If you are two people and one of you is selling, your go-to-market stack should cost a few hundred dollars a month and take a week to set up. Most early teams spend more than that and get less, because they buy the tools their last company used at a hundred people.
Here is what I would actually run.
The model is the stack
Start here, because it changes everything downstream. A frontier model with good context handling does more for a two-person commercial function than any purpose-built sales tool on the market.
Account research that took forty minutes takes four. First-draft outbound that took ten minutes takes one. Call notes write themselves. Competitive teardowns, pricing analysis, ICP definition, objection-handling documents: all of it becomes a prompt and an edit rather than a project.
This is the highest-leverage line item and most teams underuse it badly. They treat it as a writing assistant rather than as the research layer the entire motion sits on.
CRM: Attio or HubSpot
Attio if you want something fast and modern that behaves like a spreadsheet. HubSpot if you expect to hire a sales team within the year and want something they will already know.
Both are free or nearly free at your size. The mistake is not which one you pick. The mistake is picking neither, running pipeline out of a spreadsheet for eight months, and losing all the history when you finally migrate.
Data and enrichment: Clay
Clay is the most useful piece of the modern stack and the most commonly misused. It is not a lead list. It is a programmable enrichment layer that lets you build account lists from signals rather than firmographics.
Hiring for a role that implies your problem. Running a competitor. Recently funded. A specific technology in the stack. Those lists are small, specific, and they convert. A list of every Series A software company in the UK converts at approximately nothing.
Sequencing: Lemlist, Smartlead or similar
Whatever you use, the constraint is not the tool. It is volume discipline. At two people you should be sending tens of emails a day, not thousands, and every one should be good enough that you would be comfortable if the recipient forwarded it to a friend.
Deliverability matters more than features. Separate sending domain, warmed properly, sensible daily caps. Get this wrong and nothing else in the stack matters.
Automation: n8n or Make
For everything that should happen without you touching it. A new signup gets enriched and lands in the CRM. A call recording summarises itself and attaches to the account. The pipeline report builds on Friday morning whether you remember it or not.
Each of these saves ten minutes. Ten minutes, twenty times over, is most of a working week.
Call notes: Granola
Worth it from the first call, not once you have a team. The value is not the notes themselves. It is being able to search every conversation you have ever had for what buyers actually say when they object, in their words rather than your summary of them.
What I would skip
Intent data at this stage. Expensive, noisy, and you do not have the volume to act on it properly.
Anything with annual-only pricing. You will change your mind about your motion at least twice in the next year.
A full sales engagement platform with forecasting and territory management. You have one territory, it is called everywhere, and you can forecast it in your head.
The part that is not a tool
None of this fixes a message nobody understands. I have seen teams with beautifully built stacks and no pipeline, because they automated the delivery of something nobody wanted to read.
Build the stack after you know what you are saying and who you are saying it to. It is an amplifier, and amplifiers work in both directions.
Already selling, with a stack that has not kept up? That is the work in AI enablement for revenue teams.