Designing Aki

0 → 1 → Scale

Business acumen

AI product design

Complexity → Clarity

Aki

Nobody requested it. Everybody kept it.

The permission-first AI assistant
that plans tasks and calendar events by voice.

Team:

4 Designers

4 Devs

Product Manager

Marketing Manager

Management team

Timeline:

2-week sprint cycle to core release

Q4 2024

1.5 years of market validation and product evolution

Q1 2025 – Q2 2026

Product

Product Design

Marketing

Product Growth

My end-2-end role

Akiflow is an all-in-one task-and-calendar manager

With 10K+ busy professionals users, based mainly in the USA.

The Akiflow app integrating tasks from multiple tools into the Universal Inbox, displayed next to the One Calendar, that displays multiple calendars in one view
The Akiflow app integrating tasks from multiple tools into the Universal Inbox, displayed next to the One Calendar, that displays multiple calendars in one view

Universal Inbox + One calendar (Google + Outlook calendars) in one view

With a command line that lets users create tasks and meetings just by typing

With a command line that lets users create tasks and meetings just by typing

With a command line that lets users create tasks and meetings just by typing

The distinctive feature that made Akiflow known in the productivity market.

Calendar Command Line Interface with Natural Language Processing
Calendar Command Line Interface with Natural Language Processing

The request

Transform Akiflow into a productivity assistant by adding an AI layer that helps users plan deeply within their calendars.

Transform Akiflow into a productivity assistant by adding an AI layer that helps users plan deeply within their calendars.

Transform Akiflow into a productivity assistant by adding an AI layer that helps users plan deeply within their calendars.

Business goal

At least 70% adoption of the existing active users.

Technical goal

Define and release an MVP quickly, then evolve it.

The Akiflow community was not asking for it.

Users were focused on long-standing product feature requests.

Akiflow User features request

The problem

Introducing an AI assistant began as a strategic bet.

The value proposition

It was an opportunity to create one of the first specialized assistants in the entire productivity market.

It was an opportunity to create one of the first specialized assistants in the entire productivity market.

It was an opportunity to create one of the first specialized assistants in the entire productivity market.

The research

Continuous Discovery

Jobs-to-be-Done mapping

Value-gap analysis

Executive strategic influence

Context: the Competitors Map

The initial discovery phase.

The context analysis confirmed the existence of general-purpose assistants. The community confirmed real needs rather than assumptions.

The context analysis confirmed the existence of general-purpose assistants. The community confirmed real needs rather than assumptions.

The context analysis confirmed the existence of general-purpose assistants. The community confirmed real needs rather than assumptions.

30+

User interviews

3

Community surveys

9

Existing user drivers re-validated

Upon hearing questions regarding AI, users responded by asking instead for a stable and integrated experience.

Upon hearing questions regarding AI, users responded by asking instead for a stable and integrated experience.

Upon hearing questions regarding AI, users responded by asking instead for a stable and integrated experience.

Across desktop and mobile applications.

Akiflow’s community was involved.

"

The Mobile app it's almost useless. I can only check my agenda on it and create tasks or meetings in a very unintuitive way.

"

But I'm missing a feature that has been requested since four years (!): it is synchronization with apple calendar besides google and outlook calendar.

"

The app used to work decent; there were / are hiccups, but now the app doesn't sync to the to-do list, and it wants me to manually put them in.

From Trustpilot

Mobile was the unlock

Enabling planning on the go, with a personal coach.

Designing an AI SERVICE

The service identity and positioning

The service identity and positioning

Cross-functional collaboration

Brand management

Product storytelling

At the time, assistants were often character-based.

The goal was to create an intelligence, not a mascot to speak with.

The goal was to create an intelligence, not a mascot to speak with.

The goal was to create an intelligence, not a mascot to speak with.

The service levels

The service levels

Complexity → Clarity

Product strategy

Design Ops

MVP definition

Roadmapping

An intelligence that could play multiple roles.

An intelligence that could play multiple roles.

An intelligence that could play multiple roles.

01

Passive

Butler

Silently cleans up the clutter, by auto-assigning projects to new tasks. No confirmation needed, fully reversible.

02

Reactive

Assistant

Conversational.

Executes the user's commands with precision.

03

Proactive

Coach

Sending nudges to protect daily and weekly productivity, ensuring nothing slips through the cracks.

Users never became aware of these roles. This division remained for internal use only.

To be delivered in multiple releases.

To be delivered in multiple releases.

To be delivered in multiple releases.

Brainstorming output: organizing features for an MVP release.

And an MVP to be released in 6 sprints.

And an MVP to be released in 6 sprints.

And an MVP to be released in 6 sprints.

4 sprints

Aki's core shipped

2 sprints

MVP-completing features

2 sprints

Test and iteration

2 sprints

Beta launch, merging waitlist

Based on a lean and fast-prototyping process mix:

Based on a lean and fast-prototyping process mix:

Based on a lean and fast-prototyping process mix:

Identify value + Brand identity creation

User drivers map + Brand identity extension

Fast prototyping with Marketing-ready outputs

Prototype refinement

Prototype refinement: Loom walkthroughs in closed Slack channels for decision-makers

Coding & QA

Beta release

Async narration, sync decisions, written accountability.

The roadmap

A journey that basically never stopped.

Day 0

The foundations

Fast prototyping

Brand growth

Storytelling

Design governance

The promotional video

All mockups and interactive content were conceived:

1

To visualize the concept and inspire the team with demo prototypes

2

To be reused for marketing promos

Day 0

The foundations

Fast prototyping

Brand growth

Storytelling

Design governance

The promotional video

All mockups and interactive content were conceived:

1

To visualize the concept and inspire the team with demo prototypes

2

To be reused for marketing promos

The design system

Design-system extension

Code-based design

Product growth

A scalable approach, ready to evolve.

Every design layer, followed a code-based nomenclature and token-based structure legible to LLMs.

Every design layer, followed a code-based nomenclature and token-based structure legible to LLMs.

Every design layer, followed a code-based nomenclature and token-based structure legible to LLMs.

Conversational and voice-recording components + in-chat calendar previews

Sprint 4

The Butler

MVP release

Technical literacy

High-fidelity prototyping

Design foundations

"A simple animated sparkle icon and a badge applied to a task to clean up the clutter."

"A simple animated sparkle icon and a badge applied to a task to clean up the clutter."

The Butler automatically assigns existing projects to tasks, while keeping users in control: they can edit or remove assignments.

The interactive prototype was created with the development team to evaluate whether transitions should be used during implementation.

Sprint 4

The Butler

MVP release

Technical literacy

High-fidelity prototyping

Design foundations

"A simple animated sparkle icon and a badge applied to a task to clean up the clutter."

The Butler automatically assigns existing projects to tasks, while keeping users in control: they can edit or remove assignments.

The interactive prototype was created with the development team to evaluate whether transitions should be used during implementation.

Sprint 6

The Assistant

Chat-based experience

AI workflows definition

Audio > Text UX

Understand and act on requests expressed in natural language.

The first mobile assistant concept

The initial list of inspiring workflows that could be created with just a prompt

Sprint 6

The Assistant

Chat-based experience

AI workflows definition

Audio > Text UX

Understand and act on requests expressed in natural language.

The first mobile assistant concept

The initial list of inspiring workflows that could be created with just a prompt

Sprint 8

The Beta release

Product launch

Data literacy

Continuous improvement

On November 9, 2024, Aki was launched on Product Hunt.

On November 9, 2024, Aki was launched on Product Hunt.

The results

4

Sprints to core

2

New Mobile Widgets

90%

Waitlist → Activation

70%+

Active User adoption

What was validated

Users really appreciated Aki itself, starting using it every day in many ways to plan tasks in calendar, asking fro new features like:

  • element previews in chat;

  • an MCP;

  • advanced search and voice features.

What wasn't

Users did not appreciate the following features that were deleted by the product in the next releases:

  • Next (suggested) actions;

  • Clean your inbox;

  • Morning / Mid-day / Afternoon check

Sprint 8

The Beta release

Product launch

Data literacy

Continuous improvement

On November 9, 2024, Aki was launched on Product Hunt.

The results

4

Sprints to core

2

New Mobile Widgets

90%

Waitlist → Activation

70%+

Active User adoption

What was validated

Users really appreciated Aki itself, starting using it every day in many ways to plan tasks in calendar, asking fro new features like:

  • element previews in chat;

  • an MCP;

  • advanced search and voice features.

What wasn't

Users did not appreciate the following features that were deleted by the product in the next releases:

  • Next (suggested) actions;

  • Clean your inbox;

  • Morning / Mid-day / Afternoon check

Sprint 12

The Productivity coach

AI add-ons

MLP definition

Launching the first paid add-on.

Launching the first paid add-on.

Suggest action items and next steps after meetings, create tasks from emails, and more.

Ensure messaging consistency
Ensure messaging consistency

Anne

Anne

Mark

Mark

Product launch alignment

Product launch alignment

Fast prototype: a proactive meeting assistant that helps users stay present in the conversation.

Aki takes care of notes, the full transcript, and the next steps users can plan or share.

Flight check-in captured

Planned in 2 hours

Forward emails to Aki to plan tasks and events on the calendar - Control-first service

Meeting follow-up ready

3 action items for you

Get email notifications when meeting notes are ready

Sprint 12

The Productivity coach

AI add-ons

MLP definition

Launching the first paid add-on.

Suggest action items and next steps after meetings, create tasks from emails, and more.

Ensure messaging consistency
Ensure messaging consistency

Anne

Mark

Product launch alignment

Fast prototype: a proactive meeting assistant that helps users stay present in the conversation.

Aki takes care of notes, the full transcript, and the next steps users can plan or share.

Flight check-in captured

Planned in 2 hours

Forward emails to Aki to plan tasks and events on the calendar - Control-first service

Meeting follow-up ready

3 action items for you

Get email notifications when meeting notes are ready

Evidence over attachment:
Aki became a new Akiflow UX pillar.

Evidence over attachment:
Aki became a new Akiflow UX pillar.

The Desktop experience

Evidence over attachment:
Aki became a new Akiflow UX pillar.

The Desktop experience