Your startup might have an amazing point solution, but when it comes to moving up the ladder and securing those big enterprise accounts, you might find yourself falling short.
Let me share a little secret from my years in the tech industry: the company that lands the enterprise account often doesn’t have the best in class point solution.
Instead, it’s the one that’s already on the inside, trusted, and prepared when the board starts asking questions.
We'll explore the reasons behind this soon, but let's start by discussing the pressure all companies are under at the moment.
Right now, there's a universal conversation happening in boardrooms everywhere: "How are we addressing AI?" It's the most widespread directive I've witnessed in my career, and the pressure on CEOs and executive teams to provide a solid response is immense.
That isn't an epiphany; everyone can feel it. What's more interesting is what happens next, because for most organisations, answering the question is proving to be a quagmire.
AI Pilled? No, not really.
The obvious first move is to hand everyone a licence to ChatGPT, Claude, or Copilot, maybe bring in a big consulting firm to "lead the rollout", and call it transformation. No one ever got fired for buying IBM.
But those tools are individual productivity boosters. They are not an organisation changing how it works across end-to-end workflows. Making that leap takes a level of sophistication most companies simply don't have in-house.
You need someone who can say: here's how to think about context, here's how to set guardrails, here's how to get your data in order, and here's how to publish this so a whole team actually uses it.
At Firmable, we’ve been making that leap ourselves, led by our CTO Karthik and CPO Chath. We’re AI-native across engineering and operations. We’ve invested heavily in our underlying AI infrastructure (Snowflake, reporting, MCP servers) and built our own internal environment, Firmable Labs, where our team can develop and publish AI-generated apps.
We're in the middle of a shift from artifacts to applications. We use Claude to generate skills that solve workflows, then move them into an internal environment where they run reliably, like software, rather than regenerating each time. This lets our team experiment with skills and automation themselves, and when a strong business case emerges, anyone in our team can then turn that skill into a dependable application.
I'd still only put us at about six out of ten. (If you want to know where your own company sits, Point Nine's "how AI-pilled are you" diagnostic scores you on a 0–100 scale; most companies land between 1 and 30.)
Your buyer doesn’t want ten best in class tools. They want a partner who knows what best in class is and how to get there.
There's a three-speed market for AI right now.
At the fast end are the startups: adaptable, willing to test every tool that launches and self-select the winners. At the other end are the large enterprises, where the problem isn't appetite but control — whole teams are going gangbusters with AI and the spend becomes impossible to manage. Uber capped employee AI spending this month after blowing through its budget in four months.
But the middle is where 98% of businesses are.
Picture the mid-market executive who can feel the power of the technology but has no idea what to do next. They don't have a spend problem, and they don't have the context, knowledge, or bandwidth to evaluate point solutions.
Think about a CFO bringing AI into a finance team, or a head of sales at a B2B insurer trying to drive AI through their org. They do not want to evaluate ten tools, work out how to stitch them together, and personally own the integration risk. They'd much rather work with a vendor they already trust — one who can hand them a couple of obvious first wins and spare them the confusing landscape as embed AI across the organisation.
Watch what's happened to Salesforce. Plenty of people assumed the incumbents would get smashed by AI.
Instead, buyers are now turning towards them: "I'd rather just do this with Salesforce." Is their agent platform the best on the market? Probably not. But the customer has the relationship, the deployment, and the trust. And that beats marginally better features.
That's the opening for startups. If your product is already embedded in a function, you can be that partner. Your team walks into the CEO's office and says: "We're already working with you. We've put AI into this workflow, and we've noticed three more that are still manual. We can fix those too." Suddenly there's a safe answer for the executive, deployment is handled, spend is managed, and the customer looks like they're leading rather than flailing.
Now you’ve helped create an AI transformation initiative that CEO can take to their board, and you can be the partner.
Becoming that partner takes three deliberate shifts.
Shift one: reposition around the transformation agenda
This is a positioning change before it's a product change. When you attach your product to an AI-transformation agenda, your buyer changes. You go from selling to a manager to being a C-level priority.
Every good enterprise seller knows the game is to tie into the big strategic initiative, because that's where new budget is created. Right now, AI transformation is the strategic initiative. Align your value proposition with it and you stop being a cost to scrutinise and become the answer to what the board is demanding.
The clearest case is selling to a cost centre. HR software sold to HR managers is a hard internal sell. The end user has little budget and limited influence. Reframe the same product as "this is how we roll an AI-competency agenda through the organisation", and you're in a different conversation entirely. The board has told the CEO to think about the skills people will need; now your HR buyer has a story that travels all the way up.
Potentially the same product with a different value attached to it.
Shift two: the minimum viable relationship
When software was expensive to build, getting in meant a six-month implementation, and expansion meant building a whole new module. The customer wants an adjacent feature built? Go away, build one, test it, hope for the upsell.
AI collapses that. You can land an account on a single wedge workflow, and once you're in, the next workflow is a configuration of AI agents, not a build of new modules.
I recently had lunch with Hemant Taneja, CEO of General Catalyst who shared a term he coined for the strategy that Heidi (and many others) are running: the minimum viable relationship.
Let’s look at how Heidi is growing in healthcare and hospitals.
In Heidi's case, it plays out like this. They start with a single point of contact through their scribe note-taking tool, say an emergency department in a hospital adopting the Heidi app.
Once they're inside, they notice teams are drowning in discharge summaries, so they configure their agents to handle that. Then, because they're already capturing everything, they extend into billing and coding workflows too.
None of it required new products, just configuration of a flexible, agentic platform. Heidi is now becoming the AI reference point in its sector, and a hospital that has already run procurement once will find it far easier to keep expanding with the same partner than to start over with someone new.
Shift three: lean in with humans but don’t become a consultancy
What used to be called professional services is now forward-deployed engineers. It can be simple: here are your licences, and for two weeks we'll come in, set it up properly, and get your team getting value. While your team is embedded, having lunch with the team, you learn the next problem you can solve.
I recently sat down with Romilly Blackburn, the cofounder of Operata, to glean some insight on their forward deployed engineer strategy.
1. Treat every customer deployment as R&D, not services. Operata's "Applied Engineering" (Forward Deployed Engineer) function looks like services, sales engineering, and customer success combined. Companies with complex/AI products should embed real engineers with strategic customers and feed what they learn back into the core product, rather than forking the platform or racking up billable hours.
2. The "last mile" is where enterprise AI adoption actually happens. Because every enterprise stack is bespoke, the value of your product is only realized once it's wired into the customer's environment. Founders tailor the last mile of FDE work on top of a strong platform) over rebuild 80–90% per customer.
3. Operational dependency beats contractual lock-in. The stickiest position in SaaS isn't a contract, it's a system that's embedded in how the customer operates day to day. FDE-style deep deployment also powers land-and-expand (a $150K beachhead becoming a $500K account) and shortens enterprise sales cycles, because a credible technical counterpart in a late-stage eval answers the two questions that stall procurement: "does this work in our environment?" and "who will make it work?"
The big thing to call out here is not to become a consulting business. The goal is to build enough product value that you can sell in a consultative, enterprise way. Do not to become a services firm with a thin layer of tech on top. Have a point-solution wedge with a few to building a platform, but never a point-solution vision.
Where this works, and where it doesn't
It's worth being clear about who this is really for, because the forward-deployed model isn't a fit for everyone. Its sweet spot is the mid-market — that 98% in the middle of the three-speed market, with the appetite but neither the in-house capability nor the bandwidth to stitch the landscape together themselves.
That's exactly where a trusted partner who can land a wedge and grow it matters most, and it's where we've concentrated: at Firmable, 80% of our customers currently sit in this band.
Within that, it works best in three places:
Organisations that lack the internal capability to do it themselves — the B2B sales leader who knows the world has moved and knows they should be doing more with AI but has no modern revenue-operations muscle and no idea where to start. Operata and Firmable both sell into versions of this.
Regulated industries, where procurement is hard and buyers want a partner to navigate the journey rather than a net-new vendor to approve. Healthcare is the obvious one — it's a big part of why Heidi has such a strong position and the opportunity ahead of it.
Functions with deep, technical workflows, where your platform is already woven into everything the customer does. ProcurePro in construction procurement is a good example.
The takeaway
Remember software has always been how organisations change their own behaviour.
When I rolled out Salesforce at Aconex nearly two decades ago, I didn't buy it because I loved the CRM. I bought it because we needed more maturity and rigour in our sales process. The tool was the change agent. AI rollouts will follow the same story.
So the winners of this wave won't be whoever has the best point solution. They'll be the partner who was inside, trusted, and ready to act and grow when the board started asking the question.
