When big (AI) solutions meet small problems
As companies scramble to find use cases to better their businesses, they are grappling what to build and what should be built first. Unfortunately, there is a dissonance that is leading many organizations down a path to nowhere.
CMO’s, CEO’s, and CSO’s all (for the most part), recognize AI’s ability to drive better business practices. But their path to the brighter future is getting derailed by to biases. The first is the Available Hueristic bias — a tendency to prioritize things based on how easily they come to mind. The second is Self-serving Bias — prioritizing one’s own needs or perspectives over others or the greater good.
The presence of thes biases is evident in the fact that almost every organization I have worked with (~15 to date) have a backlog of AI projects that vary in scope and scale, but there is one use case that has founds its way to the top of the pile each and every time. That is an AI solution for generating and responding to Requests for Proposals (RFPs).
In the advertising world, these RFPs are issued by brands seeking marketing solutions to agencies. The brands are seeking creative platforms, campaigns, visual design, media strategies, media planning, data services, etc. They formalize this process by managing or contracting with a 3rd party company to run their RFP and pitch process. The RFP itself is usually a word document with an accompanying excel workbook. The typically RFP will contain questions about the agency — its address, number of employees, contact information, its capabilities, its clients. A lot of this is foundational information. The RFP also typcially will have a set of questions dedicated to the agencies experience in the brand or clients category and to provide cases studies or examples of such work. The last section of the RFP typically will have a battery of essay questions that give the agency an opportuity to stand out. Example questions might be “describe your agency culture” or “provide your agencies. point-of-view on the changing consumer needs” or “what is your strategy to identifying key audiences.”
Of course I am generalizing, but these are the elements of an RFP. And many organizations have a dedicated project plan to using Gen AI in responding to these RFPs. Let’s talk about how the biases affect good decision making that puts automated RFP responses at the top of the prioritization list:
Available Heuristic Bias
Organizations efforts in AI are typically driven by agency leadership. The leaders are focused on growth. They spend their time growing existing clients, but much of their attention is on new clients. They build relationships tin hopes of courting a new client. When opportunities arise, their focus goes to RFP and pitching.
Self-Serving Bias
These leaders know that they need to spend more time being corporate leaders. They want to be mentoring and helping grow the business they have, but the pressure they have to generate new revenue (i.e. new clients). If they could reduce the time and attention spent on RFPs , they could focus on more sustainable approaches to growth.
The double squeeze
You can likely already tell, but I will spell it out. Leaders are choosing to build practical applications of AI that benefit themselves, NOT the organization they serve. This is not a case of “put your own mask on first.” There is short term and long term damage that these misguided endeavors can create.
To back this up, I will create a hypothetical for you of a marketing organization of 500 people.
Total Employee: 500
C-suite: 4
Senior leadership: 20
Business development: 5 people
Pitch team: 3 + representation from Sr. Leadership and/or C-Suite
While only an example, this structure is not uncommon. So for any RFP response an agency needs to make there will be MAX 10 people who touch it, but more likely it is 1 to 2 people filling out the bulk of the document. Somewhere between 0.4% and 2% of a companies workforce spend time responding to an RFP. The typical RFP will take ~20 hours or working time to complete to complete (this does not include executive review time, copy editing, or time for the actual pitch). A good agency might get 20 RFPs per year for a grand total of 400 hours per year crafting these RFP responses. While these responses are critical for driving new revenue, they account for .04% (4. ten-thousandths) of the agencies available time.
An automated approach to responding to these RFPs will not save the agency time. But beyond the time-savings, the automation of these responses absolutely crushes an opportunity for the responding agency to put distance between itself and its competitors who are responding. The agency can use design, language, and tone. The RFP is templatized, but the agency has every opportunity in this very boring document to score points that will win them a spot in the next pitch round.
Of all of the applications of generative AI within marketing organizations, RFP responses is the biggest — and biggest waste of time. Agencies are spending time conflating AI with automation and a successful application of gen AI to RFPs will leave the RFP void of the agencies personality, culture in the written responses.
Simply put, agencies need to remove biases from their backlog. Focus on those projects that have the greatest impact to the business and client outcomes — not saving a couple hours of leadership’s time.