AI tools for Product Teams (part 2)
Let's dive in to the tools for qualitative aspect in the Sense, Listen, and Learn phase
Sense, Listen, and Learn - Qualitative tools
As we commented in the first part of this series:
Sense is about the work we do to understand our customers, our market, our technology, and how people use our product today.
You’ll do this by interviewing and observing customers. You’ll do this using product metrics that show who’s using your product, what they’re doing with it, and where they might be struggling.
In this article, we will be focusing on tools that could simplify and make efficient the interviewing and observing customers.
As customers use your product, you start to creating different mental models that simplifies the complexity, and build a shared understanding of who you customers is, their journey they go through and how your product plays a key role helping them to get the job done.
It’s paying attention to what customers and users do today that helps you identify customer’s problems and unmet needs.
Jeff Patton
Therefore, we will be inspecting and analysing tools that helps with the above; building shared understanding from interviewing and observing your customers.
Co-creation
Co-create or co-design the main artefacts that helps the team to keep the team aligned by building shared understanding are critical. It is the place where:
The team builds the idea of the main JTBD, customer profiles, segments, and user persona.
They visualise how customers are using the solution now. If they don’t use your product, how they are getting the job done today.
They visualise the ideal solution or state and how the customer will be better with it.
and more…
Figma
This post comes in with a very good timing, because by the time I’m writing this article, Config - Figma’s iconic annual event - has taken place two days ago.
They unveil a bunch of AI feature and brand-new products. For the sake of the topic we will be focusing only on the ones that regards to this topic. But I strongly recommend to see other presentations here.
What cool things we can highlight from the conference:
Everything from using prompts to generate designs, images, and components, to find a specific component and screens based on a screenshot and a drawing. The auto-renaming of the layers seemed to cause quite a positive reaction among the crowd as well.
Also you can create prototypes based on a list of screens. It will infer the order of the navigation and connect the screens automatically.
Hats off for the new product Figma Slides that also includes a sprinkles of AI as well, which I have to say it is my favourite.
Miro
Probably one of the tools, along with ChatGPT, that I use on a daily basis.
The main AI feature that Miro has introduced, and which in my opinion is much more powerful than FigJam AI capabilities, is Miro Assist. It is available once you click on any object within Miro:
The greatest thing about this feature is that it captures perfectly the object(s) you are selecting and it proposes relevant actions for the selection. For instance, if I choose a bunch of stickies, here will be the recommendation:
This is a basic User Story Mapping, and when you use the functionality “Ask Me Anything” to know the steps for a specific activity, it returns all the tasks related to it nicely.
Same if you choose panels with text. For instance, imagine you have a large text that is describing some process, here are the actions you can take with Assist:
See how the suggestions here are different from the previous selection. Future updates could include the image generation based on the text and combine it with the usage of templates available within Miro.
What happens if you select several pieces of text? No problems, Miro got you covered:
These, IMO, are the differences that make the difference. The fact that I cannot do the same with one or multiple text panels, makes me thing that someone cared deeply about the experience, and deliberately decided to reduce the number of options for a multiple selection.
As I praised the AI features for stickies and text, I don’t think it works well enough yet for images. You have the option of Ask Me Anything for an image. I introduce the prompt “Tell me what this image is about”, and I got the answer:
I tried many times with different type of images, pulled from Unsplash and icons that are within Miro. No luck this time. Although I see the benefit of having an image and get the meaning behind it, I don’t see myself using it a lot as the other features.
Whimsical*
Although this product could be in the Focus category as well, I’ve decided to include it here since it is a great alternative to the two juggernauts that dominate this space (Miro and Figma). These are the categories where whimsical competes:
Nonetheless, you can use it across the entire lifecycle to ideate, create an ideal flow, describe the parts with wireframes, and end up writing the specs for it.
In terms of AI features, you can access them very similarly as Miro. Interestingly enough, you can generate something using the Auto section, or the four highlighted before
With Whimsical AI you can:
Creating a mind map
Designing a flowchart
Brainstorming
Creating a sequence diagram:
Expanding the previous (mind map) diagrams:
Although the tool it looks very simple to use and smooth, I believe that in terms of AI capabilities is missing the connection between diagrams, and also no AI features are present in the document section.
Capturing customer insights
Dovetail
Dovetail is a qualitative data analysis. It acts as your research partner, to help you focusing on the insights by removing the tedious work. There are four features that caught my attention:
Transcription: nothing magical here compare to the tools that have transcription as a native capability because they are embedded as a bot within the call. Although it is not perfect, Dovetail can generate the transcript automatically when you upload the user interviews.
Summarisation: this feature claims to analyse the text transcript and auto-summarise key take aways, and it adds a timestamp to play back each section of the video. You can also edit the data point to incorporate anything you deem necessary.
Sentiment analysis: it analyses the transcript and identifies any words that have a strong positive or negative sentiment. I recommend to pay attention to the number of sentiments, it could be too indiscriminate and not quite accurate.
Theme clustering: allows you to group pieces of feedback that are similar either by tags you can create manually or by themes that are created by AI. This latter point has been taken to the next level by creating the concept of channels, where you can ingest a lot of data from many sources, and AI will suggest highlights that then can be clustered in similar themes.
Gong.io
Although this may be the most controversial tool to include in this post, I believe that more and more PMs should expand their reach to absorb information from Sales teams. Depending on the company you are, you will get more or less access to sales calls. But nothing stops you from pulling data from tools like Gong, if it is available.
The most relevant AI features that have recently being released in this product are:
Generative AI Insights: It automatically gives you critical insights from your customer interactions, like customer pain points, outcomes, and next steps.
AI Smart Trackers: It uncovers crucial trends happening in your customer interactions. The promise is that you can train your own AI models to listen to the most important concepts your teams and customers mentioned in only 30 minutes, and it claims to be more accurate than keyword matching.
Transcription: Gong claims their speech-to-text models achieve 85-90% accuracy, significantly better than off-the-shelf models. This makes it easy for global teams to share insights and gain a contextual understanding of customer accounts, no matter what language is spoken.
Transcription service
tl;dv
I’ve been using this tool recently among other transcription services, such as Otter.ai, and I have to say that I like it a lot.
Couple of feature that I think are great for capturing and analysing meeting notes effortlessly are:
Highlight / Timestamp a moment during the meeting: this works nicely with the transcript feature, to quickly find the information quickly about a specific topic and also with the summarisation feature, to relate questions to the moments that are pinned.
AI Tags templates: the tool automatically assigns tags to your meeting notes. You can create groups of tags and define to which kinds of meetings they should be automatically applied to. This is quite relevant for PMs since we attend meetings with a wide variety of stakeholders.
AI Coaching Hub: brings personalised scorecards to improve their skills and see how everyone’s doing, plus guides to handle tough questions easily. You can create playbooks to make sure that, depending on the conversation, you can assign a score to make sure you cover that topic:
AI Reports: allows you automate the creation of reports, based on a specific prompt, frequency, day, and time of the meeting (1). It will send out the summary based on the criteria specified (2), and also can be shared on many tools (3). Very cool! No more going back to notes or asking others about the things that have been said. It even has a feature that will allow you to compare with the previous meeting (4), which is a huge delighter to make sure you cover the items from previous meetings.
Integration: although this is not an AI feature, I love the idea of automating the workflow integration of notes into the spaces you have defined for it. This is powered by Zapier, by showing all the existing connection between tl;dv and other tools. This completely eliminates the need for a note taker.
Conclusion
There are three main ideas that we can draw from analyzing these tools:
These tools are stressing the idea of working asynchronously. You don’t need to read all the post-its and text that are there. Also, you can make it more readable before your team mates access the content.
These tools are stressing the idea of working more efficiently. You should only focus on what matters which is executing on the right insights and show customer and business results.
These tools are stressing the idea of integrating many aspects of the workflow. Although we have seen that most of them are doing similar things, the power of keep simplifying the workflow from the moment you interview someone and use that in a PRD is essential for tearing silos down.
These tools are stressing the idea of doing more with less. I’ve never been a fan of the super PM, but what these tools are bringing are the enhancing superpowers to do more in a shorter period of time and less number of people.
In the next article, we will be paying attention to tools that are related to Focus. This phase includes vision, product strategy, objectives, and roadmap: ChatPRD, ProductBoard, airfocus, Chisel, and Reforge browser extension.