Higher Self

A technical partner for brands worth building.

I help founders of purpose-driven e-commerce brands find what their systems quietly get wrong. When AI can help, I build assistants that draft customer replies and website updates from approved information, ready for your team to review.

You work directly with me, and you stay in control.

I’m David Bentley, a systems engineer based in Peru, working New York hours. I reply within one working day.

Draft ready for your review In your Drafts folder
ToJordan at Refill Row SubjectRe: Spring reorder

Hi Jordan,

Thank you for the reorder. Twelve cases of the lavender bar1 are packed and leave on Tuesday.

You asked about the new wrap. It’s home-compostable2, and it’s the same plastic-free packaging3 you’ve had from us before.

Warmly,
Ana

  1. Checked against the order in your store.
  2. From your approved claims, word for word.
  3. Held back “plastic-free shipping”. Only “plastic-free packaging” is on your approved list.
Illustration with a made-up brand and order. Drafts arrive in the email you already use and wait for your approval.

Some of my work.

Each piece says what the evidence shows, and what it doesn’t.

Email-platform recovery for a UK wellness brand

2,151 profiles recovered and verified within the platform.

An account-wide suppression had gone unnoticed for fifteen months.

Not claimed: sales impact, which wasn’t measured.

Read the case study

An assistant that drafts within approved claims

Every piece approved by the owner, with a receipt for each approval and send.

Social posts, emails and page edits for a small brand, with product and health claims limited to approved wording.

Not claimed: time saved, which hasn’t been measured yet.

Read the case study

Public tools for Sofema Aviation

A Course Finder, a Training ROI Calculator and a Knowledge Hub, all live.

Plain-English regulation pages with source links and a record of each reviewer’s check.

Not claimed: adoption, bookings or return, which haven’t been measured.

Read the case study

Three kinds of help, one partner.

Find what’s quietly broken.

Your store, email platform and apps each report that they’re working. I check the joins between them, reproduce anything odd before I report it, and fix what the evidence supports.

In one engagement, the store believed 5,490 customers could be emailed. The email platform could reach about 150. Neither system flagged it.

An assistant in your inbox.

It drafts replies to customers, stockists and suppliers in your voice, using only claims you’ve approved word for word, the green and ethical ones included. Drafts wait for your approval in the email you already use.

When it isn’t sure, it leaves the decision to you. An unknown never quietly becomes a yes.

A partner for the next bottleneck.

Your store, email, automations, data and the awkward places in between. One problem at a time, each at a fixed price agreed in writing.

Each piece comes with a handover note, so anyone you trust can look after it.

For founders who care how things are done.

You’ve built a brand around a mission bigger than the business: fair trade, honest ingredients, rights for nature, things made to last. People buy from you because they believe you. You’re also the person every inbox, app and decision still waits on.

I only work with brands I’d buy from.

A good fit if

  • You sell online, and growth has made you the bottleneck.
  • You’d use AI if it kept your voice and your claims intact.
  • You’d rather know what’s broken than hear that everything’s fine.

Probably not a fit if

  • You want more email, sent faster, with less checking.
  • You need growth figures promised before we start.

You decide what automation is allowed to do.

We begin with human approval. Once a routine has been tested and its results measured, you can choose whether particular low-risk tasks should run automatically. Customer-facing claims and sensitive decisions stay under the approval rules you set.

  1. Stage 1

    You approve everything.

    Every draft waits for you. It’s how the assistant learns your nuances, and how you learn what it gets right.

  2. Stage 2, if you choose

    You approve what’s new or sensitive.

    Routine replies you’ve rarely changed are spot-checked instead. Anything new, unusual or claim-bearing still comes to you.

  3. Stage 3, if you choose

    Low-risk routine runs on its own.

    Only the tasks you’ve picked, after testing shows you almost never change them. One switch brings everything back to you.

The evidence informs the decision. You make it.

I measure how often you edit each kind of draft, and nothing gains new permissions just because the AI seems ready.

Start with two numbers.

A free first check for a failure that can sit unnoticed for months: your store and your email platform disagreeing about who can actually be reached.

  1. 1In your store, find how many customers have agreed to marketing email.
  2. 2In your email platform, find how many contacts it can actually reach.
  3. 3If nobody can explain the gap, email me both numbers. No logins and no customer data. I’ll reply within one working day, free.

Your two numbers

Enter both numbers to see how they compare.

Email me these two numbers

Nothing you type here is sent anywhere; the button opens your own email app. This is a first comparison, not a test of deliverability or marketing permission.

How working together starts.

  1. 1

    A short email.

    Tell me one thing you’d like to be easier, or send your two numbers. Free.

  2. 2

    A one-page outline.

    What I’d build, what I’d need from you, and a fixed price. Free, with no commitment, and a call if you’d like one.

  3. 3

    The build.

    A fixed price, agreed in writing before work starts. Anything consequential waits for your yes.

  4. 4

    Keeping it useful.

    Optional. We agree who updates it, who checks it and who hosts it.

You keep control of your accounts and data, and you receive the agreed custom code, documentation and records, so you aren’t dependent on me to maintain the system.

Notes from the work.

Your inbox is the AI interface you’re not using

Why I build assistants into the email you already have, not another app to check.

Why AI drafts sound almost right

Sounding like you is the easy part. Knowing what you’d never say is the work.

Treat your AI like a bright new intern

It has read everything and met none of your customers. The nuance is where the value is.

What would you like to be easier?

Tell me one thing that’s slower or shakier than it should be. A sentence or two is enough. I reply within one working day, on New York time, and I won’t chase you. Please leave out passwords and customer records.

david@higherself.ai WhatsApp +44 7749 776805