|
Are new Chinese AI models signaling a global AI price war?; AI tool: Using Google Gemini for effective web scraping; How Agentic AI skills are enhancing collaboration tools; Cos are readying to derive value from GenAI; and more...
China is at it again, giving US-based tech companies a run for the money they hope to earn from artificial intelligence (AI) models. On 17 March, Baidu Inc. released two new AI low-cost, multimodal reasoning models, signaling a global AI price war on the back of the launches of a host of low-cost, reasoning models this year, including Chinese AI Labs' DeepSeek, Monica's Manus, and Alibaba's Qwen.
Baidu introduced Ernie 4.5, a foundation model that builds upon its predecessor, and Ernie X1, a reasoning-focused AI designed for deep thinking. It has also made its AI-powered chatbot, Ernie Bot, free for users. Baidu confirmed that Ernie 4.5 is now accessible via the chatbot platform, while Ernie X1 is expected to be available soon.
"As a deep-thinking reasoning model with multimodal capabilities, ERNIE X1 delivers performance on par with DeepSeek R1 at only half the price. Meanwhile, ERNIE 4.5 is our latest foundation model and new-generation native multimodal model. Plus, our AI chatbot ERNIE Bot has now been made free to individual users ahead of schedule. Both models are now freely accessible to all ERNIE Bot users...", the Baidu post on X reads. Baidu added that the Ernie 4.5 model will be open-sourced on 30 June, providing developers with access to its source code and model weights.
Ernie 4.5 is a multimodal AI model capable of processing text, images, audio, and videos. Users can input these formats and receive responses to queries related to them. Baidu has emphasised improvements in the model’s language capabilities, enhancing its ability to engage in conversations, respond to reasoning-based queries, and retain memory-based context. Additionally, the model can interpret contextual content such as satire and memes.
In terms of performance, Baidu claims that Ernie 4.5 surpasses OpenAI’s GPT-4.5 across various benchmarks, while being priced at just one per cent of its competitor. These performance metrics are based on Baidu’s internal evaluations. Beyond the chatbot platform, Ernie 4.5 is also accessible via application programming interfaces (APIs) on Baidu AI Cloud’s MaaS platform, Quianfan.
The pricing structure for Ernie 4.5 has also been detailed, with input costs starting at CNY 0.004 (approximately ₹0.04) per thousand tokens and output charges set at CNY 0.016 (around ₹0.019) per thousand tokens. With Ernie 4.5 at 1% the cost of GPT 4.5, and Ernie X1 matching DeepSeek's R1 at half the price, it could likely trigger the start of a global AI price war, pressuring US tech companies to reduce rates.
In this context, you may also read: 'Why Chinese AI agentic model Manus has caused another global AI stir'. AI agents, often referred to as ‘Agentic AI’ systems, are models capable of making decisions and taking actions to achieve specific goals without human intervention, making them truly autonomous. Generative AI (GenAI) models are making AI agents smarter by helping them understand, generate, and respond to users in a more natural manner and in their own language.
How Agentic AI skills are enhancing collaboration tools
On 18 March, Zoom announced more than 45 innovations including AI enhancements for Zoom Meetings, Zoom Phone, Zoom Team Chat, Zoom Docs, and the Zoom Contact Center. Zoom AI Companion, for instance, enhances agentic skills across the entire Zoom platform, using reasoning and memory to take action and orchestrate task execution, conversational self-service, custom agent creation, and more. Zoom customer experience (CX) innovations include enhancements to Zoom Virtual Agent chat and the introduction of Virtual Agent for voice, AI intent routing capabilities, and Advanced Quality Management. Additional new agentic skills include calendar management to help schedule meetings and find a time that works for everyone, clip generation for fast clip creation, writing assistance for advanced document creation, and more.
Meanwhile, Cisco, in a bid to strengthen its agentic AI journey, has expanded its collaboration portfolio to include the general availability of Webex AI Agent to improve customer experiences. With new Cisco AI Assistant for Webex Contact Center updates and Webex AI Agent availability, Cisco hopes to transform contact centers into customer experience centers. New solutions for end users and IT teams include Webex Calling Customer Assist, new features within Webex Control Hub, and new workflow automation solutions for the Cisco AI Assistant for Webex.
Adobe, at its annual summit on 19 March, unveiled its strategy and product offerings for agentic AI. The Adobe Experience Platform (AEP) includes built-in agentic functionality that makes it easy for marketers to use purpose-built AI agents to optimize websites, handle repetitive content production tasks such as resizing, refining target audiences, creating and optimizing channel experiments and scaling content and digital media production. Adobe also unveiled Brand Concierge--a brand-centric agentic app that can leverage a company’s unique brand attributes and customer data, "creating a brand concierge experience personalized for every customer", according to Adobe.
By 2028, more than 20% of digital workplace applications will use AI-driven personalization algorithms to generate adaptive experiences for the worker, according to a new report by Gartner Inc. According to the report, digital workers satisfied with work applications are nearly three times more likely to report being much more productive compared with those who are not satisfied. Yet, a Gartner survey of 5,141 employees conducted from April through July of 2024 found that only 23% of digital workers were completely satisfied with their work applications, down from 30% in 2022.
That being said, a March survey of more than 40 customers of AI agent products by CB Insights reveals three primary pain points:
Reliability--many customers reported a gap between marketing and reality;
Integration headaches--lack of interoperability poses long-term challenges;
Lack of differentiation--horizontal markets like customer support and coding are becoming highly saturated.
AI tool of the week
How Product Marketers Can Effectively Web Scrape Using Google Gemini
by AI&Beyond, with Jaspreet Bindra and Anuj Magazine
What challenges do product marketers face when gathering market insights?
Product marketers rely on up-to-date insights to refine messaging, track competitors, and understand customer sentiment. Traditionally, this involves manually visiting multiple websites, copying data into spreadsheets, and sifting through scattered information—an exhausting and inefficient process. This manual effort is prone to errors, lacks scalability, and delays decision-making. This is where web scraping can help.
Web scraping is the automated process of extracting data from websites and structuring it in a useful format. However, traditional web scraping tools like BeautifulSoup and Scrapy often require coding knowledge or complex setups, making them inaccessible to many marketers.
Google Gemini makes this process effortless, allowing product marketers to extract structured data from any webpage—including dynamic content like scrolling reviews—without writing a single line of code.
How to access:
Access Google Gemini via https://aistudio.google.com/apikey
Google Gemini can help you:
-
Monitor Competitors: Extract pricing, features, and reviews from competitor sites.
-
Understand Customers: Collect product reviews and social media discussions for sentiment analysis.
-
Track Market Trends: Gather industry reports and real-time insights from multiple sources.
Example: Scraping Customer Feedback for Product Positioning
A product marketer at a SaaS company wants to analyse customer reviews from a competitor’s pricing page to understand pain points and positioning opportunities. Here's how they can do it with Google Gemini:
-
Open Google AI Studio and log in.
-
Go to Stream Realtime option in the left panel.
-
Select Enable Screen Sharing to allow Gemini to read visible data.
-
Navigate to the review section in the competitor’s website.
-
Use Voice Command: “Extract all visible customer reviews into a structured format.”
-
Scroll and Capture: Gemini automatically extracts data as you browse, structuring it with reviewer name, date, rating, and full review text.
-
Export and analyse: Use the cleaned data to analyse and identify gaps in the competitor’s offering.
What makes Google Gemini special?
-
No Coding Required: Unlike traditional web scrapers, Gemini works through natural language commands.
-
Handles Dynamic Pages: Scrapes scrolling content without manual intervention.
-
Structured Data Output: Provides organized, analysis-ready data for immediate insights.
Note: The tools and analysis featured in this section demonstrated clear value based on our internal testing. Our recommendations are entirely independent and not influenced by the tool creators.
Firms readying to derive value from GenAI: report
Organizations are beginning to build structures and processes to derive "meaningful value" from GenAI. Further, companies with at least $500 million in annual revenue are changing more quickly than smaller organizations, reveals a new report by McKinsey titled 'The state of AI: How organizations are rewiring to capture value'.
Key findings of the survey:
|
|
71% of respondents say their organizations regularly use GenAI in at least one business function, up from 65% in early 2024.
|
|
|
|
Organizations are most often using GenAI in marketing and sales, product and service development, service operations, and software engineering.
|
|
|
|
63% organizations say they are using GenAI to create text outputs.
|
|
|
|
More than one-third of respondents say they are generating images, and more than one-quarter are using it to create computer code.
|
|
|
|
27% say employees at their organizations review all content created by GenAI before it is used.
|
|
|
|
Respondents report increasing mitigation of inaccuracy, intellectual property infringement, and privacy risks related to use of GenAI.
|
|
|
|
13% say their organizations have hired AI compliance specialists, and 6% report hiring AI ethics specialists.
|
|
|
|
50% whose organizations use AI say their employers will need more data scientists over the next year.
|
|
|
|
Many respondents expect to undertake more AI-related reskilling in the next three years than they conducted in the past year.
|
|
|
|
28% say their CEO oversees AI governance.
|
|
|
|
17% say its overseen by their board of directors --on average, AI governance is jointly owned by two leaders.
|
You may read the full report here.
NOTABLE QUOTES OF THE WEEK
“No matter what the world does with AI, it will remain incomplete without India."
Note: Lex Fridman’s recent podcast with Prime Minister Narendra Modi is available with AI-generated multilingual dubbing by ElevenLabs. The conversation has been translated into Hindi and English while attempting to maintain the natural tone of both speakers. ElevenLabs Productions matched Prime Minister Modi’s voice in English and Lex Fridman’s voice in Hindi in a bid to ensure authenticity and clarity.
"This is the year that AI gets better than humans at programming forever. And there's no going back."
OpenAI CPO Kevin Weil while highlighting the rapid improvement of AI models in competitive coding, predicting one will reach the number 1 spot in 2025.
YOU MAY ALSO WANT TO READ:
Google to snap cybersecurity firm Wiz Inc., for $32 bn
The deal marks Google's largest all-cash acquisition to date, as the company aims to include Wiz Inc. in its Google Cloud business. Google aims to use Wiz's services to protect its entire cloud business from foreign threats and fuel rapid growth using its resources and products. The collaboration aims to boost the design, operation and automation of the security services while providing an end-to-end security platform for customers amid the rise of AI.
Google’s acquisition of Wiz adds a fast-growing revenue stream to its cloud business as it competes with Microsoft and Amazon. Wiz, expected to surpass $1 billion in annual recurring revenue this year, helps enterprises manage security across multiple cloud platforms. Beyond revenue, Wiz strengthens Google Cloud’s ability to attract large corporate clients, even those primarily using Azure or AWS. While Wiz will remain accessible to other cloud users, Google can leverage the deal to promote related services like Mandiant, its cybersecurity consulting arm, creating deeper integration within its cloud ecosystem. You can read more about the deal here.
“Is AI going to affect my job? What’s it going to do to the economy?”
At every dinner party, Aaron ‘Ronnie’ Chatterji, OpenAI's first chief economist, faces this typical question. He says it's hard to give a straight answer simply because of the widely varying predictions. “Some Nobel Prize-winning economists estimate AI’s impact on GDP in the low single digits, while others expect double-digit growth. That’s how uncertain things are," Chatterji told Mint in an interview during his recent visit to India.
The effect of AI on jobs is similarly debated. “Some studies suggest AI will affect only a small fraction of jobs, while others say up to 80% of job tasks could change. We’re still figuring out how AI will move through the economy.”
You may read the full interview here
Inside the mind of Intel's new CEO: 'Disrupt and Leapfrog'
Semiconductor veteran and former board member Lip-Bu Tan has returned to the company in a time of great need. The path ahead is rocky. Read more
Apple AI crisis? Top exec admits ‘ugly and embarrassing’ Siri delays
Online paradox: In need of a digital detox? AI assistants could be of help
Nvidia’s next act needs to be even bigger
ChatGPT can now replace Gemini as default assistant on Android
How the AI talent race is reshaping the tech job market
What the dot-com bust can tell us about today’s AI boom
AI development faces risk of entrenched bias due to underrepresentation of women: Microsoft exec
|
