The More Things Change, The More They Stay The Same - AI in B2B Sales
Setting The Stage
2023 changed the game for predictive models by focusing on chat based interactions, language processing and a consumer-facing GTM strategy.
Rather than staring at a computer and feeling lost, people all over the world suddenly had a personal assistant with whom they could engage in natural language, ask questions and generate content.
This minimally inhibited access to some of the most expensive and ‘powerful’ technology in the world created a global cultural revolution as children, adults and elders were exposed to bots that can mimic human conversation patterns & execute tasks.
To laypeople this truly felt like magic, no more keyword searches or reading articles and papers to gather intelligence. With a couple of sentences, you can mimic a Monet, write a poem for your cat or even build a monetizable application.
This sensational wave also had a major impact on something I did not foresee - board room discussions.
CEOs in 2021 or 2022 who dismissed me when I challenged them to transform their companies into data-centric organizations that provide customers with analytical and predictive intelligence all of a sudden began to have a change of heart.
AI began to be mentioned in every panel, blog post & earnings call as market fervor exploded.
However, if you spend a lot of time telling your board and investors about AI’s massive market opportunity, eventually they’ll want to see results.
As I reflected on the impact AI has had on my work the last few years, I thought it would be interesting to explore what has changed & what has stayed the same.
Automation In Sales
The only thing harder than operating a high-functioning sales organization is building one! It requires high quality data, functional processes, escalation mechanisms, competent leaders and flexibility. As such, countless businesses have been built to automate sales processes, helping maximize seller productivity and revenue.
The SaaS age gave us wonderful tools focused on things like Company & Contact Data Collection, Customer Relationship & Deal Management, Auto-Dialers, Campaign Orchestration & Proposal Generation.
All of this was tremendously powerful for sellers as you minimize the amount of ‘dead time’ where you’re waiting on other functions or processes to execute before you can serve the customer.
Instead of making 50 phone calls a day to maybe connect to 2-3 people, auto-dialers would empower you to make 200 phone calls and 5-10 connections.
Managing deals also became significantly easier as leadership, sales, engineering and marketing adopted tools that integrate with each other through a central data hub - the CRM.
This helped establish more consistent, repeatable & measurable sales cycles (to a certain degree) to manage the chaos of navigating complex deals & organizations.
AI Versus SaaS
I would argue that AI models are essentially an extension of the CRM where the system can take on automation at scale, with more personalization.
When doing a ‘Mail Merge’ for example, where you use a tool to run mass email campaigns, you can have a model tailor content that’s specific to the person or organization you’re engaging.
Instead of sending an email to 5000 organizations, where the only things you customize are {Company Name}, {First Name} & {Last Name}, you can have tailored messaging based on their website, industry or recent news stories.
This approach is interesting because it’s trying to simulate meaningful human research and personalized engagement without either of those things.
While relatively more effective than minimal customization, this can be expensive, slow and runs the risk of the models generating AI slop.
Therefore, in high stakes and strategic customer engagement, I’m still a proponent of direct human engagement & relationship-building. Executives want to engage with human beings, establish trust and buy from an organization that will help them drive relevant business outcomes.
By contrast, for more consumer-facing engagement or volume B2B businesses, AI automation for personalization is quite powerful as you can target the right customer with the right product at the right time with the right message.
Beyond outbound, one of the areas I’ve been really excited about is tailoring pre-built content for clients. I would spend hours exploring slide decks, proposals and sometimes even engaged internal, specialist resources to build content for customer meetings. We’re talking about 2-3 week turnaround for customer presentations that were customized to their needs.
In the age of AI, you can point an agent at their website, 10-K and earnings call - and have it tailor the content for you. I would still advise reading through or listening to those things yourself and establishing your own foundational customer knowledge but it’s much easier to edit a first draft than to build something from scratch.
Training is another area where AI can have an impact, particularly as Sales orgs scale back meaningful training & development programs. In my first tech sales job, I was part of a small cohort of 3 sellers who had 3 months of dedicated training across Industry, Product & Prospecting. This means we were in the classroom every day, for 8 hours a day, as the training team would bring in experts from across the company to educate us. This was extremely effective in turning even recent college grads into Business Development Reps and Sellers who could sell highly technical products to the C-suite.
I would argue that AI is not a replacement for human-led training, however it can be a powerful tool to replace less engaging learning methods. Instead of role-playing with another seller that’s new (and also barely knows what they are doing), you can role play with an AI model that has built-in parameters, needs and expectations. Grading and rubrics continue to be a challenge in this space but having new reps exposed to simulated customer conversations is a great way to get them up to speed faster. Rather than failing on real customer calls, which costs revenue & time, reps have a playground in which to learn and grow - without repeating the same canned conversation every time. Chaos from Generative models can actually be a feature than a bug in this use case, as the same role play could go in different directions depending on what questions are asked & when.
The More They Stay The Same
While we see Agents and Models provide meaningful value in Sales, the question on everyone’s mind right now is how much value & at what cost. Sure it would be great to send out 1000 highly tailored outbound emails but if the response rate is 2% & you drive $500,000 in pipeline but the token cost is $250,000 - the AI model may have eaten your entire profit margin and catapulted your Customer Acquisition Cost. Besides, for $250,000 you could hire 5 early career BDRs who may be able to generate $5,000,000 in pipeline for the same cost annual - 10X ROI.
Fundamentally, AI has not transformed sales cycles, but it has encouraged people to skip steps. FOMO in the technology space is very real, as we see global competitions kick off around data centers, energy & space.
However, if you skip steps, you may learn the hard way why those steps were there in the first place.
Comparing the stories of Anthropic and OpenAI is a really interesting exploration of this train of thought. OpenAI bet that AI was a revolutionary technology that should have minimal guardrails and be available to everyone - no matter the cost to maximize market share. However, Anthropic understood that giving a super computer to 14 year olds may not drive the billions in ROI that they need - particularly in a market where supply is significantly constrained.
Anthropic decided to focus on B2B use cases where organizations would be willing to pay millions to use their models. They were intentional in building their B2B sales channels, partnering closely with both Amazon & Google, unlocking their sales teams to help support their sales cycles.
As we look at their IPOs, it seems like Anthropic will surpass OpenAI in market cap despite OpenAI having a massive first mover advantage - likely in large part thanks to their responsible & focused GTM strategy.
The market is starting to wake up from its AI-driven delirium and a few things are becoming more clear:
Humans want to buy from humans, specially when the human knows what they’re talking about
Large, complex deals & technology implementations require the coordination of human resources across multiple organizations & business units
Trust is essential, earning it is hard but losing it is very easy
Show don’t tell, people want to see it in action
ROI matters, specially as geopolitical risk skyrockets
Conclusion
AI is a powerful tool with many applications but you must be thoughtful & responsible to be successful in the long run. While usage is free, feel free to share tools broadly to help people understand how they work and where they create meaningful value.
However, be mindful of usage, costs & security requirements. While larger organizations may be able to absorb large inefficiencies, lawsuits and cost, smaller companies often cannot. One massive AI bill could be the end of an early-stage startup or small business.
At the end of the day, things usually aren’t as great as the hype or as bad as they feel - it’s somewhere in between. As Founders and business owners, we need to understand what that in between means for our business & industry.
What are the tools that you can get for free? What are tools that you’re willing to pay for? What is your methodology for collecting internal or customer use cases, validating them & launching alpha/beta or EAP programs?
If those questions start to bounce around in your head and you feel like some structure and experience can help make sense of the chaos - shoot me an email! bruno@atlasreforged.com
