OpenAI's New AI Agent Can Run Tasks Autonomously

Author

Marketer In The Loop
January 26, 2025

In partnership with

Hi, marketers in the loop,

Last week, I attended the AI Agents in B2B event in San Francisco, featuring Jeremiah Owyang (General Partner at Blitzscaling Ventures) and Akash Saraf (CEO of TheAgentic), along with 10 startup pitches. With ~70% of the audience being AI agent founders, it was packed with bold ideas and insights into the future of autonomous AI.

A key takeaway? AI agents are set to transform the economy and social norms, autonomously trading, buying, selling, and evolving with minimal human input. Imagine a world where agents represent most web users, closing ads, shopping for books, groceries or concert tickets, and selling your unused items. That future? It’s already starting.

I tested OpenAI’s new agent, Operator, this weekend. While it blocked ads effortlessly, it struggled with poorly designed websites (even humans would). What can marketers do to address these challenges? I’ve shared some key insights below.

Last but not least, I’m hosting a live session on Friday to explore agents’ current capabilities and their impact. Invites are coming soon. Stay tuned.

Onward,
Rei

In this issue:

  • Autonomous agents are here

  • AI can help with competitive analysis

  • How HexClad boosted ROAS by 20%

  • And much more

Reading time: 10 minutes

MY FAVORITE FINDS THIS WEEK

Must Read

▪️ICYMI: The Future of Jobs Report 2025 (World Economic Forum)
Notable quote: "On average, workers can expect that two-fifths (39%) of their existing skill sets will be transformed or become outdated over the 2025-2030 period. Extrapolating from the predictions shared by Future of Jobs Survey respondents, on current trends over the 2025 to 2030 period job creation and destruction due to structural labour-market transformation will amount to 22% of today’s total jobs."

Industry News

Industry Reports

Events & Webinars

AI Agents

Practical Tips

Thought Leadership

Drowning In Support Tickets? Maven AGI is here to help.

Maven AGI platform simplifies customer service by unifying systems, improving with every interaction, and automating up to 93% of responses. Seamlessly integrated with 50+ tools like Salesforce, Freshdesk, and Zendesk, Maven can deploy AI agents across multiple channels—text, email, web, voice, and apps—within days. Companies like Tripadvisor, ClickUp, and Rho slash response times by 60%, ensuring quicker support and exceptional customer satisfaction. Don’t let support tickets slow you down

DEEP DIVE

OpenAI's New Agent: A Glimpse into the Future of Work

After much anticipation, OpenAI has launched Operator, an AI agent built to handle repetitive tasks.

Picture an assistant capable of booking your flights, filling out forms, purchasing gifts, and even navigating complex web workflows - all while you supervise or step in as needed. This isn't a far-off idea. Operator offers this functionality today, leveraging its own browser to mimic human-like actions in real time.

In addition to Operator’s announcement, Perplexity also launched their agent this week: Perplexity’s Assistant for Android devices. It’s free and allows users to set reminders, send messages, play media, or book rides through app integrations.

All major players are working on their versions of AI agents, and given the popularity of ChatGPT, I will focus on Operator today.

Why Operator Matters

Operator shifts the paradigm from simply answering questions to completing tasks. It has access to tools and can act autonomously. For example, tasks like researching flights or booking a reservation - each often requiring 15–20 minutes of clicking and navigating - can be completed simultaneously in just minutes.

Looking back at the rapid adoption and impact of ChatGPT since its launch two years ago, it’s clear how quickly AI technology evolves. As Operator develops, its potential utility will only grow, making this an exciting development for businesses and individuals alike.

What It Can Do

Here’s a quick overview of Operator’s key features:

  • Autonomous Web Navigation: Operator can browse websites, fill out forms, and complete transactions. Use cases highlighted by OpenAI include booking flights, reserving tables at restaurants, or automating repetitive online tasks.

  • Workflow Automation: Users can save workflows and run them as needed to handle recurring tasks, such as updating a sales spreadsheet or checking inventory.

  • Multi-Workflow Management: Several workflows can be run simultaneously, streamlining time-intensive processes (at least until you reach your daily limit, which happened with me on Saturday ☺️).

  • Real-Time Oversight: Operator’s actions can be monitored in real time, allowing users to step in when necessary. This provides a layer of control, particularly for sensitive or complex tasks.

  • Session Sharing: Every task generates a session video, enabling users to review or share workflows with colleagues, supporting collaboration and transparency.

Operator is truly autonomous, but prompting still matters. Based on early results, more comprehensive prompts tend to yield better outcomes.

Challenges to Overcome

Despite its potential, Operator isn’t perfect. During testing, the agent occasionally misclicked, struggled with unfamiliar websites, or deleted spreadsheet content, requiring human intervention to correct errors. It also avoids handling sensitive data, like payment details, which means users must step in for certain actions.

These limitations highlight that Operator isn’t seamless yet. It remains in research preview mode, offering a window into the possibilities of task automation rather than a fully polished solution.

Why Should We Care

The arrival of Operator and similar tools signals a shift in how businesses and marketers will work. Here are three reasons to pay attention:

  • Workforce Transformation: Autonomous agents like Operator are set to transform work dynamics by automating repetitive tasks. This shift may create new roles, such as AI workflow designers, to optimize collaboration between humans and AI. If you haven’t yet, read the World Economic Forum report on the future of work. 🤔

  • Marketing Automation: Operator demonstrates the potential to streamline workflows like lead qualification, competitor analysis, campaign management and others without requiring specialized tools - I’ve been hearing a lot that agents will kill SaaS, and I suspect that’s true. Starting to experiment now can help teams prepare for wider adoption.

  • Future-Proofing Digital Strategies: Marketers must adapt their digital presence for AI agents. Data, AI-friendly content, human-generated social proof, and seamless user experiences will be key. Is your website ready for AI agents to consume content and use? What about ads being integrated or skipped by these tools? What happens when most web users are agents, not humans? These questions will shape the future of digital marketing.

A Small Experiment - and a Resource for You

To explore Operator’s potential, I tasked it with compiling a list of 20 reports on AI agents for business and marketing leaders. After some back and forth and adjustments, it generated insights from reputable sources like McKinsey, Deloitte, Accenture, BCG, and others. This list will be updated weekly and is available for anyone interested.

Final Thoughts

Operator may not yet be perfect, but it represents a significant step forward in AI’s ability to handle real-world tasks. Its current limitations reflect the early stages of a technology that could save countless hours, redefine workflows, and transform how we work.

Whether you’re curious or cautious, Operator is a glimpse into the not-so-distant future. Businesses that start exploring the potential of AI agents now will be better prepared to thrive in the age of automation.

YOU ASK, I ANSWER

How Marketers Can Use AI for Competitor Analysis

Over the holidays, I worked with a brand to run some competitor analysis to inform their 2025 plans. One key focus was web presence, where we leveraged tools like ChatGPT 4o, Semrush, and SimilarWeb to provide a comprehensive report in hours and save us days of manual work. For more in-depth design analysis, you can experiment with ChatGPT 4.0 or 4.1 and Gemini Advanced by uploading screenshots, or use Google AI Studio to navigate sites, discuss findings, and gather real-time feedback on the user experience.

AI tools will save you time and help to analyze competitor initiatives, identify best practices, and create actionable recommendations to improve your own.

Here’s a streamlined approach for websites analysis:

  1. Goal Setting and Benchmarking: Define objectives and identify KPIs like engagement or search rankings. Use tools like Similarweb to compare competitors’ traffic and performance metrics.

  2. User Experience and Design Audit: Assess site navigation, mobile responsiveness, and design aesthetics. If prompted correctly ChatGPT can provide a comprehensive report, and specialized AI tools like Crayon can help monitor broader competitors’ updates.

  3. Content Analysis: Besides ChatGPT, evaluate content relevance and quality with platforms like BuzzSumo to identify top-performing topics and gaps in your strategy.

  4. Technical SEO and Performance Evaluation: Analyze competitors’ technical SEO and performance using tools like Ahrefs or SEMrush.

  5. Reporting and Recommendations: Summarize findings into an actionable report with visual insights. Use ChatGPT to help you craft the narrative and slides.

Ready-to-use prompt that you can customize for your needs:

Objective: Evaluate and compare multiple websites to identify best practices, strengths, and areas for improvement. The goal is to inform the update or creation of a similar website that delivers an exceptional user experience for [insert target audience].

Websites to Compare:

  • Website 1: [Insert Website URL]

  • Website 2: [Insert Website URL]

  • Website 3: [Insert Website URL]

  • Website 4: [Insert Website URL]

Evaluation Criteria:

1. Core Content and Functional Pages

Placeholder: Replace with the specific pages or sections relevant to your industry/project.
Examples:

  • Test Directory or Product/Service Catalog

  • Resource Library or Knowledge Base

  • Industry-Specific Tools or Calculators

  • FAQ or Support Pages

For each page:

  • How well is the content structured and organized?

  • Is the information comprehensive and clear for the target audience?

  • Are there supporting visuals or interactive elements (e.g., diagrams, videos)?

2. Navigation and User Experience (UX)

  • How intuitive is the website’s navigation?

  • Are menus well-organized, with clear categories and subcategories?

  • Does the website provide quick links or search functionality for commonly accessed sections?

  • How consistent is the branding and design across pages?

3. Personalization and Target Audience Segmentation

  • Does the website offer tailored experiences for different user groups (e.g., [Insert Specific Roles or Expertise])?

  • Are there features that dynamically adjust recommendations based on user preferences or behavior?

  • How does the website address the specific needs of its target audience?

4. Accessibility and Compliance

  • Does the website adhere to [Insert Relevant Standards (e.g., WCAG 2.1, GDPR)]?

  • Are accessibility features implemented (e.g., alt text, high contrast, screen reader compatibility)?

  • Is the website responsive and usable across devices (desktop, tablet, mobile)?

5. Design and Aesthetics

  • Is the overall design modern, professional, and visually appealing?

  • Are the layout and typography user-friendly?

  • Does the website maintain consistent branding (e.g., logos, colors, style)?

6. Updates, News, or Notifications

Placeholder: Adjust this section to focus on industry-specific updates.
Examples:

  • How are recent updates (e.g., new tests, product launches) communicated?

  • Is there a dedicated section for announcements, or are updates embedded within relevant pages?

  • Is the update format user-friendly (e.g., filters, search, or downloadable summaries)?

7. Engagement and Interactivity

  • Does the website include interactive tools (e.g., calculators, live chat, or dynamic filters)?

  • Are there resources that encourage user engagement, such as webinars, white papers, or videos?

  • How effective are the contact or support features?

8. AI Optimization

  • Voice Search Readiness:

    • Is the content written in a conversational, question-and-answer format to match voice search queries?

    • Does it target long-tail, intent-driven keywords (e.g., "How do I use [product capability]?")?

    • Are local and "near me" phrases included where relevant?

  • Intent-Based Content:

    • Does the content address user intent clearly, focusing on informational (e.g., FAQs), navigational (finding resources), and transactional (e.g., bookings, purchases) needs?

    • Are headings and subheadings optimized to align with likely user search phrases?

  • Technical AI Optimization:

    • Is structured data (schema.org) used for rich results like featured snippets?

    • Are page speed, mobile optimization, and clean URL structures prioritized for better rankings?

9. Localization and Regional Customization

  • Does the website support multiple languages or regions?

  • Are there region-specific resources, guidelines, or contact options?

Output:

For each website:

  1. Provide a summary of strengths and weaknesses for each evaluation category.

  2. Identify unique features or best practices that can inspire the new website design.

  3. Suggest specific recommendations for improving the [Insert Target Website/Project Name].

CASE STUDY

How HexClad Enhanced Marketing Performance with Predictive Analytics

HexClad, known for its innovative cookware, faced a familiar challenge in today’s e-commerce world: achieving scalable growth while maintaining efficient ad spend. To address this, they implemented marketing mix modeling and predictive technology to supercharge their marketing efforts. This case study highlights how HexClad boosted revenue while cutting costs, offering a roadmap for other brands looking to scale efficiently.

The Challenge: Balancing Growth and Efficiency

HexClad experienced significant growth in recent years, driven by its unique hybrid cookware design. However, like many brands scaling rapidly, they encountered diminishing returns on ad spend and inefficiencies in campaign management. Their team needed a solution to optimize ad investments, predict revenue outcomes, and reduce wasted spend - all while making data-driven decisions faster.

The Solution: Predictive Analytics in Action

HexClad utilized an AI-driven analytics solution called Prescient AI to optimize marketing spend. Unlike traditional tools that rely on historical data, the platform’s forward-looking models predicted the revenue impact of marketing decisions in real time. This allowed HexClad’s team to:

  • Optimize Marketing Spend: Prescient’s platform helped HexClad build a more cost-efficient marketing mix by analyzing the potential impact of different budget allocations and channels.

  • Long-Term Projections: By forecasting ROAS several months out, HexClad made informed decisions that aligned with their growth goals.

  • Real-Time Testing: The team performed on-demand gut checks of ad spend, allowing them to adjust campaigns dynamically for better performance.

The Results: Revenue Growth and Cost Efficiency

In just a few months, HexClad saw impressive results:

  • Improved ROAS: Their return on ad spend (ROAS) improved by 20%, proving the efficiency of predictive decision-making.

  • Cost efficiencies: HexClad achieved a 40% more cost-efficient marketing mix, delivering better results and proving the value of predictive decision-making.

  • Time Savings: Automated insights eliminated over 20 hours per month of manual analysis, allowing the team to focus on higher-value strategic tasks.

These results demonstrate how marketing mix modeling and predictive analytics can help brands scale effectively without sacrificing efficiency - a critical advantage in today’s competitive e-commerce landscape.

Lessons for Marketers

HexClad’s success offers valuable lessons for marketers:

  1. Invest in Marketing Mix Modeling and Predictive Analytics: Forward-looking tools enable marketers to anticipate outcomes, not just react to past performance.

  2. Focus on ROI: Prioritize campaigns that drive measurable business results, and don’t hesitate to cut underperformers.

  3. Leverage Automation: Save time by using tools that handle complex data analysis, allowing teams to focus on creative and strategic tasks.

HexClad’s journey highlights the potential of AI-powered analytics solutions in modern marketing. By optimizing ad spend and leveraging data-driven insights, they achieved scalable growth while maintaining efficiency. For marketers looking to replicate this success, the key takeaway is clear: predictive analytics is not just a tool for analysis - it’s a strategic advantage in building high-performing campaigns.

ICYMI

How was today's newsletter?

Login or Subscribe to participate in polls.

Did you like this email? Share this link with a friend for a virtual high five!