AI Weekly Briefing: Lovable Raises Big and Salesforce Buys Qualified

AI is getting embedded everywhere and this week proved it. Salesforce is buying Qualified to power always on AI agents for pipeline. Lovable just raised big money at a massive valuation as vibe coding explodes. Microsoft and Google are leaning on partners to push AI from pilots to production. Notion reveals that AI now drives a huge chunk of revenue, and Zoom launches AI Companion 3.0 to turn meetings into actions.

But there’s a catch: Reuters reports that many companies still aren’t seeing the ROI they expected. So what separates hype from real value?

Episode Transcript

(00:00):
Hi everyone.

(00:00):
I'm Andy and this is the AI Breakdown.
Welcome to your weekly news edition where I'll cover what happened in AI last week, why it matters all in about 10 minutes. 4 00:00:14,88.988928748 --> 00:00:15,828.988928748 Alright, let's do the weekly scan. 5 00:00:16,398.988928748 --> 00:00:18,678.988928748 And honestly, this week had a really clear theme. 6 00:00:19,188.988928748 --> 00:00:24,768.988928748 AI is getting embedded as the thing that actually runs parts of revenue, workflow, and infrastructure. 7 00:00:25,278.988928748 --> 00:00:32,148.988928748 First up, Salesforce is buying qualified, a San Francisco startup that does what they call a agentic marketing. 8 00:00:32,598.988928748 --> 00:00:38,718.988928748 The headline product is an AI agent called Piper, and Piper's job is wonderfully unglamorous. 9 00:00:38,718.988928748 --> 00:00:50,508.988928748 It sits on your website and talks to inbound visitors over text, voice, even video answers, questions, figures out intent, nurtures the lead, and then books, meetings for your sales reps. 10 00:00:50,928.988928748 --> 00:00:53,718.988928748 Salesforce says it'll fold this into agent force. 11 00:00:54,63.988928748 --> 00:00:58,233.988928748 So customers can deploy always on agents for pipeline generation. 12 00:00:58,653.988928748 --> 00:01:08,223.988928748 The deal's meant to close by early 2026 and it's a pretty tidy fit because qualified was already a Salesforce ventures portfolio company and an app exchange partner. 13 00:01:08,553.988928748 --> 00:01:12,363.988928748 So this isn't some random bolt-on from Left Field SASters. 14 00:01:12,363.988928748 --> 00:01:19,743.988928748 Jason Lemkin called it the smartest AI GTM deal we've seen so far in 2025, and he's not exactly known for underreacting. 15 00:01:20,298.988928748 --> 00:01:31,728.988928748 He even said that in his own organization, combining qualified ai, SDR with Agent Force had already booked 100 plus meetings and over $1 million in closed deals in a few months. 16 00:01:32,58.988928748 --> 00:01:35,928.988928748 And Salesforce's own Steve Fisher put it in Peak corporate speak. 17 00:01:36,588.988928748 --> 00:01:42,103.98892875 The agent ification of the enterprise continues to accelerate, but translate that into English and it's. 18 00:01:42,753.98892875 --> 00:01:47,703.98892875 Your top of funnel is getting automated and revenue teams might scale without scaling headcount. 19 00:01:47,733.98892875 --> 00:01:55,83.98892875 At the same rate, if you're a SaaS founder, this is the direction of travel AI that touches pipeline and pays for itself. 20 00:01:56,732.43425842 --> 00:02:03,602.43425842 Next up, lovable the Swedish Vibe coding startup just raised $330 million at a $6.6 21 00:02:03,602.43425842 --> 00:02:05,222.43425842 billion valuation. 22 00:02:05,672.43425842 --> 00:02:07,442.43425842 That's more than triple in five months. 23 00:02:08,72.43425842 --> 00:02:13,352.43425842 Lovable lets non-programmers build web apps and websites using natural language prompts. 24 00:02:13,772.43425842 --> 00:02:14,822.43425842 Just by speaking. 25 00:02:15,392.43425842 --> 00:02:22,712.43425842 It's a fast growing and super interesting trend that I've previously covered on the podcast with lovable being the poster child for this movement. 26 00:02:23,102.43425842 --> 00:02:25,472.43425842 The growth numbers are the bit that makes you sit up. 27 00:02:25,802.43425842 --> 00:02:35,192.43425842 They hit 100 million ERR within eight months of launch and then cleared 200 million ERR just four months later, and they're not just a toy. 28 00:02:35,522.43425842 --> 00:02:40,712.43425842 Over 25 million projects have been built on the platform and customers include Uber. 29 00:02:41,87.43425842 --> 00:02:44,117.43425842 Klarner Zendesk and Deutsche Telecom. 30 00:02:44,777.43425842 --> 00:02:54,857.43425842 Whilst I still don't believe vibe coding tools like this are ready or safe for C production systems, they can make a meaningful difference contributing to the product development process. 31 00:02:55,367.43425842 --> 00:03:06,197.43425842 For example, they have reference one case study with a global ride sharing platform, most likely Uber that cut a design testing cycle from six weeks to five days using lovable. 32 00:03:06,827.43425842 --> 00:03:09,227.43425842 Investors are framing this as a fundamental shift. 33 00:03:09,632.43425842 --> 00:03:20,12.43425842 Menlo Matt Murphy said, lovable has done what was previously unimaginable by turning a latent market of tens of millions of people into web developers and content creators. 34 00:03:20,552.43425842 --> 00:03:27,992.43425842 Capital G's Lay aur said, demand from Fortune 500 companies signals a real shift in how software gets built. 35 00:03:28,532.43425842 --> 00:03:33,482.43425842 My take on this is the valuation assumes lovable becomes a proper platform. 36 00:03:33,872.43425842 --> 00:03:37,922.43425842 Integrations, hosting databases, reliability. 37 00:03:38,192.43425842 --> 00:03:39,92.43425842 That's the hard bit. 38 00:03:39,662.43425842 --> 00:03:40,802.43425842 The direction is clear. 39 00:03:41,102.43425842 --> 00:03:51,572.43425842 Developer shortage is turning into builder abundance, and if they get this right, it opens the doors for a new age of technology entrepreneurs without the need for real tech skills. 40 00:03:53,138.19157815 --> 00:03:53,468.19157815 All right. 41 00:03:53,468.19157815 --> 00:03:59,258.19157815 Partnerships, and there were two big ones that I think fit into the same narrative Everyone's trying to solve. 42 00:03:59,258.19157815 --> 00:04:00,968.19157815 The pilot two production problem. 43 00:04:01,298.19157815 --> 00:04:05,228.19157815 First, cognizant of Microsoft announced an expanded multi-year partnership. 44 00:04:05,603.19157815 --> 00:04:25,763.19157815 To help enterprises implement generative AI at scale, they're targeting industries like financial services, healthcare and life sciences, retail and manufacturing, and the focus is very much on embedding Microsoft copilot and agent AI into actual workflows, think claims processing, supply chain tools, that sort of thing. 45 00:04:26,423.19157815 --> 00:04:33,323.19157815 Cognizant, CEO, Ravi Kumar said, today, AI underpins and shapes every transformation program we drive. 46 00:04:33,533.19157815 --> 00:04:38,753.19157815 And frame the goal as solving the last mile challenge of scaling AI across the enterprise. 47 00:04:39,203.19157815 --> 00:04:47,393.19157815 Microsoft's Judson Altoff said they want to accelerate the creation of industry specific solutions that are embedded in the flow of work. 48 00:04:48,83.19157815 --> 00:04:57,863.19157815 The bit that stood out is Cognizant plans to upskill 1000 plus engineers on Microsoft's AI tech and roll, Microsoft's copilots out internally as well. 49 00:04:58,223.19157815 --> 00:05:00,533.19157815 That's essentially build in delivery capacity. 50 00:05:01,193.19157815 --> 00:05:01,973.19157815 Second partnership. 51 00:05:02,648.19157815 --> 00:05:07,538.19157815 Tech Mahindra is teaming up with Google Cloud to help enterprises adopt Gemini 2.5, 52 00:05:07,958.19157815 --> 00:05:17,168.19157815 and specifically to build next generation AgTech AI solutions with governance, security, and industry customization baked in tech. 53 00:05:17,168.19157815 --> 00:05:26,558.19157815 Mahindras Amal said, enterprises need secure, scalable, and enterprise ready intelligence as they move from experimentation to transformation. 54 00:05:27,158.19157815 --> 00:05:33,38.19157815 Google's Victor Morales called a agentic ai, a powerful opportunity to reshape business models. 55 00:05:33,488.19157815 --> 00:05:43,28.19157815 My opinion on both, this is the consulting industrial complex, retooling itself around foundation models, and honestly, enterprises want that. 56 00:05:43,568.19157815 --> 00:05:45,458.19157815 Most companies don't need another demo. 57 00:05:45,878.19157815 --> 00:05:50,888.19157815 They need someone to wire this into their data, their compliance, and their messy reality. 58 00:05:52,265.40356138 --> 00:05:54,515.40356138 ​Two more product stories and then I'll zoom out. 59 00:05:55,130.40356138 --> 00:05:59,180.40356138 Notion quietly dropped one of the most important business metrics in AI this week. 60 00:05:59,660.40356138 --> 00:06:04,910.40356138 They're now over $600 million in annual recurring revenue and roughly half of that. 61 00:06:04,910.40356138 --> 00:06:08,900.40356138 A RR comes from their AI powered products and features notion. 62 00:06:08,900.40356138 --> 00:06:10,310.40356138 Launched Notion ai. 63 00:06:10,310.40356138 --> 00:06:16,490.40356138 In late 2022 weeks before chat, GPT went mainstream and adoption just kept climbing. 64 00:06:17,180.40356138 --> 00:06:20,780.40356138 It started with 10 to 20% of customers paying for the AI add-on. 65 00:06:21,125.40356138 --> 00:06:27,755.40356138 Then 30 to 40% and now over 50% of Notion customers are paying for AI capabilities. 66 00:06:28,325.40356138 --> 00:06:39,155.40356138 Jason Lemkin take was when over half your customers pay for ai, you bundle it into business and enterprise tiers to expand average revenue per user without raising list prices. 67 00:06:39,515.40356138 --> 00:06:45,910.40356138 They're also doing a $300 million employee share sale that values the business at $11 billion post money. 68 00:06:46,505.40356138 --> 00:06:49,715.40356138 And there's chatter about a potential IPO in 2026. 69 00:06:50,495.40356138 --> 00:06:52,685.40356138 Whether or not that happens, the lesson is clear. 70 00:06:53,75.40356138 --> 00:06:58,85.40356138 AI that's properly embedded in an existing workflow can move revenue in a very real way. 71 00:06:59,395.18751664 --> 00:07:02,905.18751664 And then Zoom rolled out AI Companion 3.0. 72 00:07:03,385.18751664 --> 00:07:04,765.18751664 This is more than meeting summaries. 73 00:07:04,765.18751664 --> 00:07:15,325.18751664 Now it's got help me write for drafting documents or emails based on meetings and files, and an AI workbench where you can create custom workflow automations with drag and drop. 74 00:07:16,0.18751664 --> 00:07:17,200.18751664 Zoom's, CTO. 75 00:07:17,350.18751664 --> 00:07:21,850.18751664 Sue Don Huang described it as helping people from conversation to completion. 76 00:07:22,90.18751664 --> 00:07:30,430.18751664 And he even called the low code automation piece like agentic coding, but for everyone, Zoom's doing something smart under the hood too. 77 00:07:30,790.18751664 --> 00:07:38,290.18751664 It uses a federated mix of models, lightweight ones for quick context and heavier models when needed to keep costs manageable. 78 00:07:38,860.18751664 --> 00:07:41,740.18751664 And they're shipping a lot of this at no extra charge on most paid plans. 79 00:07:42,685.18751664 --> 00:07:45,925.18751664 Which tells you how competitive this space has become. 80 00:07:47,212.51465512 --> 00:07:53,812.51465512 Before we wrap, I want to add one more story from Reuters that I think is the reality check sitting underneath all of this. 81 00:07:54,232.51465512 --> 00:08:01,702.51465512 The headline is basically, after the hype companies are finding AI's return is falling short and their adjusting expectations. 82 00:08:02,302.51465512 --> 00:08:09,712.51465512 According to Forrester, only about 15% of firms reported any profit margin improvement from AI in the past year. 83 00:08:10,72.51465512 --> 00:08:15,322.51465512 And a BCG study found just 5% saw a widespread value from AI initiatives. 84 00:08:15,742.51465512 --> 00:08:28,552.51465512 There are some relatable examples, Klarna, which famously talked about in ai chat bot replacing 700 agents had to bring back more human support because the bot handled simple queries but struggled with nuance. 85 00:08:28,882.51465512 --> 00:08:33,802.51465512 And Verizon found around 40% of consumers still demand a human for complex problems. 86 00:08:34,657.51465512 --> 00:08:38,137.51465512 Forrester analyst Brian Hopkins said, tech firms spun this tail. 87 00:08:38,407.51465512 --> 00:08:50,407.51465512 That change would happen quickly, but we humans don't change that fast, and Forrester even forecast that in 2026, about 25% of planned corporate AI spending will be delayed by at least a year. 88 00:08:51,37.51465512 --> 00:08:52,597.51465512 I actually think this is good news. 89 00:08:52,717.51465512 --> 00:08:54,697.51465512 It means we're moving from hype to craft. 90 00:08:55,57.51465512 --> 00:08:58,87.51465512 The winners will be the teams who pick narrow high impact use cases. 91 00:08:59,47.51465512 --> 00:09:06,7.51465512 Do the integration work and keep humans in the loop where it matters, rather than trying to replace entire departments overnight. 92 00:09:06,397.51465512 --> 00:09:10,867.51465512 So if you connect the dots across all of these stories, the picture is pretty simple. 93 00:09:11,197.51465512 --> 00:09:15,127.51465512 AI is becoming less of a novelty and more of an operating layer. 94 00:09:15,667.51465512 --> 00:09:18,67.51465512 Salesforce is buying pipeline automation. 95 00:09:18,367.51465512 --> 00:09:21,247.51465512 Lovable is turning nons into builders. 96 00:09:21,637.51465512 --> 00:09:25,807.51465512 Microsoft and Google are leaning on integrators to get AI into real workflows. 97 00:09:26,212.51465512 --> 00:09:30,112.51465512 Notion is proving you can monetize embedded AI at scale. 98 00:09:30,532.51465512 --> 00:09:42,772.51465512 Zoom is trying to turn meetings into actions, and the ROI reality check tells you what matters next, not who has the fanciest model, but who can ship AI into messy organizations and make it pay. 99 00:09:44,539.66906073 --> 00:09:46,429.66906073 That's all for this week's AI roundup. 100 00:09:46,819.66906073 --> 00:09:50,839.66906073 If you found value in this breakdown, please leave a rating and hit subscribe. 101 00:09:51,229.66906073 --> 00:09:53,629.66906073 Have a great festive season and catch you next week.