Product Sense: Reverse Engineer Duet AI as a Case

Google recently announced Duet AI for Workspace - a chat assistant to improve your productivity. Looking at some of the features, I am skeptical it can really help. But three features in the list below stood out to me. I could imagine the interview prompt: Improve Google Meets.

First, I will evaluate items on the list against the price point: $30 per month. And then, I will imagine what an interview answer would look like that would get me the last three items on the list, the ones I believe will be most effective immediately. (They are impressive enough to possibly get me to prefer Meet over Zoom!)

Strategic Analysis

Would I pay $30 per person per month?

A list of some of the features:

• Generate images.

• Write email responses.

• Summarize documents and long texts.

• Scan apps like Gmail and Drive to find relevant information.

• Create presentations from text, charts and images in your files.

• Create notes, summaries, and action items during meetings.

• A “summary so far” for people who join a meeting late.

• Attend meetings on your behalf and take notes or deliver messages.

Let’s Look at the Features

The first five features in our list seem to be aimed at competing against products in the market or improving current solutions

• Generate Images

  1. Yes. It would be nice to have inline. I would need to test to see if it works like Midjourney. If the quality is weak, this is not a selling point; I would prefer an API to Midjourney.

  2. Competitors: Midjourney, Dall-E, Canva, Stable Diffusion

• Email Responses

  1. I am skeptical it will be much better than short responses in Gmail. But if it can anticipate the right longer-form answers, that would be great, but it also makes me worry about the value I bring. It would save me from a number of my existing templates. I have soo many it is difficult to find them all.

• Summarize Documents

  1. ChatGPT and other tools I pay for always fail me. I was willing to pay $10 per month for this, so if it works, they are 1/3 of the way to winning me mover.

• Using Drive Documents to Find Info

  1. Wow! If is truly works, will a manual search of my Google Drive content get better too? I will believe it when I see it. The fact that Google Drive search is awful makes me seriously wonder. I don’t know how I feel about having to pay for it, I have an expectation it should come with Drive.

• Create Presentations

  1. I seriously doubt this one. I imagine there is a limited set of topics they can really help with. Curious as to the set of problems they can solve here. Probably inline with they core user cases in most Slide Deck template solutions right now. I doubt V1 is great but imagine V3 will make my life easier.

Now let’s get into what appears to be identification of a specific set of problems: Meetings

• Create notes, summaries, and action items during meetings.

  • They have been doing this for a while. It's probably just as good as the competitors, which will get better but often leaves out things I really need. But reduce the load on me to keep track of the basics.

• A “summary so far” for people who join a meeting late.

  • Life-saver! I would pay $10 a month for this just to not have to tell people who join late where we are. I run coaching sessions where people come in late and I am focused on others only to find they were lost because they arrived late.

• Attend meetings on your behalf and take notes or deliver messages.

  • I can’t imagine how this will work unless the meeting is 30+ and I bet they end up having to build tools to prevent this behavior, but it is what I a user of the meeting software really want. The “Deliver Messages” seems fun to start but will probably get overused by executives and hurt morale.

In short, I think most small to medium-sized businesses will find the $30 per person a fair price given the increased productivity if they work as advertised. I think for a small business, I likely just need one or two features to work at least as well as the competition for me to make the move. I will need to try but I imagine it would be worth the $30 per month up until i start dishing out $100; then I start to ask: Is it worth it?

Reverse Engineering

Let’s imagine we got the prompt: Improve Google Workspace for Meetings.

To start, let’s map out the framework. This is how I would use the S.U.S.S. Framework.

• Strategic Setup

  • Clarifying

  • Insights

  • Goal Setting

• Users

  • Segment

  • Pain Point Identification

• Solutions

  • 3 Feature

• Success Metrics

  • 3 Metrics

Now, I will flesh out my answers for each stage of the framework:

• Strategic Setup

  • Clarifying: Workspace is Gmail, Meet, Docs, Sheets and Slides for SMBs as well as enterprises. A productivity suite. Is that correct?

  • Insights

    • AI Arms Race

    • Competing generally against Microsoft

    • Great opportunity to up the game and level the playing field.

    • Google is slower than startups but might be able to create more delightful experiences faster than Micosoft

    • Expectations outside of tech professionals are lower, basic feature improvements might wow them.

    • Great opportunity to innovate in this space against: Otter.ai, Slack, Grammarly, ChatGPT, and startups too numerous to name trying to use AI to improve White Collar productivity

  • Goal: Improve worker productivity

• Users

  • Segment

    • Secretaries

    • Project Managers While project manager is a profession, most office workers do some aspect of project management. Therefore, the TAM (total addressable market is high).

    • Managers (front-line)

  • Pain Point Identification

    • Calendar Management for meetings

    • Taking Notes in Meetings as a reminder and to keep everyone on the same page and allow for easy follow-up

    • Keeping up with people who arrive late - some of whom drag the meeting back by revisiting something already covered.

    • Being double-booked

    • sending follow-up meeting notes or specific emails

• Solutions*

  *for companies like Meta, you have to pick a specific pain point, but at Google you can get away with multiple pain point solutions. Let’s be a little more liberial with this case as the point is to think about how problem > solution pairs can be identified

  • 3 Features Built on Existing Products

    • Notes - AI that summarizes notes during the meeting, not just after. Allow the to be editable as you go.

    • People Arriving Late - Leverage real-time notes in a specific use case. Give those arriving late a summary to avoid interruption

    • Being Double Booked - When you need to be in two places at once, if you have a meeting where you can pre-plan what you say and/or just need highlights from a meeting, we can develop and AI that takes note and sends emails to attendees as a follow-up

• Google can already take notes after a meeting, add doing it during. Playing off this core functionality

• Success Metrics

  • Goal: Improve worker productivity

    • # people being shown a summary upon entering the room

    • # meetings with satisfaction store above 4.5

    • # users asking in chat or verbally about something already covered

I prioritized the first metric because it is something that can be measured objectively. There are too many other factors that can influence meeting satisfaction. # users question after arriving late will be very difficult to measure.

It is my hope this has shown you how to think about real-world solutions and reverse engineering the brainstorming process for your favorite/commonly used products.