AI has been good for Section.

The moment I started using ChatGPT Plus in February 2023, I knew I had to figure AI out – for myself and our business. We created our first AI course that summer, and the first cohort had 500 students. Our team built the content in-house and I taught it – we couldn’t find anyone better. (Now we have lots of great instructors who are creating AI courses with us).

Almost two years later:

• We’ve trained over 15K knowledge workers in AI, across 3 formats (workshops, sprints, and bootcamps).
• We have 18 AI courses, 8 more coming in H1, and will release our AI coach (ProfAI) in Q1 of next year.
• Every employee at Section is an AI Expert or Practitioner and we have over 50 AI-assisted workflows augmenting our work inside Section.
• We have built and deployed 10 GPTs to automate some of our most important workflows which has saved us hours of time weekly (and a lot of money).
• Our virtual AI:ROI conference in November had 14K registrations and speakers including Moderna’s Head of AI Brice Challamel, Writer.com’s CEO May Habib, Salesforce’s Marc Benioff, and more.
• We hosted over 30 other AI-related conversations, with over 60K registrations – on topics from the backlash to AI to how humans and AI will collaborate going forward.

Our goals for next year: Train 100K employees, from over 1000 companies, ranging from start-ups to Fortune 500; Host over 50 events, with over 75K registrations, all about how to get ROI from AI.

Bottomline: I want Section to be a trusted partner for leaders that want less AI hype and more business value.

Here’s the kicker, and what keeps my CFO and board happier: next year, we will serve about 7X more customers, and increase headcount by only four. And in some teams, we’ve been able to reduce contractor and vendor expenses by 10-20% – while serving more customers.

We are INSANELY more productive. And we should be. At our core, we are a content company that’s IP- and language-intensive, so generative AI was made for us.

As we head into 2025, AI has driven increased company productivity, robust customer demand, and more revenue per employee (which is at its highest since we started Section). We’re ending the year on a high.

But, since I am older than the average startup CEO, I know that we are now due for some unpleasant surprises – the start-up rocket ship usually looks more like a roller coaster. Maybe the inference cost of ProfAI per session is too high, or OpenAI and others start giving away all the training their enterprise customers need to prevent user churn, or CFO’s push back on funding new AI projects.

By the way, AI was also a gift to me. In exchange for my conviction and effort, I have earned a little boost of relevance, late in my career. I will take it. And remember: I am 63 and still type with 2 fingers – so if I can figure this out, you can.

What we’re seeing on the ground

For the last 20 years, we’ve all been living and working in a world ruled by applications that can do a few things very well and very reliably. Now we’re being asked to work with generative AI, which can do almost anything, but gets a lot of things wrong and it’s on you, the user, to figure that out.

This is the challenge and the opportunity. Add in general anxiety on AI’s impact on jobs and self worth and you’ve got a recipe for low adoption, stalled pilots, and elusive ROI.

In most organizations, that’s what we’re seeing: generally low adoption (or adoption that’s very inconsistent across teams), superficial adoption (a good start but not generating much business value), or already failed AI deployments. And it will get worse, before it gets better.

But there are bright spots:

• A few organizations have successfully seen widespread adoption that is driving real business value (like Moderna, whose head of AI told us at AI:ROI that they’ve achieved 100% AI adoption among desk workers).
  
  
• There are specific, high value use cases emerging – Blue Cross Blue Shield Michigan saves $10M by using AI for contract management, and Bell Canada partnered with Google Cloud to transform their call center and reports $20M in savings across customer operations.
  
  
• Three million developers build with the OpenAI API, so besides being overwhelmed with new AI apps (read: gold rush), there’s no doubt we will also get some brilliant new use cases that create real business value.
  
  
• I talk to people inside big companies everyday who say they can’t work without AI – so we know that early adopters get real value.

So what can we learn from 2024 – and what is needed in 2025? Change management initiatives that lower anxiety and increase baseline proficiency, advanced proficiency for language-intensive roles and functions, workflow redesigns to embed AI at the job function level, and leaders that get their own hands on AI.

Our 2025 roadmap for supporting AI operators

Next year, we’re doing more of what worked in 2024 - to lower employee anxiety and increase brand-safe AI adoption. But that’s not enough, so we are expanding our offering for the enterprise.

1. Building out our AI for business curriculum: We’re adding more courses (and in multiple languages for global customers), focused on more advanced and more role-based AI use cases. ‍
2. More custom learning experiences: We’re also continuing to offer personalized courses for whole teams based on their policies and LLM of choice – and adding an emphasis on the parts of their orgs that AI will substantially change (e.g. PR & communications, consumer insights & research, frontline sales etc.).‍
3. Change management program design and management: Starting in January, we’ll offer change management advisory services to help leaders build workforce transformation programs that support the deployment of their AI platforms and LLM’s. We have already hired three consultants (to support EMEA, and west and east coast USA) and have started our first client engagements.‍
4. Scaling AI proficiency more cost-effectively: I believe that our AI coach, ProfAI ,will be the best way to offer personalized AI training at scale. Once leaders start the change management process, they’ll need to set ambitious adoption targets – which means they need a way to train their people fast in a way that works for each individual. ProfAI will be made available to a few of our enterprise customers in Q1 and then wider release in Q2.

It feels like 2025 will be a critical year for enterprise AI – does the hype and experimentation turn into more bright spots of real business value? Do all these new heads of AI inside the enterprise get any budget or just the title?

For most of 2024, and even this month with the 12 Days of Shipmas from OpenAI, it feels like tech nerds releasing half-baked AI capabilities, tone deaf to the realities of the rest of us – we have jobs to get done, with limited resources and time. My request to Sam Altman and crew: less AI features, and more AI use cases and case studies, please.

We will do what we can to help turn these amazing AI capabilities into employee adoption and value – for the ambitious IC and the committed team or organization.