From bolt-on to backbone: SaaStr’s verdict on AI in SaaS
SaaStr’s verdict was blunt: AI has stopped being a feature you add and become the thing you build around. Charlie Haddrell, NZTE’s Trade Commissioner in San Francisco, was there with a dozen-odd Kiwi companies and the NZTE on-the-ground crew. Here’s his read.
The drumbeat all week was unmistakable: everyone’s done experimenting with AI - now they’re operationalising it. Across teams, workflows, go-to-market, infrastructure, the lot.
The stack,that’s actually winning
Forget the keynotes; the gold was hearing exactly how teams have wired their go-to-market stack to sell into the US, and which tools are pulling their weight. The same shape kept surfacing: one CRM, one data layer, one call-intelligence layer, and an AI brain tying it all together.
For a US-focused BDM team, the example was:
HubSpot or Salesforce = the CRM backbone
Apollo or ZoomInfo = contacts and account data
Clay = enrichment
Gong or Attention = call intelligence
Claude = research, content, enablement, the brain
Here’s how it snaps together.
Nail your ICP. Point Claude at your customers, win/loss notes and market data and it spits out ICP tiers, buying triggers, US-specific messaging and competitor positioning, the foundation everything else is built on. (One catch: it’s only as sharp as your CRM data. As the old adage goes: garbage in, garbage out.) Anthropic was everywhere at SaaStr, and it’s obvious why; it can do your account research, campaign writing, call prep, objection handling, proposal drafting, internal playbooks, and all from one relatively low-cost layer. Claude could be the cheapest high-leverage move you’ll make this year.
Feed the pipeline. Clay sits between your CRM and every other workflow, fusing data enrichment with AI research agents to find prospects, mine dozens of sources (LinkedIn, Apollo, job boards) and build razor-targeted pipelines, no army of SDRs required. One Aussie SaaS outfit runs it at ~US$80k a year across a 20-person BDM team. Sounds steep, until you realise that’s an entire team’s grunt work, automated, driving outcomes and creating value.
Personalise at scale. Claude and Clay together draft hyper-personalised emails that name the prospect’s role, their company’s latest news and recent events, a competitor’s offering, an industry-specific pain point, the outreach that used to swallow a rep’s whole afternoon, done in seconds.
Listen to every call. Gong records the lot and flags objections, competitor mentions and pricing signals, reading buying intent straight off the transcript. It’s the benchmark, but it bites on price: a team of six lands around US$20k a year. The challenger turning heads is Attention, it transcribes and analyses calls in real time and coaches reps live, mid-conversation. That live element is gold for teams selling into the US from afar, nudging reps to American comms norms on the fly. And at about US$100/user/month, it’s (to use a kiwi term) capital efficient.
Close the loop. The CRM soaks up Gong’s insights, Clay’s enrichment and AI notes; Claude summarises every call and drafts the next move; follow-ups, industry-matched case studies - straight back in. Zoom out to a more macro level and it reads pipeline health, win patterns and individual rep performance to fire out forecasts, deal-risk alerts and messaging tweaks. Claude becomes the brain of the whole machine.
This isn’t a moonshot. It’s a handful of tools, most of which you already know, assembled with intent. Not all are needed from day one, so get advice on how to build and right-size the stack as you scale.
What this means for NZ SaaS
The bigger shifts are the ones to plan around.
US buyers have serious app fatigue. CIOs aren’t bloating their stacks like they did 3–5 years ago; they’re slashing them. And AI isn’t just replacing software; it’s cutting the number of people needed to run the software. Budgets are now being traded directly between SaaS subscriptions and AI spend. The “running the machine” layer of business is being re-priced in real time. Strategy, judgment, creativity, relationships, trust still mean there needs to be humans in the loop. But the operational layer is being rebuilt under everyone’s feet.
For NZ exporters, bolting AI onto the side won’t cut it. US buyers are turning AI-native: they want fast implementation, measurable ROI, self-serve onboarding and clear outcomes - not a feature list. The winners will build AI into the core product and operating model, compete on speed-to-value, run AI internally across sales, marketing, onboarding and support, and defend real moats such as proprietary data, workflow integration, compliance, switching costs.
One warning worth repeating: AI doesn’t fix broken go-to-market. If outbound wasn’t working before, AI SDRs certainly won’t save you; they just accelerate the mess. Increasingly too, smart players are building their own lightweight internal tools instead of buying off-the-shelf.
The bottom line? AI-native companies; your Lovables, your Anthropics, are running at efficiency levels traditional SaaS can’t touch, with a fraction of the headcount. Analysts tip AI to swallow over US$1 trillion in IT spend by 2029, most of it flowing to a tight pack of leaders.
NZ SaaS companies have always adapted to changing market dynamics over the last decade. But this is moving at breakneck speed, and the US competition increasingly has AI as the backbone of their product and GTM, so we need to build AI-native, prove value fast (and clearly), adopt the latest tools to create efficiencies and drive growth – and we can take on the US with products that don’t just compete, they redefine the category!
By
Charlie Haddrell, NZTE’s Trade Commissioner in San Francisco,