Work-Bench | From Augmentation to Automation: How AI Is Reshaping Work and Creating New Opportunities to Build

For better or worse, AI is quickly changing how we work.

This past year we’ve seen the mainstream adoption of AI code editors like Cursor and Windsurf along with the rise of revenue management provider Clay and note taking app Granola. Engineers are breezing through code changes and and sales teams are effortlessly personalizing outreach at scale. These AI-enabled tools are breaking out faster than ever, and driving real behavior change in how we work every day

As any product builder will tell you, changing user behavior is the hardest thing to accomplish, but it’s clear there’s a powerful current shifting the way people work.

Recently, Shopify CEO Tobi Lütke mandated that all employees integrate AI tools into their daily workflows, going so far as to tie performance reviews to AI adoption. It’s one of the first public signals from a major tech leader that AI isn’t optional anymore. The message is clear: companies that aren’t AI-forward will be left behind.

So, why does this matter?

Automation Readiness Checklist

I believe that over the next year we’ll see a dramatic shift from augmentation to automation.

Employees who rely on AI copilots today may soon find the scope of their roles shrinking, as companies shift more responsibility to context-aware agents. Many roles already operate with duct-taped AI workflows or experimental tools. It's only a matter of time before those functions are fully automated.

This shift also creates opportunity. There’s real potential to build venture-scale businesses by targeting roles that meet the following automation-readiness criteria:

Where Augmentation Becomes Automation

We’re already starting to see signs of full automation in roles that were once merely augmented by AI. Below are three areas where AI is moving from assistive to autonomous and changing outcomes in the process.

Insurance Claims

For decades, the insurance claims process has relied on a high volume of human coordination: file intake, damage assessment, policy verification, customer communication, and payout. It’s a deeply operational role, but one that has long been constrained by inconsistent documentation, subjective judgment, and bureaucratic lag.

For the past decade, scaleups have been quietly embedding AI into insurance workflows, while incumbents scramble to retrofit their platforms with similar capabilities. Scaleups like Tractable use computer vision to assess auto damage from smartphone photos while Shift Technology applies anomaly detection to flag potential fraud. Meanwhile, established players like Guidewire, Snapsheet, and CCC Intelligent Solutions are in a race to infuse AI across the entire claims lifecycle through partnering with startups or by building their own solutions in-house. Some challengers, like Lemonade, have gone even further, fully automating claims for select products and delivering payouts in minutes, entirely without human intervention.

Given how (relatively) easy it is to embed AI into products, insurance claims has emerged as the holy trinity of automation readiness:

Structured data flows (policy documents, claim forms, repair estimates)
High repetition at scale (millions of similar claims every year)
Clear outcomes (approve, deny, pay X)

As AI takes the bulk of claims processing, human-labor will be increasingly confined to edge cases: complex liability issues, litigation exposure, or emotionally sensitive claims (e.g., life insurance).

I’m searching for founders who are rethinking how technology can drive improved outcomes across the insurance stack

Deal Desk

While CPQ (configure, price quote) tools like Salesforce CPQ, DealHub, and Salesbricks help sales teams generate customer quotes, there is still no purpose-built software for deal desk professionals — one of the last functions in the Go-To-Market tech stack without a dedicated solution.

The deal desk is a central point of contact for the sales team to help optimize the contract process and is intended to expedite high-priority sales and push through red tape on behalf of the salesperson. The general rule of thumb is to have 1 deal desk person for every 30 sales reps. In a high velocity sales organization selling complex deals, deal desk can be an important and stressful role. People in this role can be inundated with emails, meetings, and AEs hounding them to push deals through.

However, deal desk professionals can be put in a tough position. Everyone knows that time kills deals, but rushing the process or bending procurement guidelines to accelerate closes can result in unfavorable terms and bigger problems down the line.

But what if there was an “always on” deal desk representative that could work across multiple deals, didn’t get bogged down by email backlogs, and was sure to always stick to your sales guidelines? With AI, teams could quickly upload and analyze historical contracts, price deals effectively, initiate discounts for high-value customers, commit automated redlines to contract negotiations, and more.

RevOps could be the first company function to have a fully autonomous business unit.

User Experience Research

User Research as a dedicated headcount function is rapidly declining across B2B and B2C companies. AI will be a key enabler for the future of product and user research.

While AI-native tools will contribute to the decline in traditional user researcher roles, they will simultaneously open new horizons to democratize insights across an organization.

AI can rapidly synthesize data, understand visual and contextual cues in human conversation, and generate sophisticated, targeted surveys, dramatically accelerating the time to meaningful insights, compressing research timelines from weeks to hours.

That means that in the near future, product managers, marketers, designers, and other key stakeholders will have access to insights directly at their fingertips without multi-week studies.

Today, companies like Qualtrics, Medallia, Sprig dominate the space, but often their panel recruitment leads to poor outcomes leading end-customers to resort to lengthy panel recruitment processes from expert networks likes of GLG, Guidepoint, and more.

We’ve already met companies in the space and are looking for differentiated viewpoints about how autonomous insights engines lead to real value for end-customers.

What’s next?

We’re moving from a world where AI helps you do your job, to one where AI does the job and humans supervise. As this shift accelerates, the most valuable companies won’t just adopt AI tools, they’ll rebuild their workflows around them.

The next generation of category-defining software won’t just support how we work, it will redefine it entirely.

If you’re building in business process automation and have ideas for how to reshape the future of work, I’d love to chat.