More about our values…

Equity & Inclusion

What does “Equity & Inclusion” actually mean in your consulting?

It means we look at who an AI tool works for, and who it quietly doesn’t, before we recommend it.

AI systems learn from the data they’re trained on, and that data carries the blind spots of whoever collected it. An image generator trained mostly on white faces will struggle to represent everyone else. A resume screener trained on a company’s past hires will learn that company’s past biases. Research has shown that AI chat tools can respond differently to people based on how they write: misreading tone, giving lower-quality answers, or quietly treating some dialects as mistakes to be fixed. None of this is hypothetical, and a tool can have these problems while still looking impressive in a demo.

So when we evaluate AI for your business, equity is part of the evaluation, not a separate conversation. Here’s what that looks like in practice:

•       For your employees: We ask who the tool actually serves. If AI is screening applications, reviewing performance, or shaping schedules, we push to understand how it treats people who don’t look like the training data, and we insist a person stays involved in decisions that affect someone’s livelihood.

•       For your customers: We test tools against the customers you actually have, not just the average one. If a chatbot serves some of your customers worse than others, or your marketing images only depict one kind of person, that’s a business problem and a fairness problem at the same time, and we’ll flag it as both.

•       For your community: We think about who’s affected beyond your walls. That includes being honest about when automation would displace work rather than improve it, and favoring tools and vendors that take representation and accessibility seriously.

And when we find a problem, we don’t just shrug and say “that’s how AI is.” There’s real work that can be done:

•       We test before we launch. We try tools out using the kinds of messages, names, and faces your real customers and employees bring, so problems surface before they reach anyone.

•       We give the AI better instructions. These tools tend to follow directions well. Telling a tool to treat every customer with the same care, no matter how they write, genuinely changes how it behaves. So does showing it good examples to follow.

•       We pick the right tool for the job. Different AI tools have different weak spots. If one doesn’t treat your customers fairly, another one might, and we test rather than assume.

•       We keep people in the loop. Everything is overseen by a human. And after launch, we help you spot-check how it’s doing, because problems don’t always show up on day one.

We won’t pretend any AI tool is bias-free, because none of them are, and we can’t rebuild what’s inside them. What we can do is help you choose tools with open eyes, test them against your real employees and customers, catch problems early, and keep a person accountable for the decisions that matter.

Inclusion isn’t a feature you buy. It’s a practice you build, and we help you build it.

Ethical AI

We get asked what “ethical AI” means at Ethos a lot, and we think you deserve a real answer, not a marketing slogan. Here’s where we stand on four main pillars, in plain language.

What’s your take on AI’s environmental impact?

We won’t pretend this isn’t real. Training and running large AI models consumes significant energy and water, and the data centers behind them have genuine environmental costs. We’re not going to tell you otherwise, and we’re skeptical of anyone who does.

Here’s our honest thinking: the environmental cost of AI comes from the companies building and running the data centers, and that’s where the pressure needs to go. If you personally stop using AI, those data centers don’t run any less. But that’s not a free pass, and it doesn’t mean every use of AI is worth its cost.

So here’s what we actually do:

•       We recommend the right-sized tool, not the biggest one. Plenty of business problems don’t need AI at all, and we’ll tell you so. When AI is the answer, simpler and lighter often beats bigger. Wasteful use is bad strategy and bad stewardship, and we treat it as both.

•       We factor efficiency into our recommendations. Setting things up so AI does the work once, instead of over and over, saves you money and wastes less energy.

•       We support accountability efforts pushing data center operators and AI companies toward renewable energy, water efficiency, and real transparency about their footprint.

•       We put our money where our mouth is. One percent of the revenue Ethos earns goes to carbon removal through Stripe Climate, funding technologies that pull carbon out of the atmosphere permanently, not just credits that promise someone else will emit less. ‍

We don’t consider this solved, and we’re not going to dress it up as if it were. If you know of strong organizations working to shrink AI’s environmental footprint, we’d genuinely like to hear about them.

How do you think about privacy when using AI tools?

Simply: your sensitive information should never go into an AI tool, period. That includes Social Security numbers, dates of birth, financial account numbers, medical records, or anything you wouldn’t want stored or potentially exposed. No convenience is worth that risk.

Beyond that hard line, we help clients understand that most reputable AI tools offer real privacy controls: settings that let you opt out of having your conversations used to train future models, delete your data, or limit what’s retained. Most people never open these settings because they don’t know they exist. Part of our work is walking clients through exactly where these toggles live in the tools they’re already using.

Privacy isn’t an afterthought. It’s a setting we should actually choose on purpose.

What about AI and copyright? Isn’t that still a mess?

It was a genuinely murky area in AI’s early years, and we don’t dismiss that history. Real questions about training data and creator compensation were raised, and some remain unsettled in courts and legislatures. We’re not going to tell you that’s all fully resolved, because it isn’t, everywhere.

That said, the landscape has matured considerably. Today, we encourage clients to engage with AI tools in ways that respect original work:

•       Ask the AI to cite its sources, then actually check them. Good tools increasingly make this possible, and it’s a habit worth building rather than blindly trusting an unsourced answer.

•       Direct which sources AI can draw from when the tool supports it. Most platforms now let you scope research to specific, credentialed sources rather than an open-ended web crawl.

•       Treat AI output as a starting point for your own verification, not a finished, unattributed product. This protects you as much as it respects the original creators. And it’s another human-in-the-loop strategy.

We think of this as good practice, not a fully solved problem. The tools and norms are still evolving, and so is our guidance.

Where do humans fit into all of this?

At the center. That’s not a tagline for us — it’s the actual design principle behind how we advise clients.

We believe in human-in-the-loop AI: a real person reviewing, questioning, and ultimately deciding, rather than an AI system running unsupervised on anything that matters. AI is very good at drafting, summarizing, and surfacing patterns. It’s not good at carrying the judgment, accountability, or context that a human brings to a real decision. We help clients build workflows where AI supports that judgment instead of quietly replacing it.

That’s a related but distinct point from something else we feel strongly about: AI should make people more effective, not make them unnecessary. We’re not opposed to automation. Automation itself isn’t the issue. The question we help clients ask is how a given tool changes someone’s work: does it take repetitive, low-value tasks off their plate so they can spend more time on the judgment calls, relationships, and creative problem-solving that actually need a human? Or is it being used simply to eliminate the role altogether? Those are very different outcomes from the same technology, and we help clients to implement the first: using AI to free people up for higher-value work, not to replace them out of it.

None of this is about pretending AI is risk-free. It’s about using it with open eyes. Honest about the tradeoffs, careful with what we share, respectful of where information and content actually came from, and always keeping people, not just the technology, at the center of the decision.

That’s what “ethical AI” means to us at Ethos: not a perfect answer, but a genuinely thoughtful one.

Growth

What do you mean by “Growth”?

More than revenue.Revenue matters, and we won’t pretend otherwise: if an AI investment doesn’t eventually show up in your numbers, something’s wrong. But we’ve seen what happens when revenue is the only number a business watches. Teams burn out, service quality slips, and the “growth” turns out to be borrowed from somewhere else in the business.

So when we do an assessment or build a roadmap, we’re looking at four kinds of growth at once:

•       Revenue. Is this saving real money or opening real capacity? We push for AI investments you can trace to the bottom line, not tools adopted just because everyone else has one.

•       Customers served. Can you say yes to more clients without stretching thinner? The right AI often means serving the next customer without the next hire being an emergency.

•       Customer satisfaction. Are the customers you have getting a better experience, faster answers, fewer dropped balls? Automation that frustrates your customers isn’t growth, it’s a slow leak.

•       Worker happiness. Is your team spending more time on work they find meaningful and less on the tedious stuff? This one gets left off most consultants’ scorecards. We think it belongs on every one, partly because it’s right, and partly because it’s practical: the businesses that keep their people keep their knowledge, their relationships, and their momentum.

These four areas reinforce each other more often than they compete. Work that frees your team from busywork tends to improve service, and better service tends to show up in revenue. But when they do pull against each other, we’ll tell you, because a roadmap that grows one number by quietly shrinking the others isn’t a roadmap we’d put our name on.

Growth, to us, means your business gets stronger in every direction that matters, not just bigger in one.

Community

Why is “Community” one of your values?

Because technology is only worth adopting if it makes people’s lives better, and people don’t live in businesses. They live in communities.

When we help a business implement AI, the effects don’t stop at the office door. A local business that runs better serves its neighborhood better. A nonprofit that reclaims hours from paperwork puts those hours back into its mission. A team freed from busywork goes home with more left in the tank for their families and their neighbors. That’s the version of AI we’re working toward: not technology for its own sake, but technology that gives time and capacity back to the people and places around it.

It also means being careful, because the harms of badly implemented AI don’t stop at the office door either. When an AI tool serves some people worse than others, that erosion lands hardest on communities that already get the short end: misread by the chatbot, unseen by the image generator, screened out by the algorithm. Our Equity & Inclusion value covers how we prevent this inside your business. Community is about why it matters beyond your company: trust, once lost, is lost across a whole community, not one customer at a time.

And we’ll be plain about where we stand, because this is not a moment for mumbling. Ethos is a queer-woman-owned business, and we proudly support the queer community, women, immigrant communities, and communities of color. We know “diversity” is being treated as a dirty word in some rooms right now. We think those rooms are wrong. Diverse teams, diverse customers, and diverse communities are not a compliance checkbox; they’re where resilience, creativity, and honest feedback actually come from. AI built and implemented as if only one kind of person exists is harmful for the very fabric of community, and we won’t pretend otherwise to make anyone comfortable.

In practice, this shows up in choices:

•       Who we serve. Small businesses and nonprofits are the true fabric of their communities, and they’re exactly who big-budget AI consulting overlooks. We built our practice for them on purpose.

•       How people can access us. Not everyone can start with a full engagement, so we offer mini assessments, workshops, and group programs that make real AI guidance affordable, because a smarter and more empowered community of small businesses is good for everybody, including us.

•       How we design. We treat each AI roadmap as an opportunity to serve these communities well, as a design requirement, not a footnote.

We won’t claim to have this all figured out. Being a good steward of the communities AI touches is ongoing work, and we expect to keep learning it out loud. But you’ll never have to guess where we stand, and we think that’s exactly the point.

Empowerment

What do you mean by “Empowerment”?

Every value in this document so far has been about doing AI carefully. This one is about why we’re excited to do it at all.

We believe AI can genuinely make human lives better. Not in the vague, billboard sense, but concretely. It can help a two-person nonprofit take on a problem that used to require a grant they’d never get. It can help a business owner finally crack the operations puzzle they’ve been stuck on for three years. It can put analysis, drafting, and research capacity into the hands of people who could never have hired for it. Used well, AI doesn’t just do tasks. It multiplies the impact each and every person can have.

That’s the fork in the road every business faces with this technology, whether they see it or not. AI can be pointed at your people, as a way to need fewer of them. Or it can be pointed at your problems, as a way to help the people you have be more capable than they’ve ever been. Same technology, opposite futures. We are unambiguously in the second camp, and every roadmap we write reflects it.

Here’s what pointing AI at your problems looks like:

•       We start with what you can’t do today. The first question in our assessments isn’t “what can we automate?” It’s “what have you been stuck on, putting off, or doing badly for lack of time and hands?” That’s where AI earns its keep: not by shrinking your business, but by extending its capabilities.

•       We aim it at the ceiling, not the floor. Plenty of AI projects just shave costs off existing work. The interesting projects let your team attempt things that were previously out of reach: the analysis nobody had time to run, the follow-up that never happened, the service you couldn’t afford to offer.

•       We build your team’s wishlist into the roadmap. When we interview your staff, we ask what they wish they had more time for. Those answers don’t get filed away as nice-to-knows; they become roadmap items, so the time AI reclaims is deliberately pointed at the work your people have been wanting to do. Enrichment is designed in from the start, not hoped for at the end.

We’ll admit “empowerment” is one of the most abused words in technology marketing. Every vendor claims it, usually while selling headcount reduction with better fonts. So don’t take the word on faith. Look at the roadmap itself: your team’s wishlist will be in it, and you’ll be able to trace exactly which parts of the plan exist to make those wishes real. Empowerment you can point to in a plan beats empowerment on a billboard.

The measure of AI done right isn’t how many people you no longer need. It’s how much more your people can do, and how much more they enjoy doing it.