Agentic AI will transform most categories of enterprise SaaS software in the next 5 years. A lot of the excitement we have seen so far is centered around generating cool images, videos, deep fakes or fast “vibe coding.” While hundreds of startups are chasing these applications, a true transformation is happening in SaaS software, via the use of Agentic AI.

Most enterprises today run their core operations on traditional SaaS software, whether for sales, marketing, manufacturing, logistics, or eCommerce and so on. The challenge with SaaS software has traditionally been that it requires extensive training and knowledge of the domain to operate it. For example, having implemented Salesforce CRM in each of my 3 companies so far, each time we have required hiring expensive consultants to configure it just to make it work for us. It’s not just Salesforce CRM but Accounting software, Support software, etc.

Agentic SaaS on the other hand uses powerful Large Language Models (LLMs) to capture and learn the domain while handling all of the decision making and the sub-tasks needed to complete a task. Take approving a loan: what would traditionally have required thousands of lines of code to be written to check several external systems, such as checking credit score, verifying income, identity and so on. Each time these systems changed, such code had to be updated. Agentic SaaS can accomplish such tasks by autonomously completing each of these sub-tasks and coming back to the user with an approval or denial.

The key to making all of this work seamlessly is AI Agents which can perform tasks like connecting into other systems, generating a request, analyzing results and making a decision on the next task. The first versions of Agents “hacked” these systems by acting like humans, for example, using headless browsers to mimic user operations in these systems. Given the security and other risks of this approach, a new protocol that is already in use and evolving rapidly called MCP (Model Context Protocol) which many systems (e.g. credit check, identity check, hotel booking systems, restaurant reservations systems) will eventually implement — it is essentially a way for an Agent to securely get information (of course with the appropriate authentication) or complete a task without the requesting application or the receiving application having to write the code to do so. Such “codeless” operations means new SaaS software can be written with fewer resources. MCP is likely to be the way these Agentic SaaS platforms will scale and solve for the challenging workflows in large enterprises.

Enterprise workflows like approving a purchase or checking for compliance are significantly more complex than, say, booking a restaurant, which means domain expertise is still needed to make these Agentic SaaS systems work well for the enterprise. However, designing the workflows to follow an AI model of interaction and workflow automation will significantly reduce the cost and resources needed to deploy and maintain such software.

The world of Commerce Media, historically, has been extremely people intensive: agencies often work with brands to launch campaigns with hundreds of agency staff members. A large retailer has recently shared with me that it is taking them about 60 days to launch a single campaign due the largely manual creative production work, compliance checking, approvals and then getting it live across the many media platforms. Very often, promotional campaigns could not be launched within the promo window, leading to loss of sales or excess inventory. In addition, the work required to analyze campaign results and send reports to each advertiser and optimize campaigns to improve performance likewise has been done by many people at the various agencies and brand teams.

Jivox IQ DaVinci leverages AI to achieve its core vision: launching Commerce Media campaigns in under 5 minutes. AI facilitates the rapid integration of product or lifestyle content into pre-approved retailer templates. This content can be AI-generated or user-uploaded, with AI then used for customization, such as background removal, aspect ratio adjustments through image generation, and the creation of lifestyle images or copy. While the templates are inherently compliant, user-uploaded content undergoes an Agentic compliance check. This AI-powered system flags any compliance issues, by using an agent to retrieve and scan a retailer’s creative guidelines and also scans the creative to flag any compliance violations, eliminating the need for manual human review.

Reporting and analytics have also been automated to use an Agent that can take plain english requests and generate reports and charts without requiring a user to manually generate them. This involves an agent fetching the relevant data and then inferring the best way to visualize the data and then using a reporting engine to generate the reports.

There are several parts of launching a commerce media campaign that are complex and can benefit from the use of Agentic AI. For example, Agentic AI can be used to generate creative templates using a campaign brief, a brand’s native content, and learnings from prior commerce marketing campaigns and product content to generate reusable templates that can be used to quickly launch campaigns.

Agents are poised to revolutionize the shopping experience, which is currently a time-consuming process. Today, users often navigate through search engines, sponsored ads, and individual retailer websites, only to encounter out-of-stock items or limited color options. In contrast, agentic shopping, while still in its early stages, will allow users to discover products, refine preferences (color, style, etc.), and complete purchases with a single click, all within an LLM or a retailer’s shopping agent like Amazon Rufus. This transformative shift is expected to fundamentally change consumer shopping behavior.

It is clear that Agentic AI will transform almost every category of software, digital advertising and commerce marketing software has traditionally been very human powered and is ripe for disruption by a new generation of Agentic AI powered Commerce platforms.