There is no doubt that AI is transformative. In sales and ad operations, it has the potential to reduce or even eliminate the friction that is part of the everyday process of getting a campaign from quote to cash.
However, there are some limitations. And it’s less about the technical limitations and more about the limitations imposed by human beings.
We can break out the potential benefits that AI could offer for a publisher’s direct sales and ad operations as follows:
SLAM DUNK
Media planning
AI vendors have already introduced the ability to automatically create proposals and media plans, which are informed instantaneously by historical results for that particular client, and can help predict the optimum plan moving forward.
STRETCH GOALS
Workflow
Designing and optimizing workflow, including approval processes, user roles and permissions, could all be streamlined by AI. Ingest a publisher’s current workflow and make recommendations for improvement and even configure the optimal workflow in the OMS.
Ad Product Taxonomy
Similar to the potential improvement above, AI could translate a publisher’s current taxonomy and make recommendations for improvement which could include making a product catalogue more “findable” for proposals and more efficient for inventory management and trafficking.
Inventory forecasting
I would hope that AI can do a better job of forecasting than the tools we use today. This could include a stronger link to the real time performance of content, which has a real time impact on advertising inventory. Additionally, creating forecasting that is instantaneously linked to historical inventory trends from seasonal and tentpole events would increase accuracy.
Optimization
Automating the optimization of direct sold campaigns and recommending improvements would be a boon for inventory managers. The “stretch” in this goal is the need to manually confer with an advertising client and obtaining their approval for the changes in an optimized campaign, or setting up the permission beforehand to make changes on the fly.
Salesforce.
Any AI solution that hopes to create a single point of entry to manage the quote to cash process is going to have to navigate around Salesforce. In some cases, adding yet another application between the CRM and OMS will be a hurdle. In other cases, proposing to replace SF with a new AI enable solution which would be truly end to end will run into the need for sales to cling to their current CRM.
BLOCKERS
Third party ad tags
If every ad tag received by an agency was 100% functional, this would not be an issue. However, it’s not the case. AI “could” ingest, process and traffic ad tags automatically. But because there are human operators on both sides, the troubleshooting process will continue to be manual. This is a blocker for an AI enabled trafficking process.
Billing and Invoicing
If every campaign billed on publisher delivery numbers, this would not be an issue. However, it’s not the case. Every billing cycle is an exercise in verifying the invoice reflects the ad agency’s delivery numbers. There is a highly manual verification exercise that takes place each billing cycle, to the extent that the “real” accounting is not locked down for several days. This is a blocker for an AI module that would allow invoicing at the touch of a button.
In summary, I think that AI can do sales and ad operations a world of good, but the barriers to entry will be less about the technology, and more about legacy processes, institutions, and of course, people.
DMW MediaBlog 