Menu
News

THE AI SALES WAR IS HERE, AND IT’S BEING FOUGHT ON SALESFORCE

Avatar photo
Damian

Microsoft turns its sales floor on Claude and GPT, Anthropic and OpenAI hire humans for the jobs AI “killed”, Salesforce ships its agent army & how Carles Reina would rebuild ElevenLabs’ GTM…

If you’re selling AI into the enterprise right now, the rules of engagement just shifted under you.

This issue isn’t theory. Microsoft just turned its sales floor into a weapon, Anthropic and OpenAI are hiring humans for the exact jobs everyone says AI killed, and the data underneath all of it keeps pointing at the same unglamorous truth: your GTM infrastructure matters more than your smartest model. Let’s get into the trenches.

Microsoft trains its army to talk down Claude and GPT

Microsoft just handed its sales army a script, and the whole script is about your product.

Bloomberg (via TechCrunch) reported Microsoft pulled its sales organization into an internal strategy session to teach reps how to position against OpenAI, Google and Anthropic for fiscal 2027. EVP Jay Parikh’s mantra for the year: “Everyone else is selling parts. We’re selling the full end-to-end system.” EVP Jacob Andreou got specific, putting Copilot next to Claude directly: slower, less accurate, lacking the security integrations inside Office. The whole script leans on two themes reps can repeat anywhere, cost and security.

The backdrop matters more than the talk track. This is the same Microsoft quietly swapping OpenAI and Anthropic models out of Word and Excel for its own, reportedly to cut costs, after April’s amendment stripped exclusivity from the OpenAI deal. What used to be a partnership is now a competitive account.

Strip away the branding and it’s the oldest incumbent play in enterprise software. “We sell the whole system, they sell parts” works because it reframes the buyer’s decision from “which is best” to “which is safe”, and safe means the vendor already wired into your stack. If you sell AI into the enterprise, that’s the wall: not a better competitor, but an incumbent already in the building. The counter isn’t a benchmark. It’s proof, fast enough that the “whole system” pitch never gets room to breathe.

Read more here!

 

Damian’s insight: Distribution is king, and Microsoft just proved it: this isn’t a bet that Copilot is smarter than Claude, it’s a bet that being already installed beats being better. Same law, bigger point, there’s no stronger candidate for strategic GTM infrastructure than the CRM, the one system your whole revenue team already lives in. So don’t fight on model quality, prove ROI fast.

Anthropic runs its own GTM on Salesforce, and builds it with Claude 🧩

This is the tell of the entire AI cycle.

Anthropic, the company behind Claude, is quietly hiring humans to stand up its revenue engine on Salesforce. A live req, “Salesforce Developer, Partnerships” ($220–270K OTE), asks someone to build the platform running Anthropic’s Partnerships business: deal registration, partner attribution, co-sell, cloud-marketplace ops across AWS, GCP and Azure. And it’s not a one-off, a search returns a dozen Salesforce roles. Real platform work (Apex, LWC, Flow), staffed like core infrastructure.

Two details make it a market signal. The team is explicitly AI-first: developers use Claude to write and ship the code that powers the org, Claude building on Salesforce. And the posting is about scaling Salesforce “in a high-growth GTM environment”, a polite way of saying the CRM is where growth gets managed, measured and paid out. The move worth watching isn’t whether Claude out-benchmarks Copilot next quarter. It’s whether the biggest GTM platform absorbs Claude natively, or the labs keep renting the CRM as their operating system.

See the roles here!

 

Damian’s insight: Notice the quiet respect the AI leaders now pay to the old, boring software, Salesforce, Oracle, SAP. Those platforms already hold the customers and the trust, so the winning move isn’t to replace them, it’s to drop a feature into an install base that never has to switch. Distribution isn’t a channel, it’s millions of accounts you don’t have to earn twice.

Even OpenAI is hiring human BDRs, and the irony is the whole point 📞

Then OpenAI, the company literally behind every “AI SDR” on the market, quietly posts a role for a human Strategic Business Development Representative in Dublin. The meme writes itself.

And the job is the punchline. A human BDR inside OpenAI’s Enterprise GTM org, financial services to retail, 4+ years of enterprise experience, turning product usage and pilots into qualified opportunities. Buried in the responsibilities: “build AI-enabled GTM workflows (GPTs and agents) to scale prospecting.” OpenAI wants a human to operate the AI that does the prospecting. That’s the entire future of this role in one bullet.

The data backs the instinct, not the hype. Yes, ~36% of B2B companies cut SDR headcount last year and SDR-to-AE promotion fell from 34% to 16%. But reps who use AI well are 3.7x more likely to hit quota, and signal-personalized outreach pulls 15–25% reply rates versus the 3–5% cold average. The role isn’t dying, it’s splitting: the reps who compete with the tool, and the reps who operate it. So copy OpenAI’s own spec, build the workflows instead of just running them, get religious about signals (an active trigger wins 37% of deals versus 19% for cold), and protect the human discovery conversation AI can’t fake.

Read the OpenAI role here!

Damian’s insight: The tooling keeps getting simpler and AI more accessible, so a sharp BDR can now do a big slice of the GTM engineering themselves, tuned to the exact product they sell. Their feedback from the market loops straight back into the workflow and retunes it in real time. An AI-native BDR/SDR turns that into fresh revenue fast, but only if the GTM infrastructure underneath is actually built for it.

 

This is the room where we pull apart how a real operator actually runs go-to-market.

Carles Reina on rebuilding ElevenLabs from scratch 🔥

Carles Reina, who runs GTM at ElevenLabs and is Managing Partner at Baobab Ventures, dropped a post on what he’d do differently if he rebuilt ElevenLabs’ GTM from zero. Not theory, a GTM leader marking his own homework in public.

The headline admission is a regret, not a flex: hire quicker. Reina spent the first nine months as the only person selling, convinced ElevenLabs didn’t have product-market fit yet and slowed by 2023’s “AI tourist” skepticism, because the company was still mostly PLG and self-service. He didn’t boost the GTM team until demand forced it earlier this year.

The full list is a masterclass:

→ Hire quicker.
→ Build an enablement team earlier.
→ Open more markets, sooner.
→ Make bigger bets.
→ Be louder about ROI.
→ Hire senior sellers.
→ Trade cash commissions for equity.
→ Open more offices.
→ Build a fun brand.
→ Outbound, outbound, outbound.
→ Build pipeline with more whales.
→ Not all deals are good deals.

Notice how little of the list is about the product. It’s markets, enablement, ROI, comp, brand, outbound and deal quality, the motion, not the machine.

Read the full post here!

Damian’s insight: Modern GTM gives you something rare: speed of validation. You can test markets, outbound angles and positioning in days, not quarters, and that forces fast decisions. You’d assume a product as hot as ElevenLabs pulls leads in on its own, yet Carles repeats outbound X3 in one list. Outbound lands the whales.
Found this useful?
Subscribe GTM Club
EN Nie odkładaj na jutro. EN Zoptymalizuj sprzedaż z Dealme dzisiaj!
Sign up Book a meeting