Generative AI will not kill the sales profession...
...At least, not anytime soon
Generative AI will have profound impacts on almost every part of our economy, especially in the b2b sector. However, to everyone breathlessly predicting a sales apocalypse, it’s not going to happen – or at least for a while.
If you were to map out every single interaction within a full sales cycle, you’d realize that it’s actually quite complex- and requires so much human risk management, that it cannot be completely overhauled by AI.
The sales cycle cannot afford AI “lossiness”
Consider the many steps in a successful sales cycle:
generating new leads
engaging new leads
qualifying leads
initial discovery call
product demo
stakeholder management and consensus building
negotiation
deployment planning
deployment & training
perpetual renewal
In order to replace sellers with AI, each step of the process must be able to substitute AI for the human without losing efficacy.
This is simply not going to be possible anytime soon. There is too much “lossiness” between each step. At the very least, within each step, AI would need human guidance to keep it on task, to nail the nuance of each interaction, make the right reads, and execute correctly.
In sales, incorrect execution can cost you your brand, turn off your customer base, lose you deals, and can ultimately kill your company.
You cannot, and should not, trust AI with your revenue goals
There are a few major reasons the sales profession is safe from full AI automation, at least for the next decade or so.
AI is not equipped to take business risk or bear the brunt of bad decisions. If a company misses their number, the board and investors will not find it acceptable to blame AI.
There are already rules against robocalling, and autonomous AI spam emails without a human in the loop could see similar regulation if it gets too out of control.
There is a heavy social element to buying and selling big ticket SaaS contracts. An enterprise sales rep can exert leverage to make a sale predictable. They do this by making give-to-get trades, deeply understanding needs and how they map to product capabilities, looping in stakeholders, project planning, and overall consultative selling. AI will not be able to exert buying pressure on a human in a way that actually works.
If a customer makes a buying decision based on hallucinations or information that turns out to be incorrect, it’s a massive problem. It’s risky to trust AI to completely take over a sales cycle. Especially for big purchases, humans feel most comfortable getting guarantees of key functionality from real people who are accountable.
There are 1,000 micro-interactions and micro-decisions throughout a sales cycle. AI can speed up some aspects, but would also botch many of these if left to its own devices.
If your competition has humans + AI, and you only have AI, which would you bet on to win?
Opportunities and non-opportunities for AI in sales
If you look at the sales cycle, AI will be more useful/autonomous closer to top of funnel, and less useful downfunnel.
AI can help with lead gen, lead enrichment, lead scoring, lead routing, and perhaps even simple qualification actions.
But AI is a bad fit to fully run discovery calls, demos, and negotiations, and would lead to very frustrating experiences for both the sales org and buyer. Discovery is based on rapport building, expectation setting, deep product knowledge, ability to ask probing questions, ability to stitch together pains with product gains, and a consultative approach. It’s far too complex for an LLM to replicate reliably on its own.
AI is much better equipped to assist the human responsible for these complex actions. I’m looking to humans assisted by AI copilots to lead better sales calls, that equip the seller with a call structure, reminders for key questions that haven’t yet been asked, and automatically drafts follow up emails, or proposes follow up steps, which the human in the loop can act upon.
AI will be able to search complex repositories for just in time product knowledge and answers to complex questions. AI will be able to help humans write personalized sequences and email campaigns much faster, but will require human approval so as not to improvise/hallucinate value props that could be inaccurate.
I’m also excited about AI powered CRMs to help remind sellers about actions that need to be taken throughout the sales cycle to close the deal, as opposed to AI “just doing it for you.”
AI has the opportunity to eliminate a lot of clicks and drudgery in the sales process, take bulk actions quickly, stitch together data more effectively, and assist the human seller, but not fully replace them. Along with compelling new products being born, every product a GTM team already uses will become supercharged with AI feature sets.
AI impact on customer success
For similar reasons, I think that AI copilots for customer success will become valuable. It’s very difficult for humans to manage massive books of business, and often times, the seeds of churn are sown well before a human is able to identify.
AI will be able to automatically detect usage and engagement gaps, and will automatically call that to the customer success rep’s attention along with suggestions for next steps. It will also be able to identify upsell signals and profiles in a similar manner.
The AI trend will also lead to products that are easier to deploy. Generative AI is already leading a time to value revolution within SaaS products. It collapses steps in the process of achieving great outcomes, and this trend will lead to products that are easier to set up.
Finally, when end users have issues with a product they’re using, instead of waiting 24 hours for a customer experience rep to get back to them, they will simply be able to talk to a built in support chatbot, which can read entire product and feature repositories (along with repositories of questions and answers from other users) to return the right answer and send a valued customer on their way.
AI is a copilot for GTM teams, not a terminator
While sales headcounts may start to gradually decrease, there are still plenty of places in the GTM cycle where the human matters. GTM is a game of accountability, risk management, and relationship building. AI can fill in some of the gaps, but it is a more powerful human augmenter than human replacement.
Sellers have enough problems to worry about. They do not need to worry that their entire profession will become extinct anytime soon. That’s my forecast and I’m sticking to it.
Thanks for reading EarlyGTM post #16!
Mike Marg, Partner, Craft Ventures
(for more thoughts on go to market, @mikemarg_ on Twitter)