AI has caused a seismic shift in the way we work. At Marvin, we’re adamant that AI isn’t here to replace, but enhance our workflow. And AI note-takers are the personal assistant you’ve been pining for.
Whether you’re a student or a working professional — anyone who needs an accurate and dependable conversation record will benefit from using an AI note-taker. If you’ve ever hopped on a Zoom call and juggled taking notes while simultaneously driving forward the discussion and processing what the other party is saying, an AI note-taker is for you.
We’ll begin with a short history of note-taking, before jumping into the features and functionality to look out for when choosing your AI note-taker.
A (Very) Brief History of Note-Taking
Over centuries, man has endeavored to make his life easier (jury’s still out on that one). However, we can point to feats of engineering that have made our lives markedly simpler over the years.
Take mobility (we love this topic): to move from A to B, at first, man only had his feet. Over time, he saw a horse and decided “I’m going to tame and train this creature and ride to move faster.” Then the revolutionary wheel came along, and carriages followed. Engines replaced hooves (but still had horsepower), and cars became the mainstay in mobility on land.
Note-taking hasn’t quite had the same trajectory. Ancient cave paintings were replaced by scrolls of papyrus, pioneered by the Egyptians. Note-takers reveled in the advent of the typewriter and the derivative keyboard — these instruments made note-taking markedly easier and more efficient. Fast forward to the present — note-taking still requires you to jam on your keyboard while staying focused on what you’re hearing. Until now.
Note-Taking in Qualitative Research
Traditionally, a qualitative researchers’ job has always been a manual and cumbersome method of analysis. Interviews require bundles of preparation beforehand, copious amounts of note-taking during, not to mention guiding the participant along. The real meat of a researcher’s analysis occurs post interview. That’s where the real drudgery begins — depending on who you ask, it requires anywhere from 2 hours to a day to transcribe, analyze and code an hour long interview.
Qualitative research is iterative work — topics and themes evolve over time with each revision. How can AI expedite the process, cutting time required for mundane data processing so researchers can focus more on their analysis?
What is an AI Note-Taker?
Widespread adoption of voice activated controls was observed in the automobile industry (we’re running with the mobility example). Carrying out the simplest of tasks used to drive us up the wall. Voice recognition in our cars(and otherwise) has improved significantly over time due to leaps in the advancement of AI.
AI note-takers use Natural Language Processing (NLP) and machine learning to transcribe speech into text, capturing vital information from interviews, meetings, lectures and webinars. They generate a record of conversations and store it for later use. Access to a verbatim transcript refreshes people’s memory and helps attribute quotes to users that said them.
What’s under the hood of your AI note-taker?
Didn’t believe us about that mobility theme, did you?
NLP is a form of AI that allows people to interact with computers using human language (i.e. not code). Using NLP, computers analyze, understand and respond to humans using natural modes of communication – text and speech (think Siri or Google Assistant).
Large Language Models (LLMs) are algorithms trained on large datasets with a number of different parameters. LLMs can perform a host of NLP tasks — they go beyond merely transcribing interviews – they can answer questions, summarize documents and generate text (you might’ve heard of generative AI, a la ChatGPT and Bard).
Consequently, AI note-taking tools now go beyond merely transcribing your conversations and recordings verbatim. They can generate output from a transcript — summarizing lengthy transcripts and provide insights before you so much as lift a finger. Capabilities and performance of AI note-takers will only grow with the inclusion of more data points and parameters.
Interviewers of yesteryear were multitaskers – gone are the days of frantically trying to jot down notes on your laptop, capturing what someone else is saying, while simultaneously processing the content and driving the interview forward. Declutter your mindspace, sit back and let AI perform some of the tedious heavy lifting for you.
Really wish we had AI note-takers for all those Business Law lectures!
What Features Should AI Note-Takers Have?
Like buying a car, choosing an AI note-taker boils down to features that matter to you. Are you looking for speed and performance, or reliability, safety and comfort? All of the above? Think of the capabilities that matter most to your organization.
At Marvin, we love a good acronym (you may remember from our UX research repository overview). In this vein, we’ve come up with a memorization tool (although with an AI note-taker, you’ll probably never have to memorize anything again). It’ll help you decide what features and capabilities are most important for you and your team:
(DISCLAIMER: list is in no particular order of importance)
This messaging feels like the old days, where an old FBI warning would pop up before you watched the latest movie on DVD (remember those?). Some things change, and some don’t.
There’s a reason why law requires interviewers to ask participants for their consent before recording any meetings. To protect their identities and maintain confidentiality, concealing user and participant information is of utmost importance.
Sensitive information collected from interviewees is sacred — ensure your AI note-taker maintains stringent privacy and security and conforms to data security regulations. In that vein, Marvin is HIPAA, GDPR and SOC2 compliant, which covers a lot of bases. If you need additional compliance information, get in touch with our support team today.
Qualitative thematic research requires turning mountains of messy and unstructured data into neatly arranged tangible topics, metrics or themes. A good AI note-taker allows organized file management and easy sharing of files and snippets (learn more about getting the most out of your research repository). It’s helpful to be able to export and download transcripts into file formats of your choosing (such as TXT, DOCX, and RTF files).
(Editor’s note: We’re going to share quotes directly from G2 crowd-sourced reviews because, as always, we love elevating our users’ voice.)
Felicia appreciates time-saving with Marvin in handling numerous projects:
“Marvin is addressing several of our challenges by organizing multiple research projects and providing quick and accurate transcriptions of audio recordings. This saves us so much time and lets us focus more on in-depth analysis and extracting insights.”
To aid your thematic analysis, Marvin offers a couple features that’ll help categorize and organize your insights:
- AI-powered thematic analysis. Of course we have to start with the big one! We recently launched fully automated analysis to help Marvin customers finish research projects faster. Even better, you’ll unlock user insights you thought would be impossible to find because they’re locked in spreadsheet jail. Now you just have to upload emails, support tickets, call scripts, survey responses and all your open-ended feedback for Marvin to automatically analyze themes, emotions, and trends across large swaths of data. (Quick caveat that that AI-powered thematic analysis is only available for Marvin users on our Enterprise plan.)
- Discussion guides – help you navigate through interviews. Pre-interview prep usually involves jotting questions down on a hand-written note or a Google Doc. This helps facilitate discussion to unearth answers that they’re looking for. Marvin’s discussion guides allow you to prepare questions beforehand, and refer easily to them during the interview. Simply hit the “/” key to tag a participant’s response to a question, and it’s filed in the right place. No more chaotic organizing your thoughts post interview.
- Templates – Qualitative interviews don’t exclusively consist of open-ended free form type questions – researchers use likert scales and similar questions to get structured responses. These aid in visualizing the data to gauge user sentiment and thoughts. If you know some questions and responses beforehand, create templates for later use (shareable across your team and company).
File insights away in their rightful place using Marvin’s discussion guides and templates. Stick to the task and improve your efficiency — learn more about getting the most out of these features here.
AI note-takers don’t replace conferencing apps such as Zoom, Google Meet or Microsoft Teams. They supplement them. Invite your AI note-taker to your call and let them take care of transcripts and timestamps (more on this below) — so you can focus exclusively on the interview and the human being on the other side.
To hop on board, the best AI note-takers don’t require you to jump ship from your existing applications. At Marvin, we understand that everyone has a preferred toolkit that they’ve been curating for years.
With Marvin, import quantitative data from Qualtrics, Pendo and Google Forms to get the whole picture of the user journey. Conduct your usability studies with UserTesting — then bring the results into Marvin to understand what your end users are grappling with. Export insights to apps like Figma and Miro, Notion and Confluence. Convert notes to actionable items within Asana and Jira.
Efficiency is the name of the game — streamline your research process with integrations so you don’t have to wave goodbye to your current suite of applications.
All your data analysis and insights under one roof. Take a closer look at Marvin’s integrations here.
With transcription (see just below) taken care of, it frees up your time to take refined, in-depth notes while conducting interviews.
Marvin automates note-taking and synthesis in one fell swoop. Make quick notes on the fly or instead of jamming furiously on your keyboard — leverage auto-notes. Auto-notes bring many keystrokes down to one — just hit ENTER to create an auto-note. Our AI will create an editable summarized note for the block that you’ve chosen.
Marvin also facilitates live collaborative note-taking. Live stream interviews (with a few seconds delay) and invite your peers and coworkers to join in the note-taking process. The more viewpoints, the merrier (as long as you don’t overwhelm the user!).
John says, “The collaborative live note-taking frees you to focus on listening and guiding the interview.”
With LiveNotes, the ability to go back to time-stamped notes is vital. Ever scribble something down and when referring back to it — it looks like chicken scratch and you can’t place what your thoughts were? Understanding the right context is important — what did the person say exactly that triggered the initial thought? Tracking back to exactly what a user said and when is helpful to ascertain not only their words, but their facial expressions and body language — that’s path traceability.
The LiveNotes save our customers lots of time in post interview work. They can take notes that sync to the exact moment in the call’s transcript, making it easy to come back to the notes and refer to the original moment.
The bread and butter of an AI note-taker — generating verbatim transcripts from your conversations. A robust AI note-taker must facilitate smooth recording and transcription of user interviews. Tight integration with video and audio conferencing applications is a necessity. The note-taker must be able to handle any file format you throw at it.
Jessie waxed lyrical in her review about Marvin’s transcription.
“I appreciate the integration of Zoom, and the machine-learning enabled transcription is excellent. Light years ahead of any other tools in this space,” she said.
Some tools on the market offer live transcription, and while this is impressive, it can lead to errors (see “H” below). Marvin generates transcripts soon after you hit “end call” — you’ll get a notification when your transcript is ready for use.
Replacing a human task with AI frees up a researcher’s time for more in-depth analysis. John says, “Marvin lets me conduct market research and quickly transcribe and analyze. With transcriptions alone, I can get it within minutes instead of days. In the future, I hope to reduce the analysis time from hours to minutes.”
To break down language barriers, look for an AI note-taker with multi-lingual support (some only offer English). Even countries that speak the same language might have subtle differences in their dialects. Marvin supports a library of over 40 languages.
We’ve been harping on about this for so long, we’re on the lookout for synonyms of “share”, “sharing” or “shareability”.
Siloed information creates duplicative efforts and inhibits the ability to act on rich user insights. Look for an AI note-taking tool that has the capability to be encompassing (i.e., one that has the capacity to reach far and involve everyone in a company). Someone stands to gain a better understanding of users by the simple sharing of data.
This doesn’t necessarily dictate that everyone must have the tool, or be extremely proficient with it (refer to the last item on this list). Choose an AI note-taker where insight sharing spans organization-wide with no gated community for license holders only.
Perhaps its most important feature is the ability to easily clip interviews, stitch them together and share playlists with your peers (whether they use Marvin or not). Uncover insights and let others make their own inferences from recordings — it generates new ideas and encourages healthy debate. Read Lou Rosenfeld’s take on why the whole is greater than the sum of its parts.
So what are you waiting for? Circulate your conclusions amongst your colleagues. Disseminate new information to the organization. Distribute your research insights to your peers.
P.S. How’d we do with the new vocabulary?
Who wants low accuracy anyway?
This one’s not really a feature — accuracy is more about companies delivering on what they advertise at the outset. We’re all too familiar with the conundrum of overpromise-underdeliver. Ideally, you want an AI note-taker that outputs reliable and consistent performance — the best tools do what they say they do and aren’t erratic.
Ultimately, this impacts the usefulness of an end product — higher precision in AI’s interpretation leads to more valuable and actionable insights. The opposite is true with inaccurate note-takers — you end up spending more time unraveling inconsistencies, errors and incomplete information.
A sensible comparison would involve putting the claimed accuracy of all AI note-taking applications side-by-side. Look at user reviews to see whether they got what they wanted or not. Be on the lookout for an AI note-taker that does the following accurately:
- Speaker Identification – as the name suggests, it must correctly attribute quotes and distinguish between speakers. Separate blocks must be created in the transcript for each speaker.
- Filler words – not everyone’s a born orator. Filler words such as “ah”, “um”, “eh”, etc., must be removed from the transcript so you spend less time cleaning it up.
One of our users said this about Marvin’s transcription:
“The transcription is far better than I have ever encountered with other research tools. It makes the process so much easier and especially helpful if you are a small team and need to democratize. The ability to upload data (videos, texts), have videos automatically transcribed (extremely accurately), tag, analyze, and even make highlight reels of videos. It’s so awesome!”
Furthermore, Marvin’s interactive text editor allows you to correct, revise, highlight and annotate textual errors in the transcript. Marvin uses these annotations and edits to learn for future interviews.
“The transcription quality is AWESOME and available in minutes! So much better than other services. I only end up correcting words that are industry-specific or names,” quips John, another Marvin user.
Transcription aside, AI is a wide-ranging force of nature. With your AI note-taker, look for assistance in automating certain tasks and time-saving productivity features.
Interview transcripts can be lengthy documents, full of relevant and sometimes irrelevant details. Even skim-reading can be time consuming. We know from our primer on LLMs (above) that AI can generate summaries of large datasets. Use Marvin’s AI summaries to garner key insights and gauge sentiment from numerous transcripts. Skim read through mountains of interviews in no time.
Another Marvin user was quite taken with our AI:
“Every time I log in, I’m blown away by Marvin’s AI because it not only automates — it improves the outcomes of traditionally time-consuming tasks.”
AI summaries are used in auto-notes (from above), where with just one keystroke, quick notes are generated (post processing) on the fly. Use Marvin’s AI to create notes, actionable items and insights from an interview transcript. It also categorizes notes so that you don’t begin with a blank canvas – some skeletal analysis is done for you before you even begin.
Google’s spoiled us in a way. Our instincts have moved from jogging our memory banks to whipping out a smartphone to ask for answers.
Once transcripts are generated, searchability is key. What’s the point in creating pages upon pages of transcripts if they’re not intuitively searchable?
Marvin works exactly like Google when analyzing your data repository. Search your research repository of quantitative and qualitative data (e.g., notes, transcripts, recordings, surveys) with Marvin’s AI-powered search. Search for specific keywords, and Marvin will scour transcripts, tags, discussion guides and templates to return the most relevant results.
It’s far easier to search by a keyword than to find a spot where an interviewee mentioned a topic in an hour-long video. Time-stamping to the rescue.
Easy to use
Last but not least, your AI note-taker must be user-friendly with a shallow learning curve.
In the words of the late, great Mac Miller, “why does it always gotta be so complicated?”
Everyone’s purchase history contains a gadget, item or application that once opened, was never touched again due to its sheer complexity. An AI note-taker is the same — it’ll likely just sit there if it’s too convoluted to learn.
An easy-to-navigate interface, coupled with a quick onboarding process, allows users to spend less time figuring out the ropes from a manual and more time testing its features.
Ideally, you want one that appeals to the technophiles in a company, but doesn’t deter the prehistoric “Fred Flintstone” types.
Enjoyed our guidance for choosing an AI note-taking tool? Check out our comprehensive guide on choosing the right Market Research tool.
AI Note-Takers: Your New Favorite Research Assistant
Your multitasking days are over. Use your new AI note-taker to streamline your workflow and optimize the interview process.
Now that you’ve got a handle on what features to look out for, and which ones are more vital to your process, allow us to advise you on a great place to start. Marvin offers a host of features that make it stand alone amongst the crowd.
Don’t take it from us. Listen to what Travis said about using Marvin:
“In my line of work, I run lots of customer interviews. There is no product on the market that is better at helping analyze and synthesize customer interviews. Additionally, Marvin helps highlight insights that would be much harder to see without their helpful visualizations.”
We couldn’t have put it better ourselves. Make Marvin your ultimate research companion. Set up your free demo today.