GTM engineering · HubSpot + Claude
Periti builds go-to-market engines for companies with 200–2,000 employees. HubSpot holds the truth, Claude does the thinking, and your own customer history makes every message impossible for competitors to copy.
New VP Sales appointed at a 600-person logistics company · day 9 in role
Matched to 14 closed-won deals in logistics · top reason for buying: forecast accuracy
Researched the account, new leader's priorities and current stack · drafted a one-to-one message
Sequenced, sent and tracked · reply routed to the account owner
1.5M
HubSpot AI credits a month for our largest GTM client, Eventus
50%
of Eventus’ sales activity now involves AI outreach
1,000%+
ROI in our own tests of this stack, on our own pipeline
900+
customers served across Europe and the US
The shift
For years, the expert advice on outbound was to send fewer, better messages. That advice was right. Researching an account properly took an SDR 20 minutes or more, so the only way to send good copy was to send less of it.
That constraint is gone. Claude can read a company’s website, its job posts, its latest acquisition announcement and your own history with companies like it, then write a message that makes sense to the specific person receiving it. It can do that for ten thousand accounts as easily as for ten.
Before
20+ minutes of SDR research per account. Good copy meant low volume; volume meant generic copy.
Now
Every account researched, every message grounded in your customer context. Ten thousand accounts as easily as ten.
If you can send excellent copy to your ideal customer profile, volume is the only way to go. Every account that doesn’t hear from you is an account your competitor gets to first.
The catch is the word “excellent.” Volume with generic copy burns your domains, annoys your market and teaches buyers to ignore you. Volume is the reward for getting context right. It’s never a substitute for it.
The problem
Most of the companies we talk to already own the tools: Clay for enrichment, Instantly for sending, HubSpot for the CRM and a few AI writing tools on top. Each one was bought by a different person to solve a different problem. None of it shares a central theme or a business case.
01
Data sits in Clay tables that drift out of date the week after they’re built.
02
Sequences run in Instantly on messaging nobody has looked at since launch.
03
Nobody can tell which campaigns produce pipeline. Leadership sees activity, not return.
04
A decade of identical tools and playbooks taught buyers to spot a template in the first line.
Nobody else has your closed-won history, your customer relationships or your understanding of why people buy from you. When that context drives every message, your outreach can’t be copied. That’s what we build: one engine with HubSpot at the center, Claude as the intelligence layer, and the rest of the stack chosen to serve a single business case.
How the engine works
Claude is the model. Your customer history is the context. HubSpot and the sending stack are the application. Agents and automations are only as good as the context underneath them, and context is the part your competitors can’t buy.
The stack
We build with the latest tools and change them when something better comes along. What stays constant is the design: HubSpot holds the truth, Claude does the thinking, and everything else serves those two.
System of record
Sequences, workflows and reporting. Holds every account, contact, touch and reply, so every campaign can be measured.
Intelligence layer
Researches accounts, interprets signals and writes copy grounded in your customer context.
Native HubSpot AI
Monitors buying signals, enriches records and drafts outreach inside HubSpot.
Enrichment & signals
Finds the right companies and contacts, and watches for trigger events.
Sending at scale
High-volume email and social touches with proper domain and deliverability management.
Orchestration
Connects it all, with business-critical flows running on Azure or AWS.
Breeze or Claude?
HubSpot’s Breeze Prospecting Agent is a good tool, and we use it. It works best on data that already lives in HubSpot. Claude comes in when the context you need sits outside it, or when you need control over the reasoning and the copy at real volume.
Signal monitoring and outreach drafting on accounts already in HubSpot. Ideal for teams starting out with AI prospecting.
Large addressable markets, custom research, context from outside HubSpot, campaigns across thousands of accounts.
Most mature programs. HubSpot runs the record and the sequences; Claude does the heavy research and writing.
Anyone can buy this stack, and plenty of companies have. Without a strategy behind it and people executing it every week, it produces a lot of activity and very little pipeline. The value comes from the decisions: which segment to start with, which signals matter, what the messaging says and to whom, how many touches each account gets, and what to change when the numbers come back.
Signal-based outreach
The best time to reach a company is when something has just changed. We build trigger-based outreach that watches the market for those changes and acts on them within days, not quarters.
A team is growing or a function is being rebuilt
Speaks to the problem the new hire is being brought in to fix
Budget and priorities are about to be reset
Reaches the new leader in their first 90 days with a relevant point of view
Systems, teams and data need to be combined
Addresses consolidation, integration and reporting across the new group
A business needs its own systems, fast
Offers a path to standing up independent infrastructure
Growth targets just went up
Connects the new targets to the capacity needed to hit them
New teams, languages and processes
Speaks to scaling the playbook without losing control
A platform is being replaced or added
Positions your offer around the change already underway
A warm relationship at a new account
Reopens the conversation with shared history
Signals come from your enrichment tools, from HubSpot’s own data and from Claude reading public sources directly. The point isn’t to collect signals. It’s to decide in advance what each one should trigger, then let the engine act on it every time.
Case study · Eventus
Our biggest GTM engineering client is the highest-spending AI credits customer in the HubSpot ecosystem. We started by mapping their entire addressable market, then designed an outreach program built for their buyers and laid out over 12 months, so they knew in advance how many touchpoints every account would receive, and when.
Before these AI tools, there was no way to do that across thousands of accounts. The research alone would have needed a building full of SDRs. They’ve since cut SDR and BDR overhead significantly, and bookings have risen sharply.
“Working with Periti has made a huge difference in terms of speed and execution. It allowed us to get everything set up properly and start seeing value much faster than if we’d tried to do it all ourselves.”
Katy P., Head of Marketing, Eventus
1.5M
HubSpot AI credits used per month
50%
of all sales activity now involves AI outreach
~5 hrs
saved per rep, per week
12 mo
touchpoint plan covering every account in the market
Everything on this page runs Periti’s own go-to-market before we build it for a client. When a cheaper or faster tool appears, we test it on our own pipeline and switch if it wins. Our own tests have returned ROI in the thousands of percent. We tested it with our own money before asking you to trust it with yours.
The question every CRO asks
Here’s a straight answer. No SDR team can reach as many people, in as many relevant ways, as a well-built AI engine. An SDR can research and write maybe a few dozen good messages a day. The engine does that for your whole market.
What the engine doesn’t have is the personal touch. That’s where your people come in. For most of our clients the result is a smaller, better-paid SDR function with much higher output, and we’ll help you plan that transition honestly, including what it means for the people already on the team.
Manual SDR team
SDR + engine
How an engagement runs
Anyone selling GTM engineering as a three-week setup is selling you a tool installation. We form hypotheses about which segments, signals and messages will produce pipeline, test them through campaigns, read the results, gather feedback from your sales team and refine. Then we do it again.
01
We map your market, ICP and customer history, decide which signals matter and define success in pipeline terms, not activity.
02
We connect HubSpot, Claude, enrichment and sending into one engine, with reporting so every touch is measured.
03
We start with a defined segment and a small set of signals, warm up sending properly and get the first campaigns live.
04
Every month we review results with your team, retire what isn’t working and double down on what is.
05
Frontier models improve every few months. We bring those improvements into your engine so you’re never running last year’s capability.
A Periti go-to-market strategist and our team work for your team on a fractional basis. You get senior GTM thinking, engineering and execution without hiring a department.
Who this works best for
GTM engineering pays off fastest for companies with a large addressable market and a sales cycle that rewards being early.
CROs & sales leaders
The engine takes over research and writing, so fewer people produce more meetings.
Marketing leaders
We turn closed-won history into your ICP, segments and messaging.
COOs & RevOps
We tie them into one engine with HubSpot at the center, and every touch measured.
Growth-stage companies
We map every account and plan touches across all of them over the year.
New leadership teams
A measurable outbound motion in place within the first quarter.
Private equity
A GTM playbook that rolls out across the portfolio, with shared definitions and reporting.
How we work
Some requests we’ll challenge, and a few we won’t take on.
If your CRM data is a mess or nobody agrees on the ICP, more messages only spread the problem. We fix the foundation first.
Every message carries your brand. If it reads like a machine wrote it, we keep working on it before it goes out.
If replies, meetings and pipeline don’t land in HubSpot, nobody can prove what’s working. Every program reports back.
One campaign can book meetings. An engine takes a year of testing to get right, and we’ll tell you that up front.
And when HubSpot isn’t the answer, we say so. If the real problem is process and automation rather than the platform, we draw on our Anthropic partnership and automation consulting to scope the smaller project that fixes it.
Your team
Where to start
Most GTM engineering work starts with a Design and Strategy project. If your HubSpot data isn’t ready to power outreach, that’s where we focus first.
LaunchRight, our industry-specific onboarding package, gets you onto a proven HubSpot setup fast, with a foundation that can carry a GTM engine later.
Quick questions
Go-to-market engineering is the work of automating and systemizing how a company communicates with its market. At Periti, it means building one engine from HubSpot, Claude, enrichment and sending tools, fed by your customer context, and running it as a 12-month program of testing and refinement.
Yes, if the copy is excellent. The old advice to send fewer messages came from a time when good research was slow and expensive. With Claude doing the research and writing, you can send relevant messages to your whole ideal customer profile, and every account you don't reach is one a competitor will.
We use Claude through the Claude API and Claude Console to research each account, interpret the signal that triggered the outreach, and write copy grounded in your own customer history and case studies. Our strategists set the messaging framework, review output and refine it based on what gets replies.
Common triggers include hiring for specific roles, new leadership, mergers and acquisitions, divestments, funding rounds, market expansion and technology changes. We agree in advance what each signal should trigger, so the engine acts on it consistently and quickly.
Mostly, at the research and writing stage. One SDR working with a well-built engine typically delivers far more than a full team working manually. What AI can't replace is the human conversation once someone replies, so most clients keep a smaller, more focused team.
Often both. The Breeze Prospecting Agent works well on data already in HubSpot. Claude is the better fit for large markets, custom research and context that sits outside HubSpot. In mature programs, HubSpot runs the record and sequences while Claude does the heavy research and writing.
HubSpot is always at the center. Around it we use Claude, Deepline, Clay, Instantly and LinkedIn outreach tools, with Microsoft Azure or AWS for business-critical connections. We choose and change tools to fit your volume, budget and goals, because the tools are never the advantage on their own.
We typically work with clients for 12 months. The first campaigns can be in market within weeks, but the real value comes from testing hypotheses, refining messaging and upgrading the engine as models improve over the full year.
Your closed-won deals, lost deals, ERP and invoicing data show exactly who buys from you and why. When that context drives the ICP and the messaging, your outreach becomes something competitors can't copy, even if they own the same tools.
It helps a lot. If the data isn't trustworthy, the engine will spread bad assumptions at scale. Where needed, we start with a Design and Strategy project to fix the foundations before switching on volume.
Yes. The same approach can become a repeatable playbook across portfolio companies, with shared definitions and reporting so operating partners can compare outbound performance across businesses.
It can, if it's done badly. We manage sending infrastructure and deliverability properly, and nothing goes out until the copy reads like a person wrote it for the reader. Every program reports back into HubSpot so problems show up early.
If your SDR team is expensive and underperforming, your tools don’t add up to a strategy, or you know there’s far more of your market you should be reaching, tell us where you are. We’ll tell you honestly what an engine built on your context could do.