Sold With AI - Edition Four (Opportunities With AI)
Sales AI still has many shortcomings, but it also presents significant opportunities. This post shares a comprehensive list of 9 areas where it presents an opportunity and 4 areas where it does not.
There is only one primary reason why we adopt new (and risky) technology - because of the opportunity that it presents. The early adopters (or even fast adopters) often reap the most benefits (in spite of the inherent risk associated with being an early adopter), hence this is something that’s worth paying serious attention to.
I believe that this would be particularly true in Sales, as the ‘risk’ associated with AI adoption is on the lower side (compared to say adopting it in finance, manufacturing, or legal).
There is a lot of noise and froth being dished out when it comes to recommending what AI (especially Gen AI) is capable of. Some correct, and some not much. Sample this from BCG:
So today, I am going to bring out an insider’s perspective to recommend what actually makes the most sense, and what doesn’t. We are an early adopter of technology (as you might imagine!), and I am going to open up our playbook and thought process.
Here are the areas where I see the biggest opportunities:
Research & Preparation
Knowing more about your buyers and their needs is the lynchpin of selling. However, unlike engineers or data analysts, 'action' is what comes naturally to most sellers, 'learning' not as much.
Gen AI solves this problem to a great degree. As Prabh Sinha et all have noted in the Harvard Business Review, it is very good at analyzing unstructured, textual, audio, video data. It's easy enough to ask ChatGPT questions about a company, or feed it a company's 10-K or 10-Q and ask it to provide you relevant insights. You could also consider exploring tools like DataBook.
Call Summaries & Actions
Getting sellers to take meeting notes and add them in the CRM has always been a challenge. This has cascading effects - managers can’t review deals well, customer experience suffers when you ask them the same questions and seller effectiveness reduces.
Some of the current breed of ‘conversational intelligence’ tools have become good at taking high-quality notes, and even entering them in the CRM automatically. This feature is also becoming commoditized and is being bundled in by everyone from Zoom to Microsoft to Salesforce.
At Humantic AI, we have found Fathom to be very good and have a separate subscription even though Zoom provides it to us at no additional cost. Zoom wasn’t doing a good job. I have heard good things about Gong too.
Personalization
Personalization, or hyper-personalization is perhaps the most talked about benefit of Generative AI. And for a good reason - Gen AI lends itself very beautifully to this use case.
However, what I have seen is that most of the experts talk only about email personalization. That is not a complete view as personalization today can extend to every interaction - emails, calls, social selling, objection handling, negotiation and more.
At Humantic AI, our work is perhaps at the leading edge when it comes to personalization. Our ability to leverage ‘Personality AI’ (or ‘Customer Psychographics’ as Gartner calls it) along with other avenues for personalization brings in the ability to personalize meaningfully for the buyer and not just the company. It builds that human touch that Gartner says is all too vital in the age of AI. I talk more about it later.
Deal Analysis & Assistance
I have not come across any products that do this well yet. However, this is an emerging area that many companies seem to be working on - from Zoom to Gong to many startups. We are also building the next level of ‘Buyer Intelligence’ here - what we call ‘Dynamic Buyer Intelligence’. More about that another time.
The central theme here is the ability to understand deal progress from calls, emails, meetings, website visits, intent, technographics and more to understand how a deal is progressing, guide the seller, and help the manager review/coach.
Data Analysis & Reporting
One of Gen AI’s core capabilities is its ability to answer questions via natural conversations. This could make it far easier for first-line managers as well as leaders to ask for insights and reports, without having to indulge in tedious query writing, report generation or unnecessary dependency on RevOps.
Forecasting
Forecasting is not exactly a Gen AI feature. The very core of Gen AI (like focus on unstructured text, voice etc.) which makes it very good for personalization, research etc. doesn’t carry over very well to analysis of structured data (which is critical to long-term forecasting).
However, Gen AI does work very well at short-term, deal-level forecasting. Hence it is good for bolstering the forecasting capabilities for many sales teams. It is also a good time to remember that there is AI beyond Gen AI. And this edition, rather the whole series, is about AI in sales, and not just Generative AI in sales.
When it comes to tools, I have heard good things about Clari’s forecasting capabilities. (We do not use it ourselves.)
Coaching & Practice
One of the emerging applications of Generative AI is around virtual coaching by AI. Based on what AI can learn about a deal from all the data (see Deal Analysis and Assistance earlier), it can surface the potential doubts and questions to a seller and help them prepare, before they actually come up. It can also help sellers do mock calls and practise for various circumstances, as well as for various kinds of buyers. Practiss and Hyperbound seem to be doing some interesting work here.
Virtual Sellers
I haven’t seen a lot here yet, but this is a natural extension of Gen AI’s capabilities. Sales teams can create AI agents that can act as virtual sellers and potentially replace certain sellers who focus on more standardized scenarios in sales.
One such scenario can be autonomous prospecting. I know that Regie AI and Rightbound have some experimental offerings in this area.
Personally, I don’t believe that the current generation of AI is yet ready to sell autonomously, hence I don’t recommend it. As the leading innovator in personalization, this area is of interest to us too. We will likely come out with something experimental later in 2024. GPT5 might be necessary before this can become a reality.
Humanization
Perhaps the least talked about, most counter-intuitive and the highest impact application of AI would be what I will call ‘humanization’. AI for the first time has made it easy to understand not just companies and their needs, but also people and their wants.
That presents a unique opportunity to take selling back to its idealistic roots, but in a scalable manner. The core of humanization is an ability to understand the people, the buyers, and modifying your actions in a way that makes you and your product a more trustworthy option for them. Gartner analysts Alice Walmesley and Robert Blaisdell’s keynote at their last conference covered this topic extensively. Gartner has also written about ‘elevating the buyer’s feelings’ being a key skill of the future-ready seller.
At Humantic AI, our work in predicting buyer's behavior, personality and DISC styles is paving the way for this humanization of selling. As per Gartner’s 2023 Hype Cycle for Sales Technology, the current adoption of ‘Customer Psychographics’ stands at 2.5-5% of the market. But with more than 70%+ of Fortune 500 sellers being trained on DISC, the rate of adoption amongst this elite group is high and could reach as much as 50-60% over the next 3-5 years.
As you can see, the applications of AI in sales are many. It is no surprise that McKinsey expects sales to be the most impacted function when it comes to Generative AI, followed by marketing (see Exhibit 3 and 4 here).
But lest it might sound like that everything will be impacted, let me briefly call out some areas that I don’t expect to be impacted much. At least in the short to medium term.
List Building
While some have spoken about lead identification being a key application, I do not think that is true. Yes, AI can be very good at lead selection and prioritization (that is, picking up likely conversions from those who are already engaged and taking various actions). But the original list building requires context that AI might not have by itself.
And no, intent and technographics are not “AI”, in spite of what the vendors might claim. They are more simple data collection operations, by and large.
ICP Analysis
This is related to the point above. AI can help, but it cannot yet build ICP or finetune it for you effectively. Neither should it.
Data Automation
More and more data automation is happening, and is necessary as well as useful. But this is not an application of AI and is mostly a function of integration engineering across various tools and interfaces.
Outreach
While some vendors are experimenting with autonomous outreach, I don’t believe the technology is ready yet. ‘AI Assisted Outreach’ might be the best bet for now, something that Gong’s new engagement product or Salesloft’s Rhythm product claim to offer. (we have not yet tested these).
A lot is happening at the intersection of sales and AI and it is unlikely that I have covered everything. Quite possibly, it will be the most impacted function in the age of AI. Those who adopt and adapt, will flourish. Those who do not, will face decline. First slowly, then suddenly. We should make our choices and brace up.