> ## Content Index
> Fetch the complete content index at: https://www.maizeandmachine.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Honor's Approach to AI
- URL: https://www.maizeandmachine.com/honors-approach-to-ai/
- Published: 2026-09-08T04:27:59.000Z
- Updated: 2026-09-08T04:27:59.000Z
- Author: Rounak Banik
- Tags: Case Studies & Deep Dives, Generative AI, Strategy & Thought Leadership

**Disclosure:* The author of this article worked with Honor as an MBA intern between June and August 2026.*

Honor President Ian Clarkson has a simple framework for separating companies genuinely building with AI from those merely talking about it. It is useful whether you are deciding where to work or figuring out what your own company should be doing.

[Honor](https://www.honorcare.com/?ref=maizeandmachine.com) is a technology company focused on helping older adults age safely and independently at home. Through Home Instead, it combines a global home-care network with technology designed to improve care delivery.

As President, Ian oversees core operations and helps shape long-term strategy across Care, Growth and Product—putting his role at the intersection of technology and operations.

![](https://storage.ghost.io/c/53/db/53db4e13-71bc-4eb9-9864-740af9b355b3/content/images/2026/09/image.png)

Ian Clarkson, President at Honor

In 2000, Ian graduated with an MBA from the University of Michigan and joined Amazon, when betting your career on an internet company was far less obvious than it is today. He stayed through the dot-com crash and spent the next 15 years watching the internet transform from an emerging technology into infrastructure underlying huge parts of the economy.

When I met him during an orientation session in Seattle, he described a strange sense of déjà vu. To him, AI is where the internet was roughly 25 years ago.

There will be companies that use this technology to fundamentally change how they operate. And there will be companies that mostly watch, jump on a few hype trains and die.

The problem is that almost every organization claims to belong to the first group. Everyone has an AI strategy, a chatbot or an agent experiment. [Every earnings call seems to mention AI](https://insight.factset.com/highest-number-of-sp-500-earnings-calls-citing-ai-over-the-past-10-years-2?ref=maizeandmachine.com) in some shape or form.

So how do you separate the signal from the noise?

Ian uses six questions. They are not really about models or project counts. They ask whether AI has changed work, whether experiments become production systems, whether leadership has a coherent view of the future and whether the business remains valuable as the technology gets better.

## I. Are they actually using AI?

This question sounds insultingly obvious, but it isn’t. There is a distinction between having access to AI and actually changing the way work gets done because of it.

A company can give every employee access to an LLM, run a dozen hackathons, announce an innovation lab and still operate exactly as it did three years ago.

Ian’s first test is simple: look at what has changed. Have workflows changed? Products? The way people make decisions?

The strongest test is simpler still: *if AI disappeared tomorrow, would the organization become noticeably worse at doing its job?*

If the answer is no, then for now, it is probably still more talk than transformation.

## II. Can they actually ship it?

Using AI well is one hurdle. Getting it reliably into production is a different problem entirely.

A prototype can look impressive in a controlled environment. Production software has to deal with real users, messy data, legacy systems, security requirements, legal compliance, edge cases and unpredictable model behavior.

So even if an organization has obvious use cases for AI, ask: *can it repeatedly turn controlled experiments into things people actually use?*

That requires more than a good model. Teams need the right data, legal and security need to move quickly enough, prototypes must work with existing systems and impact must show up in metrics that matter.

The winners will be the companies that build the organizational muscle to turn prototypes into reliable products and workflows, again and again.

## III. Does leadership have a real point of view?

The next question is for leadership: *do you have a specific, opinionated view of how the future will change, or are you simply chasing whatever happens to be fashionable?*

If your company swore by chatbots and copilots in 2023, invested in RAG systems in 2024 and is now championing agents, that isn’t necessarily a problem. Technologies change, and good companies should change with them.

The problem is when the technology itself becomes the strategy.

What is rarer than another AI strategy is a leadership team with a clear thesis about how AI will change its industry, what becomes more valuable as a result and where the company should place its bets.

That thesis may be wrong. But at least it can be pressure-tested, debated and refined. If it is right, the company has probably been building toward that future while everyone else was deciding which shiny new tool to adopt.

## IV. Does the company still exist if AI is taken to its extreme?

Many companies begin conversations about AI by assessing what the technology can do today.

Ian asks you to do the opposite.

Assume AI keeps getting dramatically better and today’s limitations disappear. Assume it can automate everything that can plausibly be automated. *Does your company still need to exist?*

If the answer is no, you have a considerably bigger strategy problem than deciding which LLM provider to use.

If you’re a B2B SaaS company whose moat comes from software that was historically difficult or expensive to create, increasingly capable systems like Claude Code should probably make you nervous.

But if your business depends on something AI cannot conjure into existence: physical presence, trust, human relationships or real-world service delivery, the threat looks different.

Honor is an example. Even if AI becomes extraordinarily capable, people will still age and many will still need another human being physically present to care for them. Leadership has pressure tested more extreme possibilities too, including a future in which robotics takes on substantially more caregiving work.

The point isn’t that Honor knows exactly what that future looks like. It’s that leadership is asking whether the company’s core value proposition survives it.

## V. Do these leaders work the way you want to work?

At first glance, this question has nothing to do with AI. Which is precisely why it belongs here.

Technology transformations become organizational transformations. They force leaders to make decisions with incomplete information, tolerate experimentation and reconsider where humans should remain in control.

Ian’s advice is simple: watch how leaders communicate, make decisions and treat people. AI makes those traits harder to hide. A company cannot tell employees to experiment while punishing failure, claim speed is essential while requiring six months of approvals or build an AI-native organization while senior leadership delegates understanding AI entirely to the technology team.

For executives, the implication runs the other way: *your behavior is part of your AI strategy.*

## VI. Do you believe in the vision and does the company live it?

The final question has two parts: does the mission matter to you, and can you see it reflected in the work?

Honor’s stated mission is to change how the world cares for its aging population. That creates a useful constraint: AI is valuable when it helps Honor deliver better care, expand its capacity to do so or make the system around care work better, not simply because something can be automated.

This sounds soft compared with models, agents and infrastructure. It isn't.

When dozens of plausible AI use cases compete for limited engineering capacity, a clear mission helps answer a practical question: *which problems are actually worth solving?*

## How Honor Approaches AI

So how does Honor fare against its own framework?

Its AI efforts fall into four buckets: care delivery, software development, HQ productivity and a more speculative question: what increasingly capable AI and robotics might eventually do to care itself.

What connects them is Ian's sixth test: Does this help Honor get closer to changing how the world cares for its aging population?

## AI for delivering care

The most consequential applications sit inside the operations required to deliver care.

Honor’s care teams juggle competing priorities: staffing caregivers, responding to no-shows, confirming visits, taking calls, reviewing feedback and deciding what deserves attention next.

One application of AI is to synthesize the information available to those teams and surface a next best action. Instead of manually processing every signal, AI can narrow the problem to a simpler question: *What should I do next?*

Honor has seen meaningful improvements in core care-team performance metrics from applications such as these.

There are narrower applications too. AI can identify useful information buried in customer calls. For example, an incidental comment about a caregiver can be tagged as feedback for the relevant person.

## AI for building software

Honor has increasingly incorporated AI agents into software development, and the productivity gains have been encouraging.

This is particularly relevant to Ian’s second question: *Can the company actually ship AI?* Using AI to make engineers more productive creates something of a flywheel. AI does not only become another feature engineering teams build. It changes the way they build everything else.

## AI for HQ

The third bucket is broader employee productivity.

Teams across Honor are experimenting with AI tools that reduce the time between having an idea and testing it. Marketing teams, for example, can use LLM based workflows to develop landing pages and materials faster. Other applications connect AI with company knowledge, support and repeatable workflows.

None of these applications sounds as dramatic as replacing an entire job with an autonomous agent. That is probably a good thing.

The objective is not to accumulate impressive AI demos. It is to make people meaningfully better at the work the company already needs them to do.

## AI at the extreme

Perhaps the most revealing example is something nowhere near ready for large scale deployment. Honor has explored what increasingly capable robotics could eventually mean for home care, including prototyping around relatively simple physical tasks such as light housekeeping.

Today's technology is still far from replacing the breadth of work performed by a human caregiver. But that is not really the point.

Ian’s fourth question asks companies to assume that AI keeps improving and then ask whether the business still needs to exist. Honor’s answer is that people will continue to age and require care even as technology becomes dramatically more capable.

But leadership is not taking that conclusion for granted.

If robots eventually become capable of increasingly sophisticated physical tasks inside the home, what remains uniquely valuable? Which parts of care can be automated? Which parts should be? And where does genuine human presence become even more important?

Those are uncomfortable questions for a home care company to ask. They are also exactly the questions it should be asking.

And perhaps that is the larger lesson from Ian’s framework.

Being AI ready does not mean having the most agents, the largest AI budget or the longest list of experiments. It means AI changes how work gets done, experiments become production systems and leadership is willing to pressure test the company against where the technology is going.

Honor is still figuring out many of those answers. So is everyone else.

The companies worth watching will not be the ones that claim to have figured out AI. They will be the ones asking the right questions early enough and building the organizational muscle to act on the answers.

**The real test of an AI ready company is not how much AI activity you can see. It is how differently the company would operate without it.**