Adobe has launched a new AI agent platform for businesses. The goal of this tool is to make marketing work possible without people handling each task one by one. It means managing the whole process of attracting customers, leading them to a purchase, and turning them into repeat customers again inside one system. In this announcement, Adobe said it will work with major AI companies such as Amazon, Anthropic, Google, IBM, Microsoft, NVIDIA, and OpenAI. The strategy is to connect many AI models and services together so companies can use them in the way they want. It also explained that major global ad agencies have adopted Adobe's business CX tools as standard tools. Adobe said that now companies need to focus more on customers' long-term value than on clicks. This means it is more important not whether an ad was clicked a lot, but how long one customer will keep buying in the future and how often they will buy. It also presented a direction of connecting payment services too, so everything from consultation to order and payment can be handled in one flow.
원문 보기
Adobe's bold move was actually not a Photoshop story
If you only look at the article title, it is easy to read it like 'the Photoshop company added one more AI feature again'. But the real core of this announcement is not image editing. It is that Adobe wants to take hold of the whole operation of business marketing. It wants to run the whole flow on one platform: writing ad copy, dividing customers into groups, running campaigns, reading the results again, and deciding the next action.
Why is this a big story? Because companies are already using many kinds of AI now. One AI writes text, another AI makes images, and another tool analyzes customer data. The problem is that if these tools work separately, speed gets slower, the brand tone becomes shaky, and security management becomes more complex. Adobe is aiming right at that gap. Rather than becoming a 'company that makes AI well,' it wants to become a 'main command platform that connects AIs to real work.'
To say it a little boldly, in the past Adobe was software on a designer's desk, but now the place Adobe is aiming for is closer to the operating system of the whole marketing department. So this news is not just a simple new feature announcement. It is better to read it as a scene where a design company is trying to completely transform into an enterprise AI company.
Adobe no longer wants to stay only as an 'image-making tool company.'
The goal is to become a platform that links many AIs and data and manages the whole marketing flow of companies.
What is different between chatbots, automation, and AI agents?
| Comparison item | Chatbot | Traditional automation | AI agent |
|---|---|---|---|
| Main input | Answers when it gets a question | Pre-set rules and conditions | When it gets a goal, it plans the steps by itself |
| Decision method | Focused on conversation responses | Runs by fixed rules | Interprets the situation and suggests or carries out the next action |
| External tool connection | Limited | Some integrations possible | Widely connected from CRM, ads, analysis, to payments |
| Execution scope | Handling inquiries | Sending emails and running triggers | A continuous loop of segmentation, copy creation, execution, and performance adjustment |
| Good KPI fit | Response speed, resolution rate | Open rate, send automation rate | Conversion, retention rate, LTV(customer lifetime value) |
| Need for human approval | Needed depending on the situation | Needed for exception handling | High-risk decisions like brand, legal, and budget are still important |
This is how a marketing AI agent actually works
The key is not 'one answer' but 'a connected workflow.'
Step 1: Read customer data
It gathers data like web visit history, purchase history, and app usage logs, then first figures out who each customer is. Simply put, it is similar to a supermarket reading a regular customer's shopping cart habits.
Step 2: Split customers into groups
It divides segments (groups that show similar behavior) such as first-time visitors, people about to leave, and people who buy often. This step helps avoid sending the same message to everyone.
Step 3: Create messages and images
Generative AI makes draft ad copy, email subject lines, and banner images that fit each group. A chatbot usually stops after giving one or two results, but an agent tries to connect that work to the next execution step.
Step 4: Run it for each channel
It sends the content out in ways that fit channels like email, app push, ad platforms, and website banners. What companies want is not one pretty line of text, but a campaign that actually runs.
Step 5: Check results and adjust again
It checks who clicked, who bought, and who came back, then changes the next target and message. This repeating loop is what makes it feel like a real 'agent.'
How did Adobe go from a Photoshop company to a customer experience company?
It may look like a sudden change, but actually it was a path prepared for over 10 years.
1982~2005: Build a creative tool empire
Adobe originally started with printing and publishing technology, and became the creative industry standard with Photoshop, Illustrator, and PDF. In this period, Adobe was almost like 'a basic toolbox for designers.'
2009: Open the door to data with the Omniture acquisition
This acquisition mattered because it was the first time Adobe entered the world of web analytics and marketing data in a big way. A company that made images began to see how those images got responses from customers.
2012~2013: Marketing Cloud and the shift to subscriptions
The launch of Adobe Marketing Cloud and the shift to Creative Cloud subscriptions completely changed the direction. As it changed from a product sales company into a SaaS company that makes recurring revenue, it could go after long-term enterprise contracts more deeply.
2018~2019: Hold the customer journey with Marketo and Magento
Marketo was strong in marketing automation, and Magento was strong in commerce. After these two joined, Adobe changed from a company that ended with 'content creation' into one that handles the full journey of customers seeing, buying, and coming back again.
2024~2026: Complete the final puzzle with generative AI and agents
Now Adobe is not just at the stage of putting AI into simple creation tools. It is trying to place AI agents on top of Experience Cloud to connect content, data, marketing operations, and payments. So this news is closer to the final stage of a strategy than just a new feature.
What puzzle piece was each of Adobe's big acquisitions?
| Acquisition · Shift | What did it bring? | Meaning in the AI strategy now |
|---|---|---|
| Omniture (2009) | Web analytics, customer behavior data | A data foundation that gives AI the material it needs to make decisions |
| Marketing Cloud (2012) | A bundle of measurement, targeting, and marketing operations | A path that connects generated results to real campaigns |
| Creative Cloud subscription shift (2013) | Recurring revenue and a SaaS operating model | A financial structure that can support long-term enterprise contracts |
| Magento (2018) | Commerce and transaction flow | The link that closes the loop from ads to actual purchases |
| Marketo (2018) | B2B marketing automation | Stronger customer segmentation and follow-up nurturing automation |
| Attempted Figma acquisition (2022~2023) | Collaborative design and web-based workflow | It did not happen, but it shows they were aiming for a collaborative enterprise flow |
Why not push only its own AI, and instead connect models from other companies too
| Role | External models | Firefly | Adobe platform |
|---|---|---|---|
| Strengths | Style variety, specialized for certain tasks | A commercially safe base model | Permission management, workflow connection, billing, and governance |
| What companies get | Latest performance and experimentation | A basic option that lowers brand risk | Control many models in one work flow |
| Adobe's calculation | Do not lose customers even if the top model changes | Keep its own technical identity | Move the real profit point to platform contracts |
| Key message | The model itself | Basic creation engine | AI orchestration (the work of directing several AI systems) |
As Firefly usage grew, Adobe moved more openly
If you put your mouse over a dot, you can see the exact number.
Companies are adopting AI fast, but there are still not many companies using it well
If you look at just one big number, AI can seem like it is already fully established, but if you look inside, it is still a transition period.
AI fragmentation, what Adobe can reduce and what it cannot
| Issue | Where Adobe is strong | Limits that still remain |
|---|---|---|
| Brand consistency | Manage content creation and approval flow in one place | It is hard to fully unify offline channels and third-party systems too |
| Duplicate marketing operations | Connect asset management, campaign execution, and performance analysis | Company-wide ERP, customer center, and legacy systems still need separate cleanup |
| Security and access control | It can gather access permissions and audit logs inside the platform | Data silos by department and regulatory response are still company-wide governance issues |
| ROI measurement | It is easy to see one flow from content to campaign results | To measure actual sales contribution, payment, CRM, and financial data also need to be connected |
More marketers were looking at CLV
This chart does not show the same kind of trend. It shows two survey numbers related to CLV side by side. 2018~2019 is the share of marketers who said they know CLV, and the last value is the year-over-year growth rate in CLV metric use.
Old marketing KPIs and AI-era KPIs are changing like this
| Category | Old standard | Standard becoming more important now |
|---|---|---|
| Main KPIs | CTR, CPC, last click, instant ROAS | LTV, LTV/CAC, retention rate, repeat purchase rate |
| Preferred customers | People who click right away and buy right away | People who stay for a long time and buy again often |
| Goal AI learns | Short-term response optimization | Long-term profitability and relationship optimization |
| Risk | It may gather only lots of cheap clicks | If the data is not connected, calculation is difficult |
| Meaning for companies | It looks only at campaign performance | It looks at contribution to overall business growth |
AI marketing market, who has what kind of power
| Player | Strengths | Weaknesses | Relationship with Adobe |
|---|---|---|---|
| Platform companies(Google·Meta, etc) | Ad inventory and large-scale data | Integration outside their own ecosystem is limited | It is hard for Adobe to fully replace them, but it aims for the upper operating layer |
| Advertising agencies(WPP·Publicis, etc) | Brand strategy and execution experience | They are relatively weaker at turning software into products | They can standardize Adobe tools and become practical partners |
| Consulting firms(Accenture·IBM, etc.) | Integration of CRM, data, and organizational transformation | Weak at controlling everyday creative workflows | Can coexist with Adobe in large adoption projects |
| SaaS companies(Adobe·Salesforce·HubSpot, etc.) | Productized automation and recurring billing | They do not directly own platform or media data | Adobe stands out at the contact point between content and CX |
So, this news matters because AI is trying to close sales, not just do ads
In the past, digital ads mostly looked at middle scores. Numbers like how many people clicked, how many people signed up, and how much sales came right away compared to ad costs. But when consulting, recommendations, orders, and payment start to connect in one flow, the story changes. Now companies can see more clearly 'which message really led to payment' and 'who becomes a customer that buys again later.'
There is a reason payment becomes important here. The moment payment is added, marketing is no longer only about 'collecting interest.' It becomes directly connected to actually closing sales. In Korea, it is similar to delivery apps or shopping apps trying to finish ads, recommendations, and payment all on one screen. It is easy for users, and companies keep more data.
So this Adobe news is less an article about the future of Photoshop, and closer to an article showing a scene where the borders between ad agencies, consulting firms, and software companies are breaking down. From now on, the strongest companies may not just be the ones that make one AI model well, but the ones that connect customer data, content, and payment flow together into a system that actually makes money move.
Adobe's big move is closer to 'controlling the whole business marketing flow' than 'making better image generation.'
In the AI era, competition is happening not on who has the number 1 model, but on platforms that connect many models, data, and payment so people can use them in real work.
We will tell you how to live in Korea
Please give lots of love to gltr life