Run the Business vs Do the Work — The AI-Era Vertical SaaS Shift

The two categories

Run the business. Software that helps a business operate — track customers, schedule appointments, manage inventory, process payroll, send invoices, generate reports. The user is usually an owner, manager, or administrator. The software is sold on features, reliability, and operational fit. The buying process involves demos, sales engineers, contract negotiation, and procurement review. Vertical examples: DentalCloud (dental practice management), Mindbody (fitness studio management), Shopify for e-commerce, HubSpot for SMB sales management.

Do the work. Software that helps a practitioner actually do the thing they get paid for. Transcribing a vet’s session notes. Drafting a lawyer’s brief. Summarizing a doctor’s appointment. Preparing an accountant’s deliverables. Running a real estate agent’s market analysis. Composing a journalist’s story. Editing a musician’s track. The user is the practitioner. The software is sold on magical experience and immediate value delivered to the person doing the work. The buying process is self-serve sign-up, free trial, and a credit card.

The distinction sounds obvious, but it maps onto genuinely different product design, go-to-market, and pricing choices. A company building “run the business” software is not the same kind of company as one building “do the work” software, and the ones trying to be both usually end up confused.

Why this shift is happening now

Three enabling conditions came together in the 2023-2026 window:

1. Foundation models are good enough at vertical-specific tasks. Five years ago, building a tool that “drafts a legal brief” would have required deep NLP expertise and custom models trained on legal corpora. In 2026, a startup can wrap Claude or GPT with vertical-specific prompts, templates, and context and ship a product that does the work credibly. The technology bar dropped from “research team” to “two-person startup.”

2. PLG distribution is mature. The playbook for getting a product into the hands of individual practitioners — free trials, self-serve billing, viral content, SEO, referral loops — is well-understood. The infrastructure (Stripe, analytics, auth, trial management) is commoditized. A startup can spin up a PLG motion in weeks, not quarters.

3. Practitioners are drowning in administrative work. Veterinarians with hours of post-session documentation. Lawyers reviewing thousands of pages of contracts. Doctors typing notes during appointments instead of looking at patients. The pain in the practitioner layer of every vertical is substantial and unaddressed, because previous vertical SaaS targeted owners and never cared about it.

When the technology is cheap enough, the distribution is frictionless enough, and the pain is visible enough, a new category of software emerges. We’re in that moment for “do the work” vertical software.

What changes when you target the work, not the business

Five concrete shifts between a “run the business” product and a “do the work” product:

1. The buyer changes. Owners and IT buyers give way to practitioners with personal-card spending authority. See Hero User Strategy - Native AI’s Wedge Into Vertical Software. The new buyer makes decisions in minutes, not months.

2. The product aesthetic changes. “Run the business” products can be ugly as long as they’re comprehensive. “Do the work” products have to feel magical from the first interaction. Polish, responsiveness, and simplicity matter enormously. A single moment of friction loses the user.

3. The unit of value changes. “Run the business” software is priced per seat per month — a flat cost for access. “Do the work” software is priced per task, per outcome, per unit of value delivered. Seat pricing doesn’t capture the variance in how much each practitioner benefits. See Seat-Based to Token-Based SaaS Pricing Transition.

4. The feedback loop changes. “Run the business” products succeed by being feature-complete and reliable. “Do the work” products succeed by learning from every task — user corrections, user preferences, vertical-specific edge cases. The product that captures more tasks gets better faster, and this compounding is what makes the category hard to compete with at scale.

5. The integration story changes. “Run the business” software usually IS the integration — it’s the platform that touches everything. “Do the work” software often starts by bolting onto an existing “run the business” platform (via Chrome extension, API, or scrappy RPA) and only later subsumes the platform’s functionality. The integration pattern is inverted.

Why incumbents struggle with the shift

Most vertical SaaS incumbents recognize that “do the work” products are emerging in their verticals. Many try to ship equivalent features. Most fail. The reasons are organizational, not technical:

Gold plating. The incumbent’s product culture expects feature completeness before launch. The “do the work” category rewards launching a 60%-complete product that feels magical at the 60% it covers. The incumbent’s product team treats this as irresponsible.

Sales incentives misaligned with PLG. The incumbent’s sales team is compensated on annual contracts and large deals. A $30/month self-serve product pays them almost nothing. They won’t promote it, and if they try, they’ll try to wrap it into a bigger deal that slows adoption.

Marketing’s demo funnel. The incumbent’s marketing team runs a demo-request funnel because that’s how they’ve always generated leads. A self-serve funnel with no demo is treated as “lower-quality pipeline” even when it produces better conversion.

Pricing anchor. The incumbent sells the platform at $500-$5000/month and can’t figure out how to price the new “do the work” product without cannibalizing their main business. They end up pricing the new product higher than the market will accept.

Brand confusion. Customers who think of the incumbent as “our practice management software” don’t naturally think of it as “the tool that drafts my patient notes.” The brand association is wrong, and rebuilding it is slow.

Org structure friction. The new product gets built inside the existing product org, where it competes for resources with “run the business” features. It loses most of those fights and ships slowly.

The result: incumbents ship belated, underpowered “do the work” features that lose to native AI competitors who built the whole company around the shift.

What “do the work” looks like across verticals

A partial taxonomy, just to show the shape of the opportunity:

In every row, the left column is a mature (often oligopolistic) market with established incumbents. The right column is either recently emerging or still unbuilt. The right column is where the next decade of vertical SaaS value creation lives.

The strategic implication

For founders, investors, and incumbent product leaders:

• If you’re building in vertical SaaS in 2026, ask whether you’re running the business or doing the work. If you can’t answer crisply, you’re probably building something confused.
• If you’re investing in vertical SaaS in 2026, the “do the work” category has the highest-quality companies at the earliest stage. Most run-the-business incumbents are in the path of structural compression.
• If you run an incumbent, accept that you have to ship a “do the work” product inside your own house before a native AI beats you to it. This requires organizational changes most incumbents resist. The ones who make the changes will survive the transition. The ones who don’t won’t.

Counter-argument — healthy incumbents can ship “do the work” products themselves

A live tension to hold with this note: Balaji Srinivasan argues in a 2026 interview that the “SaaS apocalypse” framing is overhyped because distribution is a separate asset from code. AI accelerates incumbents and disruptors equally, which means a well-distributed incumbent with a functioning engineering culture can absorb the “do the work” transition without losing their position. Figma shipping AI-native design features in-house is not fundamentally different from a native AI challenger doing the same thing — except Figma already has the users. See Distribution as the Remaining Moat - Why SaaS Incumbents Aren’t Dead.

The reconciliation: the strong form of this note (“all run-the-business incumbents lose to do-the-work natives”) is wrong. The weaker and more defensible form is: incumbents who cannot ship at AI-native tempo lose to do-the-work natives. The shift from run-the-business to do-the-work is real and important, but it affects complacent incumbents more than healthy ones. The diagnostic is engineering velocity, not product category — a run-the-business incumbent with fast shipping culture can ship their own do-the-work product and hold the position; a stagnant incumbent cannot.

The honest caveat

Not every “do the work” product displaces an incumbent. Some verticals have genuine regulatory or liability constraints that make a new entrant hard to displace. Some have powerful network effects on the “run the business” side. Some incumbents are themselves well-run and will ship equivalent features before the native AI reaches scale. The pattern is a default, not a law.

The strong form of the claim: across most verticals, the next decade of value creation in vertical software will come from products that do the work, not products that run the business. The weak form: the distinction matters, and companies that confuse the two will underperform companies that pick a side and commit.

Related Notes

• System of Action - Evolution Beyond System of Record
• Hero User Strategy - Native AI’s Wedge Into Vertical Software
• Seat-Based to Token-Based SaaS Pricing Transition
• Horizontal Platform to Vertical Specialization - Enterprise Credibility Pivot
• Grow 10 or Earn 40 - Two Paths for Mature Software Companies
• AI + Business Model
• The Race to Become the System of Action Tidemark