For the first time, news aggregator Nuzzel will include advertising.

That doesn’t mean you’re going to see banner ads popping up all over the startup’s app and website. Instead, the Nuzzel Newswire consists of sponsored links in Nuzzel’s email newsletter, pointing to a blog post, press release or news article of the advertiser’s choosing.

So why call it a “newswire” when it’s really just sponsored content? Nuzzel founder and CEO Jonathan Abrams told me he’s pitching this as an alternative to paying for press release distribution through a service like PR Newswire, where he said “probably few people will really read” it.

“Of course you could do both,” Abrams said. “But the idea is that this is a way to get your content in front of influencers that a traditional press release service may not deliver, since they take anything and release so many press releases every day.”

In contrast, he said the Newswire will only run one sponsored story per day.

Nuzzel, for those of you who haven’t tried it, offers a number of ways to keep track of the news, whether that’s following feeds/newsletters curated by Nuzzel or other users, or by connecting your Twitter account to see the stories that are getting the most shares from the people you follow.

Abrams declined to reveal anything about the size of Nuzzel’s audience, except that “it’s fair to say that at this point, the real value of the Nuzzel Newswire is the quality of the audience vs. huge size” — a point underlined on the Newswire website, which highlights a few individual Nuzzel users.

As for why the ads are only running in the newsletters, Abrams said he might add them to other products in the future, but he’s starting with email “because we thought native ads in email newsletters would get really good engagement and not as many people were doing it.”

Aiden, a London-based startup building a machine learning-powered personal assistant to save mobile marketers time and money, closed a $750,000 seed round today from Kima Ventures and a number of angels, including Nicolas Pinto, Pierre Valade and Jonathan Wolf. The team first demoed the capabilities of its service on the stage of TechCrunch Disrupt as a Battlefield finalist. 

In recent years, marketers have begun to suffer information paralysis from the flood of new monitoring platforms on the market. Rather than create an entirely new end-to-end solution, Aiden merely wants to help marketers get at the data they need faster.

The Aiden conversational assistant aims to make querying for key performance metrics as easy as checking the weather on Amazon Alexa. From a technical perspective, the team is focusing their efforts in two places — natural language processing and expert systems.

The natural language processing work is intuitive; 90 percent of the challenge in building a system like Aiden is deriving the intent of any given user query. The time spent on NLP helps Aiden respond appropriately, no matter how a question is phrased.

The emphasis on expert systems, on the other hand, is a bit less textbook. Expert systems have been around in some form or another since the 1970s. The idea is that humans can encode their knowledge into computers that can later be recalled — creating the illusion of intelligence. But while most consider them to be out of date, expert systems embody the present ethos of human and machine collaboration.

Aiden co-founders Marie Outtier and Pierre-Jean Camillieri

Marie Outtier, co-founder of Aiden, is a marketer by trade. Her co-founder, Pierre-Jean Camillieri, built Aiden around Outtier’s knowledge of the industry. Because Aiden possesses baseline expectations for key performance indicators, it can proactively alert marketers when something seems out of balance.

When the company comes out of beta, it plans to target businesses with relatively high mobile marketing ad-spend — this tends to include a lot of startups. If it can straddle enough of the long-tail of marketing anomalies, this should allow companies to identify mismatches in ad-spend more quickly to more efficiently allocate limited resources.

Twitter is announcing a new ad unit — pre-roll ads on Periscope videos.

In a blog post, Periscope Group Project Manager Mike Folgner says the ads will appear in a way that’s “organic to how people already consume that content on the platform.” Basically, publishers will have the option to run a short video ad that plays before you watch their Periscope videos (both live broadcasts and replays) on Twitter.

The company has already been rolling out ways for advertisers to create custom campaigns around live Periscope video, but a standard pre-roll ad should be the most straightforward option. That means it’ll probably be used by wider range of publishers, creators and advertisers than what we’ve seen so far.

“For creators and publishers that use Periscope to go live on Twitter every day, these pre-roll ads present a new way to monetize their video content,” Folgner writes. “For marketers, they have a new source of video advertising beyond what was previously available, including ads during live streaming programs, Promoted Tweets with Periscope video, and custom Amplify campaigns with live content  from influential creators.”

Twitter says it will be making Periscope pre-rolls broadly available to advertisers over the next few months.

This is a crucial time for Twitter’s ad business. Despite all the attention around the U.S. presidential campaign (and Donald Trump’s Twitter account) ad revenue seemed to stall at the end of 2016 — an issue that CEO Jack Dorsey addressed in the most recent earnings call, when he said, “2017 will be about simplifying and differentiating our revenue products.”

Facebook is introducing a new ad format that should help retailers showcase their products.

The last big addition to Facebook’s ad lineup came last year, with the launch of Canvas, a fast-loading, rich media ad. Like Canvas, a collection is created specifically for mobile, and seems designed to win users over in the Facebook app — before directing them to the advertiser’s mobile website.

Facebook monetization director Maz Sharafi told me the goal with the collection format is “to really build a great new shopping experience for people and to help marketers really drive discovery and sales in mobile.” Collections are also designed to be easy for merchants to set up, with Facebook doing the heavy lifting of choosing the right products to showcase from the merchants’ broader catalog.

Sharafi added that collections take advantage of a number of broader trends: The growth of mobile commerce and video, plus the increasing importance of fast performance on smartphones.

So the ad shows up in users’ News Feeds — it leads with a video or image, followed by four product images below. If someone taps on the ad, it opens up a broader catalog of up to 50 different products. Then if they tap on a specific product, they’ll go to the advertiser’s website or app to make the purchase.

Sharafi said that while the advertiser can choose the initial four products that show up in a collection, the rest are automatically selected by Facebook, based on the merchant’s preferences and on user targeting, aiming to create an experience that’s “as relevant as possible for the consumer.”

Facebook is also announcing a change to the way advertisers can measure their results. For units like collections and Canvas, it won’t just tell advertisers when someone clicks an ad — it will also measure outbound clicks, namely when someone clicks out of the ad to the business’ website. This will also be available to advertisers on Instagram.

Social media can be hard.

I’m not speaking hypothetically here. There have definitely been days where I’ve thought, Shoot, it’s been way too long since I’ve posted on Instagram or Twitter — and then completely failed to post anyway because I had nothing to say. Maybe I would’ve done better if I’d used Post Intelligence.

I’ve seen plenty of other social media dashboards and assistants, but Post Intelligence’s assistant, which it calls Pi, offers the most direct advice that I’ve seen. After you connect your Twitter account or Facebook page, it creates a stream of links that you could consider posting. (Today, its first few suggestions for me were about Batman and Trump, which sounds about right.)

Once you’ve drafted a post, Pi will predict how well your post will do, and it might suggest what time to schedule it. It also offers analytics to see how a post did, and there’s even a marketplace to make money through sponsored posts.

Post Intelligence co-founder and CEO Bindu Reddy told me the product examines engagement data from Facebook and Twitter. Then it uses artificial intelligence to surface the topics that will do well with your followers, and to predict the performance of a post based on things like topic, length, sentiment and keywords.

“We’re making your tweets better,” Reddy said.

She emphasized that Pi isn’t actually telling users exactly what to say, and it’s not just focusing on whatever is currently the most popular — after all, sharing the exact same thing as everyone else isn’t necessarily a recipe for success.

“Instead of just these top five TV shows or 10 top news articles, we might highlight 1,000 different things,” Reddy said. “So I’m actually very positive that it gets rid of that echo chamber that we have.”

Post Intelligence was created by the same company behind social advertising platform MyLikes. Reddy said the technology is “completely new — we rewrote the whole infrastructure” and that the team is now focused solely on Post Intelligence.

The current version of Post Intelligence is available for free.