AI

As ChatGPT hype hits fever pitch, Neeva launches its generative AI search engine internationally

Comment

NeevaAI in action
Image Credits: Neeva

Challenger search engine Neeva wants to replace the familiar “10 blue links” in search results with something more fitting for the modern AI age.

Back in December, Neeva co-founder and CEO Sridhar Ramaswamy, who previously spearheaded Google’s advertising tech business, teased new “cutting edge AI” and large language models (LLMs), positioning itself against the ChatGPT hype train.

“ChatGPT cannot give you real time data or fact verification,” Ramaswamy wrote at the time. “In our upcoming upgrades, Neeva can.”

Fast-forward to January, and Neeva formally launched NeevaAI to the U.S. market, pitched as “authentic, real-time AI search.” While it was technically possible for users around the globe to access NeevaAI before now, it required a little jiggery-pokery in the account settings, involving changing language and location preferences. Today, however, NeevaAI is officially rolling out internationally to logged-in users, including Canada, the U.K., Germany, France and Spain. Additionally, the Neeva search engine itself (not NeevaAI) will be rolling out to Australia and New Zealand.

The timing of today’s announcement is particularly notable, coming a week after Microsoft reignited the search engine wars with the introduction of ChatGPT to its Bing search engine and, with it, the promise of a completely reimagined search experience.

What the ChatGPT?

Most years can pretty much be defined by at least one over-arching tech trend. In 2022, web3 was one of the big buzzwords in town, with the metaverse and tangential immersive technologies also jostling for mindshare. While there is nothing to indicate that such trends will wither any time soon, it’s clear from the first six weeks of the new year that generative AI will be the big talking point of 2023.

Generative AI, essentially, describes the process of using algorithms to create (“generate”) new content. The poster child of this movement is ChatGPT, a chatbot-style technology trained on large language models (LLMs) capable of producing scarily good (but far from flawless) work such as essays, articles, poems, lyrics and even computer programs. The handiwork of OpenAI, a Microsoft-backed artificial intelligence (AI) research body, ChatGPT has taken the world by storm since its prototype was first introduced back in November, with some arguing that its launch signals AI’s arrival into the mainstream.

With OpenAI already commercializing the service through a premium subscription, the mighty Google last week scrambled into action to unveil a new “experimental conversational AI service” called Bard, showcasing how AI might transform search engines by powering synthesized responses collated from multiple sources to provide more nuanced answers to online queries. And that is what Microsoft actually launched two days later — an all-new incarnation of Bing powered by a more advanced version of ChatGPT, customized for real-time search.

Rather than serving the usual laddered list of links, the new AI-infused search engine takes a request, scans for answers and generates a response replete with citations to the original sources.

The new AI-infused Bing search engine in action. Image Credits: Microsoft

And this, effectively, is what Neeva is now bringing to international markets, a month on from its U.S. launch.

The story so far

By way of a brief recap, Mountain View, California-based Neeva first launched a subscription-only search engine in its domestic U.S. market in June 2021, and later went on to add a free “basic” tier to the mix with certain restrictions in place. The company brought the search engine to Europe back in October, and has since rolled it out to additional markets around the world.

Neeva’s core selling point is that it doesn’t monetize through advertising, and it prevents third-party trackers from using personal data to display personalized ads — Neeva wants to make money through good old-fashioned paid subscriptions.

An ad-free search environment means that users don’t have to scroll through myriad sponsored results to get to the organic links they want. However, Neeva’s basic look-and-feel is much the same as search engines have been since before even Google arrived on the scene — row after row of links to individual sources, with a few aesthetic deviations thrown into the mix.

Neeva: Only organic search results. Image Credits: Neeva

But with NeevaAI, the startup is looking to do its part in reinvent search.

TechCrunch has been dabbling with NeevaAI over the past few weeks, and from our tests it performed pretty impressively, comfortably handling questions such as “Why were the Beatles so big?” or “What is the world 5k [running] record?,” with Neeva generating a response from multiple cited sources in real time.

These citations are key to avoiding the “black box” controversies that engulf many AI technologies. The idea here is that by showing people where it’s sourcing its information from, this not only promotes transparency, but also gives those who have published content to the web the credit they deserve and increases their chances of follow-on referral traffic.

NeevaAI in action. Image Credits: TechCrunch

But some questions it simply isn’t equipped to deal with. In those cases, it defaults to the familiar website-specific links and excerpts for users to peruse and figure out the answer for themselves.

Neeva returning normal results. Image Credits: TechCrunch

In an interview with TechCrunch for this story, Ramaswamy explained some of the rationale that went into developing NeevaAI, including how it decides when and where to generate answers from multiple sources.

“Roughly speaking, it works for questions that we can find authority websites that answer that question,” Ramaswamy explained.

So, unlike many of the quirky examples we’ve seen so far of ChatGPT being tasked with writing lyrics in the style of Nick Cave, for instance, NeevaAI won’t help you here. Likewise, if you ask it a trick question — deliberately or otherwise — it likely won’t return an answer, reverting to the usual list of links for you to research yourself.

“Our goal with NeevaAI was to — first and foremost — be fail-safe, we don’t want to be telling you incorrect things,” Ramaswamy continued. “So we opted for safety in how we retrieve (information).”

An example here might be if someone was to ask NeevaAI when Boris Johnson served as the king of the United Kingdom, rather than the more accurate question of when he served as prime minister. It’s the sort of question that could potentially trick any search engine, because there will undoubtedly be pages on the internet that contain all the words in the question — so the onus is very much on NeevaAI to ensure that it understands the question, and gives an appropriate response. Or no response at all, which is what it will do with questions where it’s not confident of the answer.

NeevaAI refusing to answer a question. Image Credits: TechCrunch

Clearly, NeevaAI’s response here isn’t perfect. A smarter response — one that a human would undoubtedly deliver — would be to tell the user that Boris Johnson wasn’t in fact the king, but he did serve as prime minister. Or at the very least, generate a more satisfactory response that steers the user to refine the question. And this is something that could come to NeevaAI in a future iteration, though what form this ultimately takes isn’t clear.

“If you ask it some some silly question, it’s just not going to say anything because there’s nothing on the web that suggests that this is true or false,” Ramaswamy said. “So when there’s not an answer to be had, we decline to answer. But we will be fixing that soon.”

State of play

TechCrunch’s conversation with Ramaswamy took place before Bing’s big reveal last week, so today’s expansion into global markets has a slightly different feel to it than it might’ve done prior to the knowledge that Microsoft was basically doing the same thing — with billions more dollars in its coffers to help. And it’s clear now that Google will be doing something similar in the future.

Elsewhere, other search engine upstarts are following a similar trajectory to that of Neeva. Back in December, venture-backed You.com launched similar ChatGPT-style AI smarts to that of Neeva.

It’s clear that Neeva has its work cut out if it’s to differentiate in a market that includes long-established billion-dollar incumbents, and other agile startups, with much the same goals. However, Neeva is hopeful that its ad-free approach will win it enough supporters to see it flourish more generally as a search engine, at a time when the world has grown increasingly weary of big tech’s big data harnessing. The likes of You.com doesn’t yet deliver ads, but it has said it may offer “non-privacy invading” ads in the future.

But more than all that, Neeva is betting on its future by developing its own search stack, rather than leaning on the same engine used by Google or Microsoft’s Bing, as most other challengers do.

None of this is to say, however, that Neeva is an entirely independent, self-sufficient entity. While it is pushing toward relying entirely on its own search stack, for now it does lean on Bing for some web search activity. And specific to NeevaAI, Ramaswamy confirmed that it works with “large language model companies” including OpenAI’s GPT-3.5 and Claude from the heavily-venture backed Anthropic, which was founded by former OpenAI employees.

“We use them for things like generating training data for us, and sometimes we will make calls to them to help with the summaries,” Ramaswamy said. “Most models we use for NeevaAI are our own — we pre-train or fine-tune them for custom tasks like question-answering or summarization.”

Founded four years ago, Neeva has taken on some $77.5 million in venture capital funding from Silicon Valley VC heavyweights, including Sequoia Capital and Greylock. So while it is up against the likes of Google and Microsoft, among other well-funded startups, it’s not exactly without friends and finances itself.

Irrespective of where Neeva goes from here, it’s clear that 2023 is shaping up to be a big year for generative AI, evidenced by the likes of Shutterstock’s recently-launched toolkit for creating stock images based on text prompts.

Similarly, Neeva has no plans to limit the scope of its AI endeavors to simple search engine-based information retrieval, and is likely to expand into related verticals in the future.

“Vivek [Neeva co-founder Vivek Raghunathan] and I are like kids in a candy shop, we don’t know what to work on next,” Ramaswamy joked. “There’s just so much — simple, natural follow-on things. Generative AI is one direction, doing site-specific search is another super-interesting direction. These are things that are easily adjacent to what we’re doing.”

NeevaAI is available today in English for logged-in users on both the free and premium plans in the U.K. and Canada, as well as local language incarnations in Germany, France, and Spain.

More TechCrunch

Google Play has a new discovery feature for apps, new ways to acquire users, updates to Play Points, and other enhancements to developer-facing tools.

Google Play preps a new full-screen app discovery feature and adds more developer tools

Soon, Android users will be able to drag and drop AI-generated images directly into their Gmail, Google Messages and other apps.

Gemini on Android becomes more capable and works with Gmail, Messages, YouTube and more

Google’s gunning for OpenAI’s Sora with Veo, an AI model that can create 1080p video clips around a minute long given a text prompt.  Unveiled on Tuesday at Google’s I/O 2024 developer…

Google gets serious about AI-generated video at Google I/O 2024

In addition to the body of the emails themselves, the feature will also be able to analyze attachments, like PDFs.

Gemini comes to Gmail to summarize, draft emails, and more

The summaries are created based on Gemini’s analysis of insights from Google Maps’ community of more than 300 million contributors.

Google is bringing Gemini capabilities to the Google Maps platform

Google says that over 100,000 developers already tried the service.

Project IDX, Google’s next-gen IDE, is now in open beta

Here are quick hits of the biggest news from the keynote as they are announced.

Google I/O 2024: Everything announced so far

The system effectively listens for “conversation patterns commonly associated with scams” in-real time. 

Google will use Gemini to detect scams during calls

The standard Gemma models were only available in 2 billion and 7 billion parameter versions, making this quite a step up.

Google announces Gemma 2, a 27B-parameter version of its open model, launching in June

This is a great example of a company using generative AI to open its software to more users.

Google TalkBack will use Gemini to describe images for blind people

Firebase Genkit is an open source framework that enables developers to quickly build AI into new and existing applications.

Google launches Firebase Genkit, a new open source framework for building AI-powered apps

This will enable developers to use the on-device model to power their own AI features.

Google is building its Gemini Nano AI model into Chrome on the desktop

Google’s Circle to Search feature will now be able to solve more complex problems across psychics and math word problems. 

Circle to Search is now a better homework helper

People can now search using a video they upload combined with a text query to get an AI overview of the answers they need.

Google experiments with using video to search, thanks to Gemini AI

A search results page based on generative AI as its ranking mechanism will have wide-reaching consequences for online publishers.

Google will soon start using GenAI to organize some search results pages

Google has built a custom Gemini model for search to combine real-time information, Google’s ranking, long context and multimodal features.

Google is adding more AI to its search results

At its Google I/O developer conference, Google on Tuesday announced the next generation of its Tensor Processing Units (TPU) AI chips.

Google’s next-gen TPUs promise a 4.7x performance boost

Google is upgrading Gemini, its AI-powered chatbot, with features aimed at making the experience more ambient and contextually useful.

Google reveals plans for upgrading AI in the real world through Gemini Live at Google I/O 2024

Veo can generate few-seconds-long 1080p video clips given a text prompt.

Google’s image-generating AI gets an upgrade

At Google I/O, Google announced upgrades to Gemini 1.5 Pro, including a bigger context window. .

Google’s generative AI can now analyze hours of video

The AI upgrade will make finding the right content more intuitive and less of a manual search process.

Google Photos introduces an AI search feature, ‘Ask Photos’

Apple released new data about anti-fraud measures related to its operation of the iOS App Store on Tuesday morning, trumpeting a claim that it stopped over $7 billion in “potentially…

Apple touts stopping $1.8BN in App Store fraud last year in latest pitch to developers

Online travel agency Expedia is testing an AI assistant that bolsters features like search, itinerary building, trip planning, and real-time travel updates.

Expedia starts testing AI-powered features for search and travel planning

Welcome to TechCrunch Fintech! This week, we look at the drama around TabaPay deciding to not buy Synapse’s assets, as well as stocks dropping for a couple of fintechs, Monzo raising…

Inside TabaPay’s drama-filled decision to abandon its plans to buy Synapse’s assets

The person who claimed to have stolen the physical addresses of 49 million Dell customers appears to have taken more data from a different Dell portal, TechCrunch has learned. The…

Threat actor scraped Dell support tickets, including customer phone numbers

If you write the words “cis” or “cisgender” on X, you might be served this full-screen message: “This post contains language that may be considered a slur by X and…

On Elon’s whim, X now treats ‘cisgender’ as a slur

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: Watch the AI reveals live

Facebook once had big ambitions to be a major player in enterprise communication and productivity, but today the social network’s parent company Meta will be closing a very significant chapter…

Meta is shutting down Workplace, its enterprise communications business

The Oversight Board has overturned Meta’s decision to take down a documentary revealing the identities of child abuse victims in Pakistan.

Meta’s Oversight Board overturns takedown decision for Pakistan child abuse documentary

Adam Selipsky is stepping down from his role as CEO of Amazon Web Services, Amazon has confirmed to TechCrunch.  In a memo shared internally by Amazon CEO Andy Jassy and…

AWS CEO Adam Selipsky steps down