.
of channel activity
Senior execs craved high-level reports that distilled channel activity without the deep dive.
Users were willing to try AI but doubted accuracy and seamless integration.
A deep dive: raw interview + research insights
The perfect storm: key trends validating the need
In early 2023 while 42% of companies were still just exploring AI, the launch of ChatGPT signaled a massive shift proving the tech was ready for mainstream business application.
SOURCE: IBM Global AI Adoption Index: 2023
A 2022 report found that 62% of companies lacked clear asynchronous-first policies, creating a major communication gap for teams spread across different locations and time zones.
SOURCE: Buffer, 2022 State of Remote Work Report
By 2022, 70% of organizations came to rely on workstream collaboration tools as their primary means of communication, leading to information overload and lost productivity.
SOURCE: CIO Magazine, 2021
“It quickly gets overwhelming. Setting up channels for specific projects / topics or DMs for specific convos helps, but I find that no matter what I do,
it becomes a chore to keep up with everything.”
- Jesse K. / comment G2.com: Slack product review page
Strategy: prepping to explore both AI & conventional solutions
The goal here was to create a versatile, strategic foundation before ideation. Using the Jobs-to-be-Done (JTBD) framework as you see below, I translated my research insights into sets of clear plans that would allow me to rigorously explore both AI-powered and conventional solutions to the problem. (Showing a brief example below)
(Story maps = research insights --> jobs to be done --> loosely defined feature sets)
Activities & outputs
Going wide: exploring ideas & reducing risk
Flexing superpowers: from dozens of sketches to x2 viable concepts
This led to two promising directions:
1. AI CHANNEL SUMMARY TOOL: An on-demand tool allowing users to actively manage channel overload
2. "WELCOME BACK" DASHBOARD: An automated dashboard for users to passively catch up after time away
Technical validation: de-risking the AI direction
To ensure my concepts were viable, I consulted two experts. A Sr. Slack Engineer confirmed the feature was feasible within their architecture. The ML Engineer provided key insights into model limitations, giving me valuable design context.
The decision: why an AI solution was the right path
• USER-DRIVEN: It directly solved the core user need for on-demand information filtering.
• TECHNICALLY VIABLE: Early engineering consults had already confirmed it's feasibility.
• BUSINESS SMART: It strategically positioned Slack as a leader in the emerging AI productivity space.
• EEXECUTIVE VALIDATED: It fulfilled a direct request from an executive for "daily channel reports" providing a clear north star.
“From a senior leadership level, I would have just liked daily reports on each channel. I wish there was a better way for Slack to report back to me at the end of the day.”
- Erin T. / Startup COO
Activities & outputs
Usability testing: validating a new AI interaction
Concept to clickable test with x5 users
I built a mid-fi prototype and ran remote usability tests with x5 Slack users to validate x3 key hypotheses:
• LEARNABILITY: Could users quickly grasp this new Al interaction?
• INTEGRATION: Does the feature feel seamless within the existing Slack experience?
• VALUE: Would it actually reduce overload, or was it just adding more noise?
Most impactful testing change: targeted info forwarding options
The most significant core change I made was the approach to shareable summary content.
BEFORE: The prototype allowed users to forward a link to an entire summary.
AFTER: Sharing was restricted to only targeted sections, like a list of action items.
RATIONALE: During testing, I realized that forwarding an entire summary risked recreating the original problem: sending colleagues more noise. The project's core goal was to reduce information overload, not just shift it. Restricting sharing to high-value, targeted content ensures the feature remains a tool for clarity.
Other key updates from testing
FINDING: The initial onboarding pop-up was a single, dense block of text, which 60% of users found overwhelming.
RATIONALE: This directly contradicted the project's core goal of reducing cognitive load, so I split the explainer into two shorter pop-ups.
FINDING: During testing, "Action Items" was the most sought-after and asked for information.
RATIONALE: To speed up the workflow and deliver immediate value, I made "Action Items" a pre-selected default, as voiced by of 60% of testers.
(80% helped)
FINDING: 80% of users expressed anxiety about missing key information. My testing also revealed a need for an easily scannable inbox.
RATIONALE: To address this, I added Al-generated keywords under each summary, allowing users to quickly triage and find relevant content at a glance.
Key testing wins
Every tester (5/5) generated a summary with no errors, in under a minute proving the feature was highly efficient and had a low learning curve.
This represents a massive time savings vs. having to scroll through multiple channels.
4/5 testers responded extremely enthusiastically and expressed interest in using this daily as part of their Slack work ritual.
Users were responding positively and genuinely felt that a feature like this could help them.
Activities & outputs
From an individual problem to a teamwide challenge
Systems thinkink in action: How one comment reframed the problem
Erin’s feedback reframed the problem for me from a personal productivity issue to a collective cultural one, rooted in the interconnected patterns of team communication.
“The big thing for me with AI is, if it becomes flat, and I can't push it forward to the next person, it's not as helpful if only I can engage with it.”
- Erin T. / Interviewee
Sanity restored: channel summaries buy back time
Key results at a glance
Every tester generated a summary with no errors, proving it was highly efficient and had a low learning curve.
IMPACT: This represents a massive time savings vs. having to scroll through multiple channels.
Slack's similar product
x9 months after this project wrapped, Slack launched its own AI channel summaries feature (2024).
IMPACT: This confirmed the problem was a high-value strategic priority and validates the relevance of my design direction.
80% of testers were excited at the prospect of actually using this feature in their day to day Slack workflow.
IMPACT: This strong adoption signal directly addresses customer churn by solving a core pain point and increasing the product's value.
x9 Months later: Slack launches their AI summaries feature in early 2024
Slack AI channel summaries were announced months after I wrapped my project. This was proof I had identified a relevant business and user need!
"Being able to pull your team's action items, share them, you're just eliminating such like a communication backlog..."
"It just saves you so much time from having to find all of those action items, write the email, send it out, etc."
- Erin T. - Startup COO
“I immediately see how this would impact me, I've got 14 channels and 50 messages each all with threads in those channels..."
"Having a tool that can capture the action items, and then summarize them would be super helpful.”
- Andrew T. - Product Manager
I learned a ton; creating a new feature end-to-end, working my way through a big, ambiguous problem
Superpowers in action
My biggest takeaway was reframing the problem from an individual to a team-wide challenge. A key insight from a user test revealed that simply sharing summaries could recreate the noise problem for others.
This "systems thinking" moment led to the crucial decision to restrict sharing to only targeted, high-value information.
My industrial design background trained has trained me to de-risk ideas early. Before committing to a direction, I validated the technical feasibility of my concepts with a Sr. Slack Engineer and an ML Engineer.
Their feedback gave me the confidence to move forward with a solution that was both creative and viable.
Future focus: next steps & measurement
If this were a live feature, the next steps would be to measure impact by tracking KPIs like weekly adoption and user satisfaction scores. I would also build a user feedback loop to continually train and improve the Al model over time.
Feel free to email blutjens@gmail.com or connect on Linkedin.
Check out my other work!
