Hi Friends,

I've been working on this so-called essay for the last couple of weeks. I started writing it in Mexico, and the more I wrote, the clearer my view got. It’s longer than my typical weeklies, so grab a coffee.

I’ve been thinking a lot lately about what a B2B buyer will look like in 2036.

In the last 9 months, I went from using AI to research ideas to running agentic systems for companies. The rate of change has been ridiculous. Looking at my own behavior, AI has made me a better seller, but it's also made me a much harder buyer. I get more answers than ever, and I still catch myself asking whether I believe any of them.

So where is this all heading?

This issue is my exploration of what a B2B deal looks like in 2036 and what lessons we can take away to apply today.

It's March 2036. Maria is COO of a 300-person company that supplies restaurants across Texas with food and kitchen equipment.

Her tenured operations AI agent catches something nobody on her team has flagged. Month-end close has slipped 3 days over the past year. The agent checks that against this year's goal (30% growth without adding finance headcount) and figures the slow close will force 2 new hires by fall.

It tries to fix the current software first. It rebuilds 2 approval workflows and wins back 1 day, which isn't enough. So it writes Maria a recommendation to replace the system, with requirements pulled from 14 months of support tickets, the contract, and the finance team's notes. That vendor contract runs out in 3 months.

It asks her to sign off on a budget ceiling and 3 milestones: a shortlist in 2 weeks, a live test by week 6, and a decision by week 8.

Maria reads it in 4 minutes. Before she approves, she forwards it to her CFO, who killed the last software project. Then she walks over to Priya, the controller who'll live inside the new system, and asks whether she'd trust it to close the books on time.

Priya says yes. Maria approves. The agent gets to work.

Across the country, a $50M software company gets a brief from Maria's agent. It arrives through her company's procurement network and goes only to the 6 vendors her agent picked from reviews and peer data. Nobody filled out a form or sat through a discovery call. And the 80-page RFP that ate 3 of your weekends, only for the buyer to pick the vendor they wanted all along? It's in a museum next to the fax machine.

The seller's agent answers within 30 minutes. It sticks to approved product facts and pricing rules, and shows before-and-after close times from 3 restaurant suppliers Maria's size.

It also sends a 5-minute demo built just for Maria. It uses tailored screenshots of the platform configured with her approval rules, plus a voiceover walking through her month-end close. Most demos today are a “spray and pray,” where a rep steers past every feature and hopes one of them matters.

Maria's agent asks for a reference. Calendar invites magically appear. The next day, a controller at one of those suppliers spends 20 minutes on the phone with Priya.

Maria needs 1 custom integration, which the seller's agent isn't allowed to promise. It hands the question to Bill, a solutions lead. Bill and Maria work out the trade-off on a 30-minute call and agree on a 60-day sandbox proof of concept. The agents handle the paperwork, and Maria and Bill sign.

The seller never ran an ad Maria saw or sent her a cold email. Maria's agent found the problem and went looking for a vendor. In that world, reaching a buyer costs almost nothing. I can see buyer and seller agents meeting in their own marketplace. Picture LinkedIn, except every profile is a business problem.

This got me thinking about the so-what. Even if that future sounds far-fetched, what can we take from it today? 3 things.

Sales claims will lose their value. When every seller's agent can write a polished, tailored pitch in seconds, buyers will stop trusting pitches and ask for proof instead: trials, results, and customers who'll vouch for you. Forrester's newest predictions say AI will speed up discovery and lengthen validation, which is beautifully put.

Outreach will get filtered. Email made outreach free, spam followed, and Gmail started rejecting bulk senders that fail its basic rules (authentication, easy unsubscribe, low spam rates). I expect buyer agents to do the same to seller agents, which is bad news for all the outbound machines being built right now.

Renewals will get tested every year. If Maria's agent can re-check her contract against the market every cycle, she stays only as long as the results hold up.

People still showed up in Maria's deal every time something could go wrong: the CFO, Priya, the reference call, Bill, and the signature. Today, only 9% of software buyers are comfortable letting an agent complete a purchase, even within approved guardrails. Someone still has to put their name on the decision.

Punchline: getting considered will get easier. Getting chosen will get harder.

The seller won Maria's deal because every interaction lined up. The agent's answer, the demo, the reference call, and Bill's call all told the same story. Sales, marketing, and customer success worked from the same picture of the customer.

Bessemer sees the same shift. Its September 2026 research on AI-native leadership teams found pre-sales, marketing, and customer success converging around revenue outcomes across its portfolio. It calls technical sellers the highest-leverage sales hires and highlights forward-deployed engineers, who work directly with customers and feed what they learn back to product. Bill is that person.

I call this the AI revenue system: 1 system that holds what the company knows about each customer, so every team and every agent works from the same facts. It's what we are building at Allenix.

Here are 3 things I’d encourage B2B founders and revenue leaders to prioritize now.

1. Turn your customer context into content. Buyers are asking AI who belongs on their shortlist, and they’re listening. In G2’s research this spring, 69% of software buyers said a chatbot steered them to a different vendor than they’d planned. GTM teams have noticed. In Kyle Poyar and Maja Voje’s 2026 State of B2B GTM report, AI search had the biggest jump in adoption of any channel, and 50% of the 300 companies surveyed plan to invest more in it next year. That’s more than any other channel.

If your content team sits in a corner brainstorming blog ideas, you’re behind. The best content now comes straight from what sales hears on calls, what customer success sees at renewal, and what product learns in the field, sorted by hyper-specific customer segmentation across the buyer journey. That’s how the right information shows up when a buyer asks their AI tool of choice. Build that function now.

2. Make your proof easy to check, and put it in public. Forrester predicts two-thirds of providers will lose deals because buyers can’t verify their claims. More than 60% of buyers now run a trial, and the winner was already on the buyer’s Day 1 shortlist 95% of the time. Public proof is how you get on that list.

So write your transformation stories, segmented by client type: the before and after, in your customer’s numbers, with their permission. Keep references ready to take a call, the way Priya got hers. Make your trial or pilot easy to say yes to. I know some of you hide results so a competitor can’t copy the playbook. I get it. But a competitor reading your case study costs you a lot less than a buyer’s agent finding nothing at all.

3. Give every team 1 view of the customer. At most companies I work with, what you know about a customer (prospects and existing clients alike) lives in 6 places: call recordings, email, a half-baked CRM, contracts, support tickets, and your best rep’s head.

The fancy term for fixing this is context engineering. It means deciding what your agents and your people know about each customer, where it comes from, and who gets to see it. It’s the plumbing under your revenue system, and investors are starting to bet on it. Bessemer partner Elliott Robinson called enterprise knowledge “increasingly fragmented,” and his team argued agents need the same permission-aware access to context people get.

Maria’s story might sound far-fetched. But the AI era is here, and there is no denying it’s moving really fast. What I know for a fact: B2B buyers' needs won’t change. They want out of a current state issue, with someone they trust to get them there without drama.

The winners will continue to be the companies that make buyers feel seen: the ones that understand their specific problem, tell the same story across the website, the sales team, and every AI agent, and have proof that holds up when someone checks.

Til next time,

--Ali

P.S. Hit reply and tell me which stage of your buyer’s journey has the most room to improve. 1 word is enough. I’ll tell you what I’d fix first.

Follow me on LinkedIn

Book a time to chat

About Me: I'm Ali Mamujee. I'm obsessed with B2B growth. I run Allenix, where we build AI revenue systems for B2B companies between $5M and $25M with complex sales cycles. Aleena made her 7th grade basketball team, and getting to watch my kids play live basketball has been a lifelong dream come true. They won their first game! Go Wildcats!