AI Meeting Assistant Reviews

I’ve been looking for the right meeting assistant for years. Excellent ones like Gong.io lead the industry but cost thousands per month and aren’t affordable for many startups or small businesses.

For the last month I’ve been testing out two new AI meeting assistants: Read.ai and tl;dv. Fully loaded, both cost $25 per month, per user (when billed monthly).

Here’s what both apps do:

  • Automatically record calls on Zoom and Google Meet by joining the meeting as an attendee.
  • Transcribe calls using GPT to make the transcriptions much more accurate than anything pre-AI.
  • Automatically save meetings with timestamps for takeaways, topics, or questions covered in the call. Click on what you’re interested in and it plays that portion of the call.

Here’s what you need to know about Read.ai:

  • Seems like it’s designed more for enterprise teams, or for managers reviewing rep performance en masse. Scores meeting performance and participation. Gives feedback on attendee engagement, talking pace, interruptions, non-inclusive terms, bias and more.
  • For each meeting it documents: summary, chapters & topics, action items, key questions. The chapters & topics are usually accurate and helpful, but nothing else.
  • Meetings often don’t record, or nothing besides the transcript is documented. Each time, support told me it was a bug they’re fixing.

Here’s what you need to know about tl;dv:

  • Simpler approach: meeting reports contain “takeaways” throughout the call. AI does a good job of recommending these, plus they place a button in your meeting client so you can manually set markers with one click during a call. Great for when you know in real-time that something is important.
  • Integrations for automatically logging calls and highlights in HubSpot, Salesforce or Slack.
  • Easily create clips of takeaways or other key moments and share them.

The Verdict

tl;dv is my new everyday meeting assistant. It’s reliable, accurate, and has already saved me hours of sifting through sales/client calls to find information.

Read.ai is good too but worked on less than 75% of my calls. It also seems like it’s made for monitoring employees. These features don’t take away from the experience, but the bugs do. I’ll try it out again in a few months to see if they’ve ironed out the issues and improved the quality of action items and key questions.

tl;dv is offering 30% off your first 6 months if you click this link.

Read.ai did not offer a discount for our subscribers.

How will AI impact the sales industry?

Artificial intelligence (AI) is all over the news. Ever since the arrival of ChatGPT and AI art like Journey, everyone is asking one of two questions:

  • Will I be replaced by AI?
  • How can I use AI to get an edge?

Human sales reps won’t be replaced by AI anytime soon. AI can’t compete with humans when it comes to making phone calls, giving presentations, or providing a personal, confidence-inspiring sales experience. At least not yet.

Today’s buyers are the primary reason sales reps are safe for the time being. The technology will come, but when will humans be willing to interact solely with AI agents to close 6-figure deals? Similar to driverless cars, even if you’re told they’re safer, are you comfortable sharing the highway with 80,000 pound trucks driven by robots? People still prefer the company of other people.

So what is sales AI good for?

Today’s AI is incredible at comprehending language. This includes spoken languages and programing languages. This boils down to conditions and rules that, with enough training, AI can master. Since ChatGPT trained its AI model using pretty much the entire history of human text, it’s safe to say it’s receiving enough training. AI is excellent at researching as well.

Understanding language means AI can make the predictable aspects of sales much easier.

Digital sales experiences will change quickly. Well-trained AI can soon replace inbound reps who only work chatbots or inbound emails.

The same thing goes for operational sales processes like targeting or workflow automation. Targeting is a research task and automation is programming.

To clarify, the AI technology is already here (and improving rapidly), we’re more so waiting for companies to harness it and train it.

You can go on ChatGPT right now and teach it about your product, common objections and ideal responses, then have reps use those chat prompts to respond to leads. But, while that may improve response quality for some companies, it’s not convenient enough to be used at scale. We need products that interface with AI in order to use it at scale.

Here are some areas to keep an eye on:

  • Live Chat & Inbound Email — One of the most frustrating aspects of inbound sales is when reps find a way to mess up easy deals that should’ve been a quick close. AI does best with specific scenarios, reading and comprehension. So if buyers are coming to you ready to commit once you answer a few questions, AI can make a serious impact right away. One of my favorite prospecting tools, Reply.io has a conversational AI feature powered by ChatGPT that’s worth taking a look at.
  • Outbound Email — Writing personalized emails or introductions is time consuming, and most sales reps aren’t great writers. AI can do this much quicker than all sales reps, but the quality is inconsistent and therefore only better than some sales reps. Projects working on this include SalesGPT and Coldreach.ai but they’re not quite there yet. HubSpot is also working on an AI content assistant for prospecting emails, marketing emails and blogs.
  • Targeting — Check out this Twitter thread that uses ChatGPT to provide similar (if not better results) than a marketing research firm.
  • Operational ProcessesRead.ai and tl;dv are AI meeting assistants that outperform non-AI competitors. Bardeen integrates in many ways that Zapier and Make can’t hack. Browse.ai replaces costly scraper bots capable of tracking your competitors websites and pricing. HubSpot’s ChatSpot (in beta) gathers reports and sends follow ups. All of these tools buy back more time for sales reps to sell.

This How Smart is ChatGPT graphic shows where AI lands in all major academic exams (99th percentile in the verbal GRE, around the 90th percentile in the reading & writing SAT, LSAT and bar exam.) I noticed a response saying this means AI is “only as good as a smart college kid.”

But hey, as long as it’s predictable and consistent, that’s not so bad.