December 2025
Written By Adabro
🚀 December 2025: From Production to Orchestration
Ever feel like marketing, sales, and digital get hit with a major tech announcement every month?
First, let's talk about AI video. Kling's new technology gives you unprecedented control—edit videos through simple text prompts, swap backgrounds, change lighting, reframe scenes from different angles. This is the Nano Banana moment for video—just like Google's new image model gave us precise control over photos with text, now we can do the same with video. What once required expensive crews and weeks of production now happens in minutes.
Disney just invested a billion dollars in OpenAI's video platform. That's not a test—it's a signal that AI-generated video is going mainstream.
But here's what matters more: how you orchestrate these new tools.
The buyer funnel is getting out of control. Buyers create their own journeys across channels you don't manage—self-educating, jumping between platforms, often skipping sales entirely. Marketing teams are drowning in complexity, multiplying touchpoints without improving results.
Orchestration means building dynamic, buyer-led ecosystems. Orange tests AI agents with control groups before scaling. Walmart uses unified frameworks with specific agents for each touchpoint. At Malt, one marketer combined workflow tools in unexpected ways and connected entire departments overnight.
This is the shift—use AI to automate the routine work—data crunching, basic reporting—so your team can focus on understanding customers and shaping strategy. Build workflows that respond to real buyer behavior, not predetermined sequences. Act as your organization's eyes and ears, anticipating shifts instead of just executing plans.
The bottom line? Technology replaces tasks, not leadership. The question isn't whether to adopt AI—it's which workflows truly reflect who you are, and how to scale those authentically. That's where the real competitive advantage lives.
🎥 Watch the video summary of this newsletter
-
Kling has launched cutting-edge AI video technology giving creators unprecedented control over video production. Their Kling O1 model lets users edit videos through simple text prompts—changing backgrounds, lighting, and characters with previously impossible precision. Users can swap objects, reframe scenes from different angles, and create predictive shots to extend scenes. The platform generates complete videos from just 7 reference images with multiple camera angles. Kling Avatar 2.0 produces highly realistic avatars with lifelike emotions and impressive lip-sync. Users can personalize avatars with their own images and voices, creating videos up to five minutes long—ideal for storytelling, podcasts, and marketing. The Kling 2.6 update adds enhanced audio synchronization, improved facial expressions, and granular control over sound levels, speech, and multiple characters. Users can directly modify video elements through prompts, achieving previously impossible results. My take: This is a Nano Banana moment for videos. Just as Nano Banana Pro elevated precision control for images, Kling O1 and 2.6 bring similar control to video. These tools will streamline production for AI video marketing and sales. We're entering a paradigm where creativity matters more than big budgets and large teams.
| #Marketing | Kling | 2025/12/08 -
Steve Jarret, Chief AI Officer at Orange, talks about the progress of AI at Orange and the challenges of managing agents on a large scale. Many examples come from marketing, where mistakes by agents can be very risky. To create good marketing workflows with agents, it is important to have control group processes. This means keeping a percentage of clients separate from marketing efforts for A/B testing. Without this, companies might work without guidance, making it difficult to measure effectiveness. My take: I would add that the agents need guardrails, a human-machine process to check the results of automated workflows at important steps. It's like in a regular factory, where we check a percentage of the products made at different stages of production to see the error rate and the reasons for the errors.
| #Marketing | Orange | 2025/12/20 -
Zara has introduced a Virtual Try-On feature in its app. This allows users to upload a selfie and a full-body photo to see how clothes will look on them. The processing takes about 2 minutes and gives real-looking images of the outfits. This technology might change the way people shop and is especially interesting for Gen Alpha and Gen Z. It shows how brands can use virtual try-ons effectively. My take: yes, it's not perfect today. 2 minutes to create your try-on is long. But eventually, these kinds of initiatives will become fast and push AI video in retail and beyond.
| #E-commerce #Marketing #Video | Zara | 2025/12/17 -
According to Denis Raveraco, Directeur Général at Arcane, marketing teams struggle with data strategies and need more independence. This requires tools built specifically for marketing, enabling teams to decide and act without heavy reliance on data or technical teams. My take: To scale marketing, teams need simple, connected systems that automate work while staying flexible. With AI and many channels, coordination is critical. Tools must connect easily, share data freely, and allow new workflows without license barriers.
| #Marketing | Arcane | 2025/12/02 -
An interesting example of automation in creating content on a large scale is podcast automation, which includes making thumbnails to attract potential guests. Nano Banana Pro helps to generate images and analyze data. With an orchestration tool, you can produce a video miniature based on a client's photo. My take: today, the best examples of image generation at scale come from independent initiatives.
| #AtScale #Marketing | Shubham SHARMA | 2025/12/31 -
According to Sébastien Rozanes, CDO at FDJ United, data analysis is becoming a universal skill, like Excel once was. BI and AI tools will be adopted across roles, freeing professionals from basic analysis to focus on deeper insights, coding, and higher-value decision-making. My take: Analytical thinking will increasingly merge with workflow automation. Designing effective automated processes requires understanding data, process mechanics, and how to use programming tools and AI to execute tasks intelligently.
| #Marketing #Organisation | Sébastien Rozanes | 2025/12/31 -
According to Morgane Dawant, CDO Engie, some organizations prioritize data governance before AI, while others start by proving value through AI use cases. Both approaches are valid. Data quality and AI initiatives must move forward in parallel, focusing on high-impact opportunities—waiting for one before the other only slows value creation and adoption. My take: Especially in marketing, sales, and digital, both top-down and bottom-up approaches are essential and must run simultaneously. Otherwise, data governance remains theoretical and disconnected from real business impact.
| #DataGouv #Marketing #Organisation | Morgane Dawant | 2025/12/31 -
According to Edouard Flouriot, Director of Data Analytics at Sorare, AI is expected to outperform traditional BI tools within 12–24 months, reducing the need for manual dashboarding as users access insights directly through AI—chatting with databases instead of consuming dashboards. However, this doesn’t remove the need for data skills: there will be more focus on critical analysis and verifying whether information is correct, requiring a strong understanding of data analysis tasks. In this context, data analysts—who today may feel somewhat lost—will refocus on what truly matters: insight. By deeply specializing in verticals and speaking the language of business leaders, they gain real impact and authority. Those who master this shift won’t just analyze data—they’ll become tomorrow’s functional leaders, with data expertise as a prerequisite for leadership. My take: as with all transformation, it's not one technology that replaces another; there will be generations. Critical analysis will still be done in dashboards, while other less important tasks will be handled by LLM.
| #Analytics #Marketing | Edouard Flouriot | 2025/12/31 -
Salesforce is competing with Google and Amazon in the AI ad tech world. They promise to provide tools to make planning and tracking ad campaigns easier. Salesforce's Agentforce promises smart features that can automatically pause ads that are not doing well and monitor campaigns on its own. My take: The MarkTech stack is going through major changes with priority optimization, campaign automation, and content generation. Whether it’s Salesforce, Google, or Amazon, the foundations need to be scoped and built now to prepare for the new marketing landscape.
| #Integration #Marketing | The Information | 2025/12/31 -
Retail leaders are pushing IA. Walmart has launched a unified AI framework featuring four super agents for various business touchpoints. The customer agent, Sparky, assists with reviews, recommendations, reordering, and event planning. Partner agent Marty helps suppliers and advertisers with catalog management and campaign setup. An associate agent focuses on benefits questions and workforce data insights. The AI aims for three-hour delivery to 95% of U.S. households by 2025. Amazon has introduced Lens Live for visual shopping in its app (you can scan any object in real word to find it on amazon).
| #Marketing | Retail Dive | 2025/12/31 -
Millennials and Gen Z now represent 71% of global B2B decision-makers. Based on a LinkedIn study and interpreted by Brendan Wong, LinkedIn Careers Editor and global B2B content expert, these buyers decide through trust, peer proof, and real stories—mainly on social platforms, especially LinkedIn. Corporate jargon and generic messages turn them away fast. They want clear, relevant, and human communication. Today, brands win not by sending more messages, but by the credibility, consistency, and authenticity of the people behind them. My take : the question now is how to scale authenticity with AI tools. This means reviewing your marketing, sales, and digital workflows to select those that support your true brand and make it scalable.
| #Marketing | LinkedIn | 2025/12/31 -
According to Juliette Duizabo, Head of Data at PhotoRoom, using a modern BI tool can have a wow effect. For example, with Omni and AI, it can summarize charts effectively based on what users need. With its connection to Slack, it offers weekly reports that highlight 3 successes and 3 challenges. My take: When we talk about how to scale marketing, it is about these kinds of automation. Instead of manually analyzing basic statistics, we can have automatic routines that give us regular reports. Of course, this needs to be deployed step by step, with guidelines and human supervision.
| #Marketing | PhotoRoom | 2025/12/24 -
An interesting prediction for 2026 says that marketing will focus more on creative content instead of targeting. Creative diversity will be very important to create content that fits different stages of the customer journey. A successful paid media strategy will need to prioritize content over targeting to connect well with different audiences. My take: The easy production of content with AI videos and images might make choosing who to target less of a problem. Today, organizations can easily share interesting content with all different groups. Instead of spending a lot of time on targeting, it will often be simpler to reach all customers and create special content for each of them. For example, a B2B company could share content on LinkedIn about its main products with IT leaders, fun facts about the company with everyday users, and inspiring videos about the company culture with the CEO.
| #Marketing | CEEK Marketing | 2025/12/19 -
A member of the marketing team at Malt, a leading freelance platform in France, has created a system to collect product feedback automatically from different departments. This system uses n8n and Dust to send feedback to Notion, where it is stored for easy access. Product managers receive weekly summaries to review. This project was done by one person in marketing on their own. My take: It shows how important it is today to have tech + business oriented profiles for fast solution creation. Technology allows for quick development, and with marketing ideas and simple workflows, an individual can have a big impact.
| #Marketing | Malt | 2025/12/17 -
Klarna's AI agent has replaced over 853 full-time human agents, saving the company $60 million and improving response times by 82%. It uses AI in sales to handle simple questions from customers and staff. However, analyst Kate Leggett pointed out that depending too much on AI hurt the customer experience, as people complained about receiving generic answers. Still, Klarna says customer satisfaction has stayed the same. The company brought back human agents and reported 114 million active users. My take: automation of marketing & sales is two races: speed up the process and scale up your authenticity. If you only speed up the process, clients will see that it's fake. If you are authentic, clients will see that you remain yourself, but you are much more efficient.
| #Marketing | CXDIVE | 2025/12/14 -
Google says it might allow paid ads in Gemini by 2026, even though some people say that won't happen. VP Dan Taylor says those reports are wrong and that Gemini does not have any ads right now. While AI Overviews and AI Mode show ads, updates about the future will come later. My take: Since Gemini is the core model behind Google’s AI Assistants, already reaching around 500 million users, it’s highly tempting for Google to monetize it through advertising. However, it’s unclear how users will react if recommendations are influenced by advertisers. This uncertainty explains why Google is running many tests and announcements before an official launch.
| #Marketing | EMARKETER | 2025/12/14 -
Adobe, a big software company for marketing, has agreed to buy Semrush for $1.9 billion. This decision aims to improve Adobe's AI skills and better help marketers. Also, the company is looking to acquire AI startup Synthesia for $3 billion. My take: Adobe's business is highly exposed to AI impact. Semrush specializes in media and SEO, and it recently started focusing on AI search. Synthesia works with AI video. Adobe wants to position itself in these new fields.
| #Marketing | TheInformation | 2025/12/14 -
In the generative AI era, brand visibility is increasingly shaped by communities. Reddit alone represents 40.1% of genAI citations, proving AI favors public conversation over brand-owned content. As AI blends search, social, PR, and content into one response, traditional silos break down. Marketers must collaborate across teams, align messages, and build shared workflows to remain visible and influential early in the buyer journey. My take: Before, it wasn’t noticeable if press releases, websites, and social media used different tones. Now, as more customers ask ChatGPT to answer their questions, AI generates unified responses by blending all available messages. Any lack of consistency directly affects how AI assistants represent your brand. Moreover, since AI assistants favor question-and-answer–based sources, having little or no voice in communities will reduce your visibility in AI-generated answers.
| #Marketing | Marketing Dive | 2025/12/14 -
Google has begun displaying ads in its AI Mode search engine, which appears at the bottom of search results marked as "sponsored." This integration is part of an ongoing months-long test. This move aligns with a broader trend as AI search engines like Perplexity and major players like Microsoft and Amazon explore ad placements. In contrast, ChatGPT has not announced any ad integration plans.
| #Marketing | Marketing Dive | 2025/12/14 -
A recent conversation at the Web Summit, raised concerns about uniformity in brand building and marketing design issues. 40% of digital video ads will be AI-generated by next year. Experts warned that the overuse of AI could lead to increased sameness in branding efforts. As AI accelerates production, achieving distinctiveness in marketing is becoming increasingly difficult, signaling a shift in the industry to tackle these challenges. My take: To scale authenticity, you first need to map all your marketing, sales, and digital workflows, then identify the ones that truly reflect who you are and what makes you authentic. Start to scale the critical once with AI and automation. Creating AI videos just for the sake of making AI videos isn’t authentic.
| #Marketing | EMARKETER | 2025/12/14 -
WPP, global advertising and communications conglomerate, uses AI to enable faster and cheaper ad production, marking a significant shift in the industry. Although AI is transforming advertising, strategic leadership remains crucial for success. AI commoditizes content creation but cannot capture cultural nuances or true creativity, highlighting the ongoing importance of human-led strategic decision-making. My take: History shows that technology replaces tasks, not leadership. That distinction matters more than ever. As tools automate roles and skills, leadership remains irreplaceable—because judgment, direction, and responsibility can’t be automated (for now).
| #Marketing | MarTech | 2025/12/14 -
The traditional funnel breaks down because buyers no longer follow a linear path. They self-educate, jump between channels, and often involve sales late—or not at all. To compensate, marketers multiplied campaigns and touchpoints, but this created complexity without improving results. At the same time, budgets are shrinking while growth expectations rise, exposing the inefficiency of funnel-based thinking. Linear campaigns can’t adapt to fragmented, on-demand journeys, leading to poor performance and wasted spend. With embedded AI now shaping real-time, personalized interactions, effective marketing must move from managing stages to orchestrating dynamic, buyer-led ecosystems.
| #Marketing | MarTech | 2025/12/14 -
Disney has invested $1 billion in OpenAI as part of a three-year licensing agreement around Sora. The partnership enables AI-generated videos featuring Disney characters and points toward Sora becoming a standalone social platform for creators. With the AI content market set for rapid growth over the next five years, the companies aim to reshape storytelling on Disney+, encourage fan-generated content by 2026, and set standards for responsible AI use. My take: If a major media group is investing at this scale in a leading AI video platform, it’s a clear signal that AI video will go mainstream. For marketing and sales teams, AI-generated content will soon be standard practice—not a novelty. The time to prepare is now.
| #Marketing | OpenAI | 2025/12/14 -
Pomelli, Google’s experimental tool for creating social content using only a website URL, has integrated Veo 3.1, Google’s video-generation model. This update enables the creation of animated visuals, making it easier to produce engaging assets for social media campaigns. My take: It's a simple experimental tool from Google that shows the future of marketing and sales. It has a smooth process that analyzes your brand using your website URL and generates social content ideas in images and now video. It will be interesting to see how it evolves.
| #Marketing | Google | 2025/12/13 -
Organizations are progressing toward an AI-driven future. However, unreliable data presents challenges that hinder AI's effectiveness. Data readiness is vital for marketing AI implementation, requiring curated, transparent pipelines with strong governance, human stewardship and data literacy.
| #Marketing | MarTech | 2025/12/11 -
The role of Analytics Engineers is evolving. According to Juliette Duizabo, they now spend more time writing documentation rather than only coding in SQL, the language used to manage data. New technologies like Large Language Models (LLMs) can handle many SQL tasks quickly, while automation tools such as autocomplete and copilots reduce the need to write code manually. However, strong documentation remains critical, as it clearly explains business needs and the purpose behind the data. My take: This shift towards documentation shows that communication is becoming more important in data work. It’s not just about coding anymore; understanding the needs of the business and explaining them clearly is key for success.
| #Analytics #Marketing #Organisation | PhotoRoom | 2025/12/06 -
Julien Richard, Senior Data Science Manager at Doctolib, explains that Doctolib’s Consultation Assistant saves doctors up to 50% of consultation time. The tool listens to consultations, automatically generates transcripts, and fills clinical software. Launched in Q4 last year, it’s now used by thousands of practitioners in France, improving efficiency and allowing clinicians to focus more on patient care. My take: We see this across many industries. AI assistants are increasingly supporting frontline client service by delivering real-time recommendations, enhancing conversations through live listening and instant information validation.
| #Marketing | Doctolib | 2025/12/03 -
ChatGPT has introduced a new image model. It is faster and more accurate, and it has a very easy-to-use feature that lets users edit specific parts of images. You highlight it with a magic wand and write a prompt to describe the results. My take: the model that wins the best control over AI image editing will win the market. Google Nano Banana Pro recently had a big impact, and OpenAI is working to compete. An advantage of OpenAI is allowing users to add notes and ask for changes. In this tough race, every little bit helps.
| #Leader | OpenAI | 2025/12/16 -
The new GPT 5.2 has improved memory, which means it can remember things for a longer time and have longer relevent conversation. It also makes 30% fewer mistakes. Additionally, it has better vision and can create websites. My take: we can see that after the launch from Google (Gemini 3.0 & Nano Banana Pro), OpenAI speeds up its roadmap development to keep up with Google’s latest push.
| #Leader | OpenAI | 2025/12/12 -
Google's Gemini is increasing user engagement, now surpassing OpenAI's ChatGPT in time spent on its app. This growth is fueled by Gemini's seamless integration with Google services like Gmail and Maps, enhancing productivity. However, ChatGPT still excels in creative capabilities. Despite this, Gemini's significant momentum makes it an emerging player to watch in the AI landscape.
| #Leader | Financial Times | 2025/12/07 -
OpenAI CEO Sam Altman declared a 'code red' to enhance ChatGPT amidst rising competition. Google’s Gemini 3 and France’s Mistral are gaining traction.
| #Leader | France24 | 2025/12/04 -
Carrefour has created an internal ChatGPT tool called “Carrefour ia”. It helps new users learn and understand how to use AI. Many people at Carrefour want to find ways to include AI in their work. The company has set up guidelines to help train employees so they can solve problems on their own. More experienced users can also create agents to assist their teams. My take: most big companies are going through this step. The next step is to build a workflow builder and workflow orchestrator system to create all steps for an eventual company operating system.
| #Organisation | Carrefour | 2025/12/22 -
Infrastructure as code is a way to manage technology that makes things easier and faster. It helps set up computers and servers automatically without needing to click on screens or set up physical machines. It relies on code that builds everything from scratch to the final solution. This means you can build an entire system (like a CRM) automatically, making it easier to track changes and versions. This automation allows people to make sure everything works well by testing it first. It also keeps a clear record of how the technology is set up. My take: very tech topic but in times when we talk about automation and its limits, one of the limits is infrastructure which is not code-based. Infrastracture as code we can feed the current setup of workflows to LLM models for debugging, maintaining, or improving. Therefore, the more your marketing and sales teams rely on open-source, code-based infrastructure, the easier it is to integrate advanced AI features.
| #Analytics #Organisation | TECHWORLD WITH NANA | 2025/12/03 -
Google has launched Workspace Studio, enabling users to build AI agents without coding. Example use cases include translating meeting action items into Spanish, generating meeting summaries, analyzing customer complaints, automating invoice processing, and using ready-made templates for common tasks. Users can review and approve outputs while AI agents execute workflows efficiently.
| #Orchestration | Google | 2025/12/06 -
Salesforce has launched Agentforce 360, integrating virtual agents directly into its platform, including Slack. Users can, for example, ask an agent to compare Offer A with Offer B and generate a clear explanation for a customer. The initiative automates routine tasks and effectively turns Slack into an operational front end for Salesforce. However, concerns remain around AI reliability, data readiness, and the potentially high costs of adoption. My take: Most companies operate with fragmented tech stacks (e.g., Shopify + Salesforce for B2B + SAP ERP + proprietary PIM/CRM for B2C). It’s an illusion to think that buying one new suite will magically fix this. The result? You ask an agent to compare two B2B offers—and it pulls an outdated B2C version. Frustration follows. Before integrating agents, mapping data and workflows is a prerequisite. The real question isn’t which solution to deploy, but which exact process needs to be accelerated—and how.
| #Integration #Marketing | MarTech | 2025/12/31 -
Salesforce and Microsoft face challenges justifying the costs of their AI features. Both companies depend on expensive state-of-the-art models from labs like OpenAI and Anthropic. Customer reluctance is evident, with Microsoft encountering skepticism regarding its AI's value. Salesforce finds implementation harder than anticipated. While Microsoft leverages cloud benefits, Salesforce struggles with limited diversification. The AI landscape may change, as improvements could substantiate costs or allow for more automation. My take: Salesforce and Microsoft don’t fully control AI models, as they rely on model providers like OpenAI and Anthropic. They mainly layer AI features on top of legacy systems and resell them as packaged solutions. Meanwhile, large organizations often build internal chatbots directly with OpenAI or Anthropic, gaining more flexibility and enabling very specific use cases—such as virtual clients to quickly onboard new sales representatives. This likely explains why it’s hard to sell solutions that clients are already sourcing through other channels.
| #Integration | The Information | 2025/12/31 -
Microsoft has lowered its sales goals for AI because customers are struggling to use the new products. Several departments did not meet their sales growth targets for AI last year, leading to changes in the features of AI agents. This decision shows how hard it is to help customers switch to advanced AI solutions. My take: we had a lot of signals that organizations, especially big ones, are adopting advanced AI features more slowly than planned. From my point of view, one of the main reasons is that advanced AI integration needs work from top and from bottom at the same time, which is a new approach compared to the traditional way. It requires strategy, orchestration, and operationalisation to happen together. The profits to address these challenges come from both tech and business, which is also something less common today.
| #Integration | The Information | 2025/12/04 -
OpenAI has launched ChatGPT Shopping Research, integrating e-commerce directly into its AI platform. Available to all users, this tool analyzes prices, reviews, and specifications from trusted sources to generate personalized shopping guides through interactive quizzes and real-time modifications. The platform now supports Instant Checkout for direct purchases within chat. Current integrations include Etsy, with expansion planned to Shopify stores like Ossie, Skims, Spanks, and Vuori. OpenAI has made the open-source code available for implementation. OpenAI has partnered with major retailers including Target and Walmart. Target retains full control over transactions and data, with features supporting browsing, cart building, and checkout for pickup or delivery. Walmart's integration creates dynamic "chat and buy" experiences with personalized multimedia interactions, though critics have raised concerns about AI's impact on jobs. Google has also updated its AI Mode with shoppable images, pricing information, product reviews, virtual try-ons, and price tracking, positioning itself alongside other major players in this space. My take: Agentic e-commerce is the new target vision for all major players—OpenAI, Amazon, Target, Google, and Walmart are racing to lead. This development may challenge Google's search dominance and disrupt traditional SEO and advertising ecosystems. Today we visit websites to research; tomorrow we'll launch assistants that prepare personalized recommendations for us. This will fundamentally change how we shop and reshape the competitive landscape. Retailers partnering early are positioning themselves strategically, though concerns about job security must be addressed. Those without a strong offering should monitor this closely as conversational selling continues its growth.
| #E-commerce | Retail Dive | 2025/12/31
