At Story99 we’ve been working with several AI native clients on their positioning (use-cases extend from marketing all the way to fund-raising).

Some clear patterns and understandings have emerged. This blog is an attempt to summarize some of that, hopefully helping founders in making the right positioning decisions about their offering / startup / product.

- Amrit, Founder, Story99.com

Story99.com is a Branding & Storytelling Consultancy for B2B, Tech & Deeptech

The Pre-AI Positioning Story

For roughly two decades, B2B software / SaaS [positioning ](/p/356011d7387b80b08eb0f2a9ce3d7b22?pvs=25)followed a script that worked because the underlying approach barely moved - you picked an ICP, found a wedge, built a tool that made one part of someone’s job faster, priced it per seat, and expanded inside the account as usage grew.

Whatever you were selling - the feature set, the workflow, the integration - was “durable”. A CRM in 2015 looked a lot like a CRM in 2020. And that’s why the SaaS land-and-expand script worked.

That assumption no longer holds for AI native solutions.

The Story Now

The ‘durability’ logic that held true earlier, doesn’t apply any more.

Today, the capability sitting underneath an AI product is moving on a curve that can outpace the roadmap, the pricing model, and sometimes the entire category it was built for.

The feature that justified the seat price earlier can now simply be absorbed into the next model release for free. And therefore, trying to position an AI-enabled product / solution as “software with an AI feature bolted on” is unlikely to work. We are noticing this for every client that we work with.

In this new context, below is how an AI native startup can approach fine-tuning their positioning

1. Sell the work, not the tool

The clearest dividing line in AI GTM right now is between products that make an employee work faster vs products that do the job directly (with or without human oversight).

The first kind sells a tool to the person doing the work. But it’s the person that’s accountable for the output. Here, the tool earns its budget by essentially saving them time (or by offering some quality advantage over manual - like higher accuracy etc).

The second kind sells the outcome itself to whoever needs it done, and the buyer stops thinking about headcount or hours entirely. Here you are taking ownership of the job (a job = a set of tasks).

This matters for positioning because it determines what you’re actually competing with.

• A tool is competing with the next tool - and with whatever the underlying model can now do for free

• But an outcome is competing with however that outcome currently gets produced, whether that’s an internal team, an outsourced vendor, or nothing at all (where the work doesn’t get done.

What we are seeing is that the second comparison is far more favorable. It changes the entire pitch: instead of “here’s how much faster or accurate you’ll be,” the pitch is now: “here’s the result, handled.” And that’s usually followed by an explanation of why it makes sense to get the work done through your offering, from money, quality, speed, reliability or forward-thinking perspectives (use of AI agents in the workflow is the standard explanation).

How far you can push toward selling the outcome, depends on how much of the work is pattern-based (high %age of standard SOP) vs judgment-based.

The more a job looks like applying complex but knowable rules (eg: coding a claim, drafting a routine contract, reconciling a ledger), the sooner you can credibly take ownership of the result.

The more it depends on taste, relationships, or accountability that can’t be transferred, the longer you’ll need to sell the tool first → earn your way toward the outcome.

2. Translate capability into the language of the operator, not the language of AI

The buyer - especially in a traditional or mid-market business (the practice owner, the operator running three locations, the CEO who’s really a clinician with a P&L etc) does not think in terms of agents, orchestration, or “outcome pricing.” They think in terms of cash flow, staffing, and whether they can take on one more location without burning out the people they already trust.

And that’s why your AI-native customer facing pitch needs to survive translation into that language. Their language.

Instead of saying “Our agent automates claims adjudication”, try “this gets your AR moving ten days faster.”

“We deploy a multi-step reasoning pipeline” → “this is the equivalent of one more person on your billing team, without the hire.”

You get the point. Basically your pitch to customer should not be describing the technology but the result (tech or AI just answers the ‘but how’ question).

This also means being honest about scope.

For example, a real operation might involve well over a hundred small handoffs between a customer walking in and money landing in the bank. So an AI product that automates a handful of those steps and gestures at “outcomes” for the rest is selling just a roadmap. And that is not the same as selling result. And so you will struggle to sell anything at all.

It will pay to be precise about which parts of the chain you can actually move and let that precision do the selling.

3. Make data ownership part of the value-prop, not a footnote

For buyers who’ve watched a parade of vendors promise “transformation” and deliver another monthly invoice, “your data trains our model” is the thing that makes them suspicious of every AI pitch they hear. Many of them have already noticed that the tools they’ve licensed know everything about their operation. But they own none of it.

You can flip this into a selling point. For example say if your product is building the customer their own data asset (could be a structured record of how their business actually runs, or models trained on their own operational history or something that has value on its own balance sheet), never forget to say so explicitly and early. The prospect will appreciate it very much.

For a traditional business, “we’re helping you become a more valuable, more efficient company, and you keep what we build” is a fundamentally different proposition than “we’re helping you use AI.” It also happens to be a durable moat - a customer who owns their own context has less reason to leave. Also, a competitor can’t simply replicate it by shipping a better model.

4. Price for alignment (not for access)

Per-seat and per-license pricing is a tool-positioning signal, and it carries a tool-positioning risk: your revenue is tied to how many people are using your product, at the exact moment AI is reducing how many people are needed for a given amount of work. See the problem?

It also ties your fate to your own roadmap in a race against frontier labs who can erase your differentiation in a single release.

Outcome-based pricing: a fee per claim processed, a percentage of cash collected, a share of the cost saved - points in the right direction.

A pricing model like this scales with the value produced rather than with seats, which matches how AI-driven work actually scales today: a single workflow can run unattended for hours and touch far more volume than any human team using the old tool ever could.

It also changes the relationship between you and model improvements. If a cheaper or faster underlying model shows up next quarter, a per-seat vendor has to worry about being undercut or commoditized. But a vendor priced on outcomes just gets more efficient at delivering the same result - the improvement becomes margin. That’s a meaningfully different story to tell a buyer who’s been burned by “the next version will be even better” promises before.

Just make sure the outcome is defined really really well.

5. Anchor on a budget line that already exists

If a piece of work is already outsourced (say to a services vendor), then all that you are asking the prospect is to replace them with you.

Your pitch here is - hey there, we can get the same job done better (typically in terms of speed and / or accuracy) at either the same or a lower price today and non-linear increase in price with scale tomorrow (where SaaS / traditional services typically struggle).

This “head to head” positioning leads to a vendor-swap conversation. And you can figure out how to then take that to conversion. I am not saying it’s easy to change the status quo, but if you can find something that’s broken & painful in the existing non-AI vendor delivery (like say linearly increasing cost or slow service or something that the customer really wants to be fixed), you definitely have a good chance to get a contract. At least for a PoC. Then show the result and win the rest.

In contrast, positioning against work that’s currently done in-house (with or without a Saas) is essentially a reorg conversation. And we are seeing that reorg conversations are going slow. They are political and not something easily won by outside pitches (unless you are friends with the founder or they seek your advice).

It can work, but only when the org is ready.

You also limit the folks you can sell to (because for most leaders, this pitch means letting go off their people and shrinking their teams; unless that’s an immediate pressure on them, you are not going to convert).

Net, net, in our opinion it makes more sense to lead your pitch with the part of the workflow that’s already leaving the building: the outsourced billing, the third-party adjuster, the contractor handling overflow IT tickets. Win that first. Find your wedge. Expand thereafter. And since sooner or later, every task in a workflow that can be automated, will be automated, all you gotta do is to be there inside the system to take over that workflow and automate-max it when the time comes.

PS: When selling to SMB’s, at least for one of our clients, we’ve seen that selling a services solution that never existed earlier at a price point that they are willing to spend - is moving the needle really well (this is neither a vendor-swap pitch nor a reorg one; it’s a “finally a service that delivers results, at a price point that you can afford” pitch). If you have something like that to offer, put all your efforts towards marketing, selling and scaling that.

6. Build distribution around people who’ve done the job

What we are noting is that generic AI sales motions (SDRs reciting capability lists, top-of-funnel content about “transformation”) tend to bounce off operators who’ve heard it all before.

But what is landing more is credibility from someone who has actually run the function being sold: the person who’s done five acquisitions in the same state, who knows why a claim gets stuck on day forty-five, who can sit in the room and translate between what the model can do and what the practice actually needs.

This is also why services and systems-integrator partnerships are showing up much earlier in AI GTM than they ever did in SaaS.

When the product itself is partly “the work done correctly,” the distribution channel and the product essentially blur together.

Thus build your GTM around domain-fluent operators. Don’t treat it like your support function, but your front line. This approach will outperform a predominantly product-led motion (more so in mid-size markets where the buyer doesn’t even have a CIO or a consultant to guide them).

Lastly, Some suggestions on how to review your current positioning

Just ask yourself the following questions and see what your answer is:

• If the underlying model got dramatically better tomorrow, would that strengthen the pitch or make it irrelevant? If the answer is “irrelevant,” the positioning is built on the tool, not the outcome. Fix it.

• Could the pitch be explained to the buyer in terms of cash, headcount, or capacity without using the word “AI” at all? If it can’t, it’s probably still describing technology rather than a result.

• Does the pricing model reward the company for the same things it rewards the customer for? If the vendor wins by adding seats while the customer wins by needing fewer people, that tension will eventually surface - usually at renewal.

• Can a prospect point to an existing line item - a vendor, a contractor, an outsourced function - that this would replace? If not, the sales cycle is fighting for new budget and a reorg decision at the same time, which as we said earlier, we believe is a relatively much harder and slower sale.

In a market where the underlying capability is a moving target, positioning that depends on what the product currently does is not a great idea. But positioning that depends on whose outcome it’s responsible for, whose data it builds, and how its incentives are aligned is much more durable, and likely to help you win in the market.

Happy building & scaling your AI company, and may the best positioned product win! 🙂 All the best!

- Amrit, Founder, Story99.com

amrit@story99.com

PS: Feel free to reach out to me for consulting on all things storytelling & communication (use Calendly to block a time).

References:

• Services: The New Software - Julien Bek (Partner, Sequoia)

• Moative Essay: The AI Trap Door - Ashwin Ramasamy (Co-founder & CEO, Moative)

• Who will set the price / intelligence? - Jaya Gupta (Partner, Foundation Capital)

Story99.com is a Branding & Storytelling Consultancy for B2B, Tech & Deeptech