Selling AI will be harder than selling most services. The hype balloon ruined it for many.
The CFO and CEO thought that they could install AI and let 1000 employees go. In fact, many CEOs are using AI as cover to layoff.
But selling AI for channel partners will be challenging. One is the lack of product knowledge, the other is the missing technical knowledge. But let’s assume a partner is going to charge into the void anyway to sell AI. This is the anatomy of that sale.
First, the partner will only be having conversations with Buyers/Prospects who self-select in, In other words, if they are already interested in an AI project and are looking for a little guidance from a Trusted Advisor. Because they self-selected, this will cut the sales time in half.
Now, the partner has to ask about business goals. What are they trying to accomplish with AI? What have they heard is working? Where do they think they can get the best benefit from applying AI?
This is when use cases for that specific vertical will be required. What are other {banks, dentists, manufacturers} doing with AI that is working? Well, Mr. Buyer, we have these 3 use cases from 3 vendors for you to examine.
After the Buyer agrees that a use case will work for them, now it is time to layout the POC (Proof of Concept) or trial. In detail, the partner will try to gather (1) What is the goal of this POC?; (2) What 2 factors will define success of the trial?; and (3) Bring in other stakeholders in order to flush out the Scope of Work (SOW). This might be when the TSB Sales Engineer is roped into the deal.
Next, the partner (or vendor team) will need to describe the responsibilities of both parties, in writing, for the SOW. This is important and not something most partners have ever done. There will be software and data requirements. There may be hardware requirements. What cloud platform are we using? What LLM? What database? Are there open APIs?
This is where the data issue rears its ugly head. Now the customer finds out they need a data project as well. Most data is in a silo without an open API. The data project consists of either scraping the data from an unfriendly software/SaaS app or replacing that SaaS/software with a friendlier alternative.
Now it is at least 3 weeks while we discuss the POC, and the data with vendor. This will be the proposal stage. In many cases, the professional services charges will break the budget and that deal will be put on hold. It isn’t unusual for the NRC to be north of $40K. Companies pay $30K for PS for CCaaS implementations. This will require more man hours to design and connect the systems and extract data, in order to feed the AI.
After the proposal is accepted, there will be 2-3 weeks for legal to redline. It will take even longer if your legal team uses ChatGPT as a paralegal.
Finally, we get to the stage where we go over the details of the SOW, the POC and the contract. One more week of dithering before it is signed (hopefully, it could be longer if the NRC is 6-figures). We had one drag out for 2 months after the “agreement” between the legal departments of the vendor and Buyer because the owners were jerks from the jump.
This is at least 16 weeks – 4 months – with a prospect that self-selects. This is a fast sale in AI.
This is like after the pandemic when everyone was re-evaluating the UC&C they paid for during the haste of the pandemic.
I know everyone wants things to happen fast, but the sales cycle is at least 4 months long — and that is for the POC!
Then after the project, there will be more negotiation over whether it was successful and if they want to move forward and how.