revelations of an AI sales consultant

👋 What’s up folks!

This is Prompt Punk — the best AI sales newsletter 😉

TL;DR - inside today’s newsletter

Let’s rip…

I follow a bunch of AI transformation consultants to learn from their experience in the field. Where are people are succeeding, getting stuck, failing.

They mostly try to get companies to adopt AI tech to improve business performance: Revenue, costs and other efficiencies.

Here is a post from Alex Lieberman, he is famous for selling his business newsletter, Morning Brew for $75M. Now he is building an AI consultancy - Tenex.

Here are his revelations of an AI sales consultant:

What stalls programs: no owner, no measurable goal, massive scope, cheap models too early. Expect quick proofs, but the organizational digestion of workflows, systems, habits can be multi-year.

Bottom line: Alex is right on direction the ceiling is higher than most think. Give it an owner, buy the good model, sequence wins, and stack proofs.

Here’s a neat trick Granola (AI meeting notes) customers are using.

Create a template in Granola for your meeting notes + push to CRM + recommend next steps + draft follow up.

I like this because you can just plug in a sales methodology like SPIN Selling, Challenger Sale or MEDDPICC, then your call notes can fit into the schema. Great visibility for leadership too.

Result

No more “forgot to log it” excuses. Notes auto-structured (SPIN/Challenger/MEDDPICC), pushed to CRM, tasks created. Follow-ups drafted in minutes, pipeline hygiene improves, deal reviews get clearer.

Customers notice

Same-day, specific follow-ups with owners and dates. Fewer “what did we decide?” Slacks. Reps stay present in the call, leaders get accurate notes.

Why it matters

The ROI isn’t from “AI notes,” it’s from shipping the workflow: capture → structure → CRM → next step → follow-up. Tools change, the playbook stays the same. Everything should move faster.

Sweden’s legal-AI upstart Legora just closed $150M at a $1.8B valuation to scale its legal AI grunt-work killer. Maybe not so good for paralegals.

I once met the CTO and liked his ambition. Customers also speak highly about the product.

Their plan: push from Nordic dominance to global enterprise rollouts and expand headcount aggressively. Obviously competing hard with Harvey.

This deal is proof that the “AI ops layer” wins when it plugs into existing workflows, not just chat. Europe just minted another meaningful AI platform in a category (legal) that touches every deal, dispute, and diligence.

Why it matters
Vertical AI that ties to revenue capture and cost avoidance keeps getting funded even in late 2025. Expect copycats in compliance and audit.

Mistral continue to ship, they just launched AI Studio, packaging the plumbing enterprises actually need: observability, evals, versioning, governance, and runtime orchestration. I am impressed.

The goal? fewer “cool demos,” more deployed workflows. With AI Studio, Mistral attempts to move from model vendor to platform where contracts are bigger, stickier, and measured on uptime and SLA, not vibes.

This is Mistral’s enterprise armor needed against hyperscaler defaults.

Why it matters
The AI war isn’t model-versus-model, it’s production versus prototype. Studio is Mistral’s bet on the only fight that counts, shipping stuff!

Anthropic announced a landmark expansion with Google Cloud: access to up to one million TPUs (Google’s proprietary AI chips), bringing well over a gigawatt online in 2026 a tens-of-billions compute commitment to train and serve future Claude models. Amazon is still named as Anthropic’s primary cloud provider but this deal suggests their largest forward-looking compute might be Google.

Beyond the headline scale, it signals serious vendor hedging against the NVIDIA dependency.

For users, this should translate to better availability and pricin. For rivals, it’s a reminder that the real moat is silicon and distribution.

Why it matters
This is Anthropic’s grand strategy vs. OpenAI: split the stack, training and distribution on AWS and Google. Make the two hyperscalers compete for Claude’s roadmap.

Where OpenAI is tightly coupled to Microsoft Azure + NVIDIA, Anthropic now has an informal alliance with Amazon and Google.

Google gets a marquee TPU showcase, AWS keeps enterprise reach via Bedrock. Net-net: buyers gain availability and OpenAI faces pressure from a cross-cloud, cross-silicon rival that’s structurally harder to block.

On Alphabet’s Q3 2025 Earnings call, Sundar Pichai said Gemini 3 lands later this year, flagging upgrades in reasoning and coding and touting 650M+ MAUs for the Gemini app.

The subtext: Google is leaning into a full-stack play (chips → models → products) and wants enterprises to default to a first-party stack for agents, code, and content.

For buyers, another top-shelf model in 2025 is good news, for procurement, it cements the multi-model reality (Claude, GPT, Gemini) with more room to mix by task.

Why it matters
A live, dated roadmap from Google keeps pressure on OpenAI and Anthropic and gives execs cover to expand “Gemini seats” before year-end. Where is xAI now?

Microsoft and OpenAI cut a new deal that removes fundraising constraints on OpenAI, paving the way for a more traditional corporate structure and an eventual IPO.

Microsoft keeps a ~27% stake, but drops key rights while securing massive Azure commitments, $250B.

The result? OpenAI can raise at will, buy compute where it wants, and sprint toward bigger models and a bigger business AFTER paying Microsoft first. For Microsoft, it de-risks governance drama while keeping the Azure flywheel humming.

Why it matters
OpenAI gets independence-ish, free to raise and go multi-cloud-ish for training, while Microsoft still clips their revenue via massive pre-committed Azure spend. That makes Microsoft the toll collector on OpenAI’s growth even if workloads spill to other clouds.

From Microsoft CFO, Amy Hood…

AI is short of power and datacenter infrastructure not GPUs.

That’s bullish for datacenter operators and power companies.

Thanks for reading! 👋

If you liked this:

📨 Forward to a biz techie friend
💬 Hit reply and tell me what you want more (or less) of
📊Vote in the poll below 👇

— John
Prompt Punk