B2B Online Chicago 2026: Four Sessions That Should Move Your Q3 Roadmap
I Sat Through Three Days of B2B Online Chicago. Here's What Actually Matters and justified the Trip
Average time for a Schneider Electric employee to get a satisfactory output from their AI tools: 117 minutes. Average prompts to get there: 43.
That number didn’t come from a pessimist. It came from the team running one of the most successful enterprise AI deployments in B2B.
I just got back from B2B Online Chicago. Four sessions stood out, and together they tell a coherent story about where B2B commerce actually is in May 2026.
The data: 40% of B2B revenue is now digital
Shopify opened with fresh survey data, manufacturers and distributors between $500M and $10B in revenue, fielded April 2026.
40% of revenue now flows through digital channels. Expected to grow 15% per year through 2030.
90% of buyers said they’d switch suppliers for a better experience. 70% are ready to place orders over $50K online.
Here’s the part that surprised me: 31% are still running custom or proprietary commerce platforms. In B2C, you’d never see that number. A chunk of the room is still on something built before Shopify B2B was even credible.
The barrier to AI adoption respondents flagged? Not budget. Not skills. Integration complexity with legacy stacks.
Which is exactly what you’d expect.
Stop building monoliths. Build bridges.
A workshop later in the week reframed how I think about AI strategy.
The argument was simple. AI models are changing every two weeks. The big four (Amazon, Google, Meta, Microsoft) are spending $700 billion on AI capex in 2026 alone. Trillions next year. When the underlying technology shifts that fast, deep platform commitment is a losing strategy.
The playbook of “pick a vendor, integrate deeply, ride it for 10 years” is dead. AI companies are becoming commoditized. You’re not going to brag about which one you use.
What replaces it: short, high-value bridges. Three steps.
Intent is the business outcome. Not “we want AI.” A specific revenue or margin number your CEO would fund.
Truth is where the data actually lives. Your PIM. Your ERP. Someone’s spreadsheet. A salesperson’s head.
Experience is what your customer or employee will do differently once you connect the two.
The exercise: pick one bridge for your business and the two or three people you’d need to lock in a room to build it. Not a 24-month roadmap. One bridge.
This is the right altitude for 2026. Long roadmaps tied to specific platforms will be obsolete before you ship them. Short bridges tied to high-value use cases will compound.
The Schneider Electric reality check
Then came the Wednesday morning Schneider Electric session, and this is the one I keep thinking about.
Schneider is a $45B energy tech company, 150,000 employees. They reframed their entire digital function around what they call DXO: Digital Experience Optimization. SEO, GEO, SEM, content, and PDP optimization stopped being separate teams chasing separate KPIs. They became one function with one job. Make sure the right buyer finds the right product, regardless of channel.
That alone is worth stealing. Most B2B companies still have an SEO team that doesn’t talk to the AI/GEO team that doesn’t talk to the paid search team that doesn’t talk to the PIM team. Each one optimizes its own slice. None of them owns the buyer’s actual journey.
Schneider started their AI transformation in early 2023. Evaluated 40+ platforms. Reported 80% reduction in content production timelines and 30% cost savings.
So far, so impressive.
Then the speaker shared the numbers nobody puts in the keynote.
Average prompts to get final long-form content: 43. Average time to satisfactory output: 117 minutes. Logins per month during pilot: 11.
100% of pilot users said they’d recommend it to colleagues. Only 55% said they were ready to use it themselves.
Read that twice.
This is the gap between AI demos and AI reality. The tools work. People believe in them. They just don’t reach for them.
What Schneider learned: tracking ROI was counterproductive. It created suspicion and ate hours. The real win wasn’t efficiency. It was that employees got time back for curiosity.
The vendor demo is not the deployment. Your team is going to need 117 minutes the first time too. Plan for that, or you’ll declare failure at week six.
A side note on DXO, and a tool to test where you stand
The Schneider session validated something I built two weeks ago and have been improving daily. A free GEO audit tool.
If your SEO, GEO, SEM, and PDP teams are still siloed, you don’t actually know where buyers are finding you anymore. You know where one channel is performing. You don’t know whether you’re in the consideration set when a procurement specialist asks ChatGPT for “industrial pump suppliers under $50K.”
The tool runs your domain through Claude, ChatGPT, and Gemini using buyer-realistic prompts and shows you the verbatim transcripts. Not a score. Not a dashboard. The actual answers AI gives buyers when they ask about the kind of products you sell.
If you’re cited, you’ll see it. If your competitor is cited and you’re not, you’ll see that too. If the AI is hallucinating about your own catalog, you’ll see that, and that one’s usually the most uncomfortable.
Run your audit at humanafterall.ca/geo-audit
Five minutes to run. Longer to digest.
The proof point: SAP’s agentic demo
SAP walked us through a concrete example of AI doing real work in B2B.
A procurement specialist needs to outfit the Shanghai office. She used the same vendor for Beijing last year. She doesn’t dig through a year of orders. She asks her assistant to pull the Beijing order. The agent finds it, lets her tweak the parameters, drops everything into the cart.
Cart comes in over budget. She requests 12% off in a note.
The quote hits her sales rep with full context. Silver tier. $102K spent last year. The system tells him 10% is his margin ceiling, then flags an active promo waiving install and delivery. Combined, the deal lands inside margin and inside her budget. Auto-approved.
Most AI demos in B2B have been theater. A chatbot summarizing a PDF. A copilot drafting an email.
This was different. Real task. Real margin math. Real approval workflow. The AI didn’t replace anyone — it removed the parts nobody wants.
If you’re running a B2B commerce P&L, this is the proof point you bring to your CEO.
But here’s what the demo doesn’t show on stage. Every layer of that workflow depends on data clean enough for an agent to use. The customer’s tier. Last year’s order with full line-item history. The active promo. The margin floor. The approval rule. If any of that lives in a spreadsheet, in someone’s head, or in a PDF below the fold of your product page, the agent can’t do its job.
What to do Monday
Three things, in this order.
One. Audit your product data before anything else. This was the single most consistent message across every session I attended. Clean, structured product data is the prerequisite, not a parallel workstream. The GPT bot doesn’t execute JavaScript. Anything below the fold on your PDP is invisible to LLMs. If your specs, part numbers, and pricing aren’t in clean structured fields, you’re not in the consideration set.
Two. Pick one bridge. Don’t write the 24-month roadmap. Write the one Intent → Truth → Experience your CEO would fund this quarter. Identify the two or three people you need in a room. Build it. Ship it. Then write the roadmap from a position of proof.
Three. Plan for the adoption gap. If Schneider needed two years and a culture program to get past 55% adoption, your $200M distributor is not going to do it in a quarter. Build the rollout assuming people will need permission to spend 117 minutes on their first useful output.
The companies that will win the next two years aren’t the ones with the most ambitious AI strategy. They’re the ones with the cleanest data, the shortest bridges, and the most patience for the human side of adoption.
Chicago made that clearer than any conference I’ve been to in years.
I gave my own session on Day 1, Beyond the Portal: Why GEO Is the New Frontline of B2B Commerce. I’ll write that one up next week. Subscribe and you’ll get it in your inbox.
A big thank you to the B2B Online team for the welcome and the production. And to every speaker who showed up with real numbers, real receipts, and real opinions. You’re the reason this conference keeps raising the bar. The depth of the conversations on stage and in the hallways was the best I’ve seen at a B2B event in years.
I write deeper breakdowns of B2B commerce, PIM, and GEO strategy on my Substack at b2becommerce.substack.com.
Human After All is a B2B commerce agency. Montreal | Miami. We help distributors and manufacturers replatform to Shopify Plus, deploy Akeneo PIM, and get their product data ready for AI. Need support on a project, evaluating a replatform, or just want to talk through what a modern PIM actually looks like? humanafterall.ca