Jobs-to-be-Done in the Age of AI

When technology, competitors, and customer expectations change every quarter, your customer's job is the one target that holds still.

Technology, your competitive landscape, and your customers' expectations are changing faster than they ever have. If a customer doesn't see you as an AI-partner, forging the future, they will take their business elsewhere. A competitor that did not exist last year can ship an agent that does the work your software only organizes.

Every executive feels the urgency to change. Fewer are clear on what to change toward.

You are a venture investor now

With change comes risk. When a market moves this fast, no single product bet is safe, and standing still carries more risk than a failed experiment.

Whether you are a PE sponsor or a company leader, you are now in the position of a venture investor. A venture investor does not put the whole fund into one company. They spread capital across multiple shots on goal, expect some to miss, and size the portfolio so the winners more than cover the losses.

Operating companies now need the same discipline. That means funding more than one initiative at once: keeping the core product ahead, launching a new agentic product, testing a new pricing model. It also means deciding quickly which shots to back and which to stop.

A portfolio of bets only works if each bet has a clear target. Without one, you cannot tell an early result from noise, and every initiative drifts toward whatever the latest competitor announced.

The job is the stable target

This is why Jobs-to-be-Done matters more now than it did before AI.

Your customer's job-to-be-done is the goal they are trying to accomplish, independent of any product. Products change. Technology changes. The job does not. Kodak's film business ended, but people kept capturing memories, first with digital cameras and then with phones.

The needs within the job are just as stable. A customer need is an action and a variable within a step of the job. For a support leader whose job is to resolve customer issues, "determine the cause of the issue" and "confirm the fix worked" were needs ten years ago and will be needs ten years from now. What changes is how fast and how accurately a product helps the customer satisfy them.

That stability is what makes the job useful as a North Star. When you target the job:

• You measure progress in terms customers care about: how much faster and more accurately they get the job done.
• You see competitor weaknesses as specific needs they serve slowly or inaccurately, not as a feature list to match.
• You can leapfrog, because you are aiming at where customers struggle, not at where competitors are today.

Teams that chase competitor features stay behind.

In May 2023, Chegg's CEO told investors that since March, student interest in ChatGPT had been hurting new customer growth. Chegg's shares fell more than 40% that day (CNBC). The students' job had not changed. They still needed to understand course material and complete assignments correctly. A new product got that job done faster, and customers moved.

You no longer need a 3-6 month research project

The standard objection to Jobs-to-be-Done has been time. Mapping a job, interviewing customers, surveying for unmet needs, and analyzing competitors could take three to six months. In a market that shifts every quarter, that timeline is hard to justify.

That constraint no longer applies.

thrvAI: an agentic platform that does Jobs-to-be-Done for you

thrvAI is our new agentic platform that does Jobs-to-be-Done for you. It is the manifestation of the method we have been refining since 2013 with companies including Microsoft, Google, American Express, and Target.

thrvAI maps your customer's job: the job steps, customer needs, emotional needs, root causes, and job metrics. It identifies unmet needs, measures competitor speed and accuracy on each one, and proposes high-growth strategy hypotheses aimed at a segment that struggles with the job and is willing to pay for improvement. It keeps that analysis current by mapping customer interviews and sales calls back to the job.

Once thrvAI identifies the market opportunity, it builds the agentic platform to seize it. It builds an agent for each job step category (Plan, Execute, Assess, Revise, and Conclude), with an orchestrator that holds the job state so each agent knows what has been done, what comes next, and what the customer needs. Then it iterates on the product itself. The same customer evidence that keeps the job map current (interviews, sales calls, customer feedback, and usage data) tells thrvAI which needs the agents still serve slowly or inaccurately, so it knows how to improve their intelligence and what new work they should take on. It runs your evals on every change and ships the improvements, reads the trace when a run fails to find where it went wrong, and trains open-weight models for each job step on your executions and your experts' preferences. Each cycle helps customers satisfy their needs faster and more accurately than competitors do. Because the agents are priced against the economic value of the job rather than the number of seats, thrvAI tracks the cost of each job execution so the product delivers economic value to customers at a healthy margin.

The result is lower investment risk. Each shot on goal starts with a defined target, a way to measure progress against it, and evidence that updates as you learn.

thrvAI is the foundation of two new offerings.

thrvMCP: your customer's job, where you already work

thrvMCP gives Claude, ChatGPT, and other LLMs the context of your customer's job-to-be-done. The job map is generated automatically in minutes, available through MCP wherever your teams work, and continuously updated by customer evidence.

With thrvMCP and our JTBD skill files, your teams can:

• Generate product ideas that target specific unmet needs, then evaluate each on whether it gets the job done faster and more accurately
• Write marketing messages grounded in the struggle customers describe
• Draft sales scripts around the needs a prospect's current solution serves poorly
• Produce strategy and progress reports for stakeholders
• And every other product development, marketing, and sales activity gets better with the customer's job, competitor weaknesses, and customer evidence from thrvMCP

thrvMCP keeps your current product ahead of competitors by aiming each release at unmet needs in your customer's job instead of at the last feature a competitor shipped.

thrvMCP+ adds thrv JTBD experts: up to 12 customer interviews a month, ongoing competitor validation, continuous skills refinement, and a monthly JTBD training session.

AI Disruption Teams: build the agent that does your customer's job

Keeping the core product ahead is necessary, but it is not enough. An agent is going to do your customer's job for them and then steal your core product's revenue. You should build it.

Legacy software organizes the inputs to a customer's job and leaves the customer to do the work. An agentic product does the job. It is priced against the labor it replaces and economic outcomes it improves, not per seat. It improves daily against evals, not quarterly against a roadmap. It is a different product with a different revenue model and different business processes.

Clay Christensen found that companies survive disruption by building the new business with an autonomous team. IBM built its personal computer with a separate team in Boca Raton, away from the mainframe business. The reason is economic: the existing product funds the company, so when one team serves both, the existing product wins every argument about resources.

An AI Disruption Team is a small, isolated team (one engineer, one product manager, and one domain expert) that runs a new agentic business, with thrvAI as its engine. The work runs in four steps:

1. thrvAI identifies the market opportunity and builds v1 of the agentic product.
2. thrvAI iterates the product from there, using customer evidence, evals, and usage data to get the job done faster and more accurately with each release.
3. The team validates customer value and product-market fit through customer interviews.
4. The team takes the product to market and scales it into a new source of equity value.

Two shots on goal, one target

thrvMCP and AI Disruption Teams are two shots on goal aimed at the same target. One keeps your current product ahead. The other builds the product that could replace it before a competitor does. Both are measured the same way: how much faster and more accurately your customer gets the job done.

Tell us your customer, and we will show you the agentic market opportunity: the unmet needs, your competitors' weaknesses, and the segment to target first. Schedule a working session.