The Change Economy

Evolving how you run a software company when your software evolves at warp speed

We're witnessing what I call the "Change Economy" - an unprecedented era where software companies can improve their products 10X faster than ever before. This shift fundamentally transforms how software businesses must operate to deliver value and compete successfully.

AI adoption is outpacing every previous technological revolution in history. It took about 20 years for 60% of Americans to adopt electricity, nearly 50 years for telephone penetration to reach the same level, and about 10 years for the internet. AI is accelerating even faster - with 39% of the US population becoming monthly users within just two years of launch.

The implications for software development are staggering:

• Products that once required massive engineering teams can be built by small, nimble teams with 10-50X less effort

• For some startups, 80%+ of their code is now LLM-generated

• Integration issues that plagued software development for decades are disappearing overnight with emergence of tools like anthropic mcp

• The traditional "Agile" product/design/dev complex is struggling to adapt to rapid iteration cycles compressing weeks of work into hours

When someone like Sean Ren (an ex-PM) can envision a feature, ask Claude to plan it, write the code, fix syntax errors, and deploy - all in about an hour instead of weeks - we're experiencing a paradigm shift that affects every aspect of building and selling software.

The New Success Formula: Speed × Insight × Talent

In this brave new world, the formula for startup success has been rewritten:

Speed of Execution × Level of Insightfulness × Density of High-Quality Talent

Technical differentiation alone isn't sustainable when your competitors can replicate your features in weeks, not months or years. The right to play is understanding customer needs as well as or better than your competition. The right to win comes from executing faster and more insightfully to tastefully solve those needs.

This isn't just about coding faster - it's about reimagining how software businesses operate when their product's value proposition changes impactfully faster than annually. Today, your product and your competitors' products likely change meaningfully enough to matter every month - and if they don't, you'll likely be roadkill.

However, most SaaS companies were built for a world where significant product evolutions happened once or twice a year. As a result, many of your functional heads probably have that “head on a swivel” look right now as they are wrestling with how best to operate in this new world - a world in which you do not have ample time to easily plan, organize and deliver without overwhelming the “normal” rate of human comprehension. They are wrestling with which best practices to bring over and scale from the pre AI SaaS playbook and which to discard and reimagine. Founders similarly are wrestling with whether to bring in SaaS scale professionals or first principles as so much of how to build world class functions requires a rethink.

Let's dig into six areas that I believe are forever changed in this new era. I believe those founders, execs, and teams that master this transformation the best will see outsized returns vis a vis their competitors.

Six Areas Forever Changed by AI's Pace

1. Product Ideation, Design, and Delivery

Traditional product management devoted weeks or months ensuring requirements were correct before coding began. Development resources were precious, so meticulous planning was essential. Because development resources were so scarce and costly, teams would spend weeks or often months making sure their product specs or PRDs were correct. This involved gathering information across multiple sources, conducting targeted user research, and holding numerous design sessions before developers actually started writing code.

The concept of Agile development emerged as a way to do this more efficiently, breaking work into sprints and ensuring development could be handled at atomic levels of assignment. Yet even with Agile, the process remained labor-intensive and time-consuming.

Now everything has changed:

• Product teams leverage AI to gather, synthesize, and prioritize what needs to be built in hours rather than weeks

• Product ideators can use off-the-shelf AI to quickly produce good enough quality software to immediately test customer feedback

• These prototypes can be handed off to developers who use tools like Cursor to dramatically reduce their own coding time

• The amount of code that needs to be written has decreased significantly—sometimes by 80% or more

• The technical capabilities released by the LLM providers evolve weekly, enabling solutions previously thought impossible

Importantly, the nature of technical innovation has fundamentally shifted. Over the last decade of building SaaS apps, no one was relying on technical breakthroughs to deliver the end software applications. The code or how to do something was relatively known—it was more a matter of prioritizing your development resources versus the work to be done.

But now, the underlying technology is changing so rapidly that we see leapfrog capabilities almost weekly from all the major AI infrastructure and model providers. This means beyond the obvious items you know must be built, R&D teams are now brainstorming with customers on new needs and experimenting to see if they can solve them with the rapidly improving stack.

I love seeing the Cheshire cat grins on developers' faces when they show me something they didn't know was technically feasible just a week ago. But this creates massive challenges for how R&D works:

• How do you balance rapid innovation with customer-requested improvements?

• What happens to roadmaps when you don't know if a solution is technically feasible until you try?

• How do you manage your backlog of small customer requests against new capabilities?

• What's the point of T-shirt sizing in Jira or Linear for situations where you don't know if something is technically possible?

• What the hell does generally available and alpha or beta releases mean?

• How do you maintain order without stifling innovation?

I don’t know what the next generation terminology will be used to describe the right way to do things - it’s not waterfall development (the on premise era) and it’s not agile development (the saas era). I don’t know but the most effective teams that I am working with are adapting to this new reality by:

• Using AI to accelerate customer research and feedback cycles

• Embracing "spike" periods for technical exploration

• Implementing monthly and often weekly release cadences rather than quarterly

• Reducing the emphasis on long-term roadmaps in favor of rapid iteration

• Reimagining product management as synthesizers rather than spec-writers

• Maintaining smaller, more senior development teams with greater autonomy

• Rethinking the role of design and UX in a world with less screen real estate but higher importance of "taste"

Interestingly, many founders are successfully using the 30 to 60 day out clause in their prospect engagements. This was previously a complete no-no in the SaaS era but it has proven, in the case of a true WTF product, to reduce sales cycles down to a few calls. The caveat when deploying this is ensuring the scope of the problem you are solving for your customer is one that you are confident works well (with some minor tweaks once you get the customers data) or else your engineering team will never get an hour of sleep. During the out, your R&D team engages with the customer to ensure they can successfully accomplish the desired jobs to be done. Technically, you are making code changes / fixes “in the field” and as such GA is when the customer says “Oh yes, I can not live without this and will not exercise my contractual out”. I don’t think this approach works for companies at scale in any fashion but its a scrappy way for start-ups to cut through all the bs.

2. Product Messaging and Value Proposition

The underlying heart of an amazing software company that crushes its competition is the combination of a great product and the organization's ability to communicate at scale their unique value proposition. The latter usually is driven by some world-class product marketers—of which our industry has too few.

World-class software companies have always excelled at articulating:

• Who For - identifying their ICP/Persona, i.e. the exact types of people at specific types of companies for whom their offering is most meaningful

• Why This - explaining how their product solves a key pain or need for the customer

• Why Us - differentiating from competitors as to how they solve that need the best, especially vis a vis customer’s incumbent solution

• Why Now - convincing the ICP why they should move heaven and earth to buy their solution immediately or risk falling behind peers or failing to achieve key outcomes

In these companies, this messaging value tree is embedded across the entire go-to-market motion—on their website, in outbound emails, sales discovery, competitive kill sheets, their sales process and methodology, in case studies, in discussions with analysts, etc. It's rare that companies get this right, but when they do, they tend to be market leaders.

Getting this right is hard but doable. And AI has been helpful in making this easier—both in crafting content better/faster and especially in synthesizing what's working versus not in practice (i.e., across all your recorded Gong calls). AI also helps if you actually sell an agentic product that prospects can clearly see and touch— in that case, one can argue that less product marketing "explaining" is needed as the user can immediately grok the value.

But this approach works well when your product has regularly scheduled release cycles and your core value proposition changes once or maybe twice a year. What happens now when your product and its value prop—and that of your competitors—changes 6-12x a year? How does this group/team/set of resources (that is usually underfunded) stay on top of all this change and, more importantly, ensure it's propagated across all your relevant content/messaging/assets?

I personally think this is a massive hurdle that most organizations will struggle to master. The bible on this, like all of these areas, has yet to be written but I think the solution requires:

• Creating frameworks that allow for rapid adaptation without constant rebuilding. In my view, the jobs to be done framework is a great one especially in the agentic world - ie what does your products assist a human in doing better and/or what does products do that no longer requires a human.

• Moving messaging from static documents (slides or Word docs in Google Drive) into systems of record with real metadata that leverage AI to easily assemble, adjust, update, and propagate messaging at scale across all touch points (like www.octavehq.com)

• Grouping innovations and your messaging updates monthly at a minimum to ensure that humans (both on your team and at your customers) can comprehend and convey it effectively

• Investing disproportionately in product marketing talent—in this new world, the value of effective product marketing becomes 5x more important than before AI (and even then, the winners had great product marketing). I have met the product marketers at companies like Anthropic and their competitors - these folks are the best of the best in the industry.

3. GTM Sales Resources Readiness

Let’s say you are able to update your messaging at the pace of your product innovation - how do you actually make sure that your sales resources—whether humans and or agents—understand the latest information about the jobs your product does better and why its urgent prospects get on board?

All of the existing ways we train and enable sales teams haven't fundamentally changed in 20+ years. New hire bootcamps, stand-and-deliver deck and demo certifications, monthly/quarterly update training webinars, sales content management systems—all were predicated on two concepts:

1. Your product's core value proposition doesn't change that much nor does that of your competition

2. The focus, especially for new hires, is getting them to understand your buyer and your playbooks

Traditional sales enablement was designed for a world of relatively stable products where sellers needed to master buyer understanding and your specific sales playbooks more than constantly shifting product capabilities.

But when your and your competitor’s product value prop changes 6-12 times a years, how is a PowerPoint deck in Seismic or some LMS course going to help? By the time an admin or content creator finishes the content for someone to use or a quiz to take, the content is no longer relevant. In this new world, everboarding is 5x more important than the initial on-boarding of new GTM hires.

With value propositions changing monthly, how do you keep your sales teams current? On the good news front, if your product is awesome and leverages AI to deliver WTF moments to prospects (“does that really work? wow!”), selling is far easier. You simply don’t need to deploy many of the old traditional sales tricks because your product no longer requires your prospects to do so much work to understand / appreciate the value your offering delivers. Unfortunately, that advantage will not last for long as this will be the new standard for any competitive offerings in your respective marketplace. Therefore, here are my preliminary recommendations for GTM organizations now:

Enablement is no longer a function to be ignored or hired at some point in the future when our company is big. You need a kick ass enablement person and I likely would have them work under product marketing given the importance of ensuring your GTM teams understand all the rapid iteration of messaging and differentiation coming their way.

For your enablement approaches, focus as much time on the everboarding plan as you do on the onboarding (i.e. the first 90 days). It sounds crazy but I would take an entire day off each month (if that’s your frequency of innovation) to train and refresh your sales teams.

Re-examine your existing approaches and tools and understand if they simply will not work in this new era. Look at heavily AI centric enablement tools which have been designed from the get go to reflect and serve this new frenetic pace of change. For example:

Traditional content management systems (CMS) were designed for a time when products changed slowly, and answers could be found parsing through static documents. Evaluate and select AI-native enablement platforms (like www.spekit.com) that make it dead simple to automatically update and instantly propagate new information wherever your sales professionals are working - ensuring they always have the latest answers or deal-specific content recommendations to effectively communicate with your buyers

Instead of a traditional LMS platform where the content will be outdated quickly and don’t drive applied learning (and therefore it is less likely to stick), instead look for roleplaying agents who play the role of your customers/prospects that your reps can meet and engage with an unlimited amount and receive instant feedback / grading on theory performance

Finally, competitive listening and intelligence is more important than ever - your value proposition matters relative to your competitor’s frame. Now is the time to likely lean into an AI native CI tool or set of agents that automatically collect and curate competitor insights, win / loss analysis you can trust (i.e no rep is ever going to mark down “I was outsold” for a loss reason) and equips sales with constantly updated battlecards.

4. Customer Education and Adoption

The concepts of customer implementation and onboarding were all rooted in the assumption of minimal changes in product value proposition over time. This was often managed by humans and/or in-product checklists that guided users through setup and initial use.

On the positive front, in theory, onboarding as we have known it should largely disappear for agentic systems that actually work—either for a user or as a replacement for a worker. In those cases, it's really a matter of activation and validation: the customer simply turns it on and confirms whether it's performing the job correctly. That being said, that does not mean onboarding goes away - I actually think it becomes a much more strategic unlock for product and success teams. The new wave of great agentic software is unlocking huge productivity gains in individuals and across organizations IF they understand how best to get the most value out of it super quickly after first using it. Not everyone can do “superhuman” levels of onboarding but the key unlock here is for your product to become an addictive habit quickly for users across an org or else I think you are dead in the water.

After onboarding, however, how do you ensure your customers are aware and take advantage of all of the 10x rapid innovation you have released since they were first “onboarded”? In this new environment, we need entirely new approaches. In my view, \two things become paramount for companies on the education front:

1. Regular intervals of updates and education—ideally monthly. Anything faster than that is too hard for customers to process. I know that there will be many that scoff at me and say “but Brett, we can ship every 9 hours”. That’s awesome but none of your customers will be able to track that rate of change nor be able to just consume that level of innovation - unless our customer replaces its users of its software with agents - then by all means, knock yourself out.

2. In-product/in-agent notifications and guidance—instead of sending an email and hosting a webinar on what's new, the agent in the product should notify users that it can now do more for them and offer to demonstrate how. This same in-product/agent guidance should serve as the entry point for any "how do I" questions. Do NOT send them to a separate FAQ—just give users the ability to ask the agent questions directly, and the agent should show them how to do something or just do part of it for them.

We must also rethink how we measure and track adoption. Most SaaS companies track usage metrics like MAU, DAU/MAU, etc.—which still matter. If customers aren't using your product, they've moved on, and you're done. Beyond usage, it has always been hard for even the best success teams to understand if their customers are achieving the business impacts they desired when they bought the product. On the positive side, when your agentic software is actually doing jobs for the user or company, it's much easier to measure how many jobs it's doing and if they're achieving the desired outputs/performance. What's different now within the Change Economy is that customers will evaluate you not just on how well you do the job they bought you for, but how you do that job better over time as well as the additional jobs you can do for them vis-à-vis competitive alternatives.

As such, the focus on tracking real adoption beyond basic usage stats should center around the usage of each job-to-be-done that your product assists or solves. If the customer uses very few of them, they're likely at risk.

I believe the best approach is primarily in-product—yes, QBRs will still matter, as will other traditional SaaS success notification techniques—but your product should engage users regularly about what more it can do for them, including providing visual summaries of which jobs they're utilizing and which they haven't yet tried.

This is critical because AI is new to people, and leveraging these systems effectively often requires understanding what prompts to use. In this combination of agentic interfaces and rapid product innovation, you need to suggest "prompts" or jobs-to-be-done.

Getting this right is essential for maintaining strong net revenue retention. In the SaaS world, so much of a company's value is tied to recurring revenue, which requires less GTM spend to maintain. In this new era, where software is incredibly easy to build and deploy, switching costs are far lower, and the decision process is clearer: is this product doing the maximum number of jobs AND doing them well AND not more expensive than alternatives? Unless you rethink education and adoption along these new lines, achieving even respectable NRR will be challenging.

5. Pricing and Packaging

Most software companies struggle to get pricing and packaging right - i.e. properly reflecting the right balance of minimizing customer confusion/friction and ensuring you get paid more where there is more value delivered to the customer. That being said, in a world where you ship real innovation rarely, most companies have time to think through the traditional discussion of "is this a feature, is this a module, or a brand new product line" and then make the proper adjustments to good/better/best packaging.

The age of AI has created two big "hmm what do I do now?" shocks to the system:

• How do you charge for AI specifically? Is it seat-based, work-based, deliverable-based? This is especially important if your offering allows companies to reduce existing or future headcount—in such cases, you should try to secure dollars from payroll spend rather than the software/tools budget. I won't delve too deeply on this as 1) I don't have any unique insights here yet and 2) there is much written on this topic already.

• How do you price capabilities you can deliver in weeks? I'm seeing companies gather a client use case and deliver a solution within a few weeks. When the customer asks what it costs, most founders say, "that's a great question..." This is the bigger challenge that requires immediate attention.

As such, we really do need to rethink and reframe how we normally would approach pricing and packaging adjustments and the frequency of said changes.

In my mind, the likely easiest way to start that framing, especially for agentic software (assistants or full-on agents), is around the jobs-to-be-done they perform for the customer:

• An improvement to how you do a job to be done for a given persona/function is a feature and is included

• A new job to be done for a given person/function triggers a decision point around whether or not you charge additionally for that (partially tied to how valuable that job to be done is for the customer)

• A new job to be done for an additional persona/function type is something that should definitely be an additional charge

Regardless of the framework you choose for deciding what to charge for and what to include, a really important decision is determining the acceptable frequency of changes to your pricing and packaging, especially for existing customers.

In theory, with all of your product innovation, you could be making adjustments each and every month—which may be great for you but likely frustrating and confusing to customers who may feel they're getting nickeled and dimed. It's important to align your cadence of product releases with your messaging updates, sales and customer training, and pricing/packaging updates.

In the pre-AI world, pricing and packaging updates should never be done after new functionality is in the customer's hands and they're using it—no one likes being surprised. But now that you can deliver new capabilities into customers' hands monthly or faster, what do you do about pricing and packaging?

It's a dilemma to wrestle through. I recommend being intentional internally and externally and classify your monthly deliverables before you release them into:

• Included for free (the feature released helps the same role do the same job more effectively)

• Potential Additional charge (the feature released helps the same role do a new job). If you do decide to charge, consider a grace period for existing customers with clarity on what it may cost at a future point if they elect to keep using it

• Definite Additional charge (the feature released helps a new role do a new job). This is clearly something customers should elect to review and purchase (i.e., an add-on sales process) before it's ever activated or turned on.

This framework won't solve every situation, but it provides a structure for thinking about the problem in a world where your product's capabilities might double or triple in a year.

6. Do Not Ever Buy or Sell Roadmap - Or Maybe Somewhat?

For the longest time, I've been maniacal with the teams I've run or the teams I advise that selling roadmap is so dangerous—it mis-sets customer expectations, likely leads to garbage deployments and unhappy customers, and causes a fundamental loss in trust—which is often one of the only moats you have in a software company.

Similarly, when buying products, I want to see what vendors have now first and foremost as that is what I'll deploy. I often ask to see roadmap presentations to get a sense of how the vendor thinks about innovation because you're hopefully investing in a long-term relationship. But I never buy based on future promises.

As an investor, I hate when early-stage startups present their CARR (Contracted Annual Recurring Revenue) number as a top-line metric—it scares me because it means their product is not deployed and the customer is not paying for it yet. This is usually a recipe for disaster for so many reasons.

But do I feel as maniacal about this today as I have for the last 25 years? We're now in a world where:

• Software companies can deliver far more innovation every month

• Software companies are encouraged to land-grab additional jobs to be done much earlier

• We often don't even know technically if we can solve for a given customer pain until we try to tackle it

• As a buyer, my renewal decision will likely be tied not just to how well the product solves my current job but how much better it does that over time PLUS what new jobs it can solve

So the answer—I feel less dogmatic now, with important caveats:

1. Customer trust is more important than ever

2. If you hand me an agentic solution and it doesn't do what I need or what you told me it would do, I can discover that much much faster than with plain old saas.

The best approach? Be careful if you sell a bit of roadmap while recognizing these truths:

• Prospects will evaluate you on your pace of innovation up to this point (so always helpful to explain all the value you've shipped in the last 3-6 months).

• Customers will evaluate you on your pace of innovation since they signed with you as a strong consideration in their renewal decisions.

• It's never a good idea to have your success and development teams completely under the gun to fulfill previously mis-set expectations.

Embrace and Adapt or Get on the Struggle Bus

This is without a doubt the most exciting time to be involved in software in my 30 years - as an investor, builder, or customer. The pace of innovation in delivering amazingly helpful software is unprecedented and will only accelerate.

Those organizations that transform their processes to keep pace with this innovation will persevere and prosper. The others will find themselves increasingly irrelevant as the Change Economy rewards those who can not only build faster but adapt every aspect of their organization to thrive at this new, exhilarating velocity.

The Change Economy is here, and it's getting spicy out there. I'm all for it as it will very quickly separate the wheat from the chaff across the industry.


Note: For full disclosure, Spekit and Octave are Bonfire investments.