Employer branding · Technology
Every company says AI now, so the word tells a candidate nothing at all.
For an AI, machine learning or analytics company, employer branding means showing what your models do, on what data, and whether any of it runs in production.
Scope, fee, reporting and term are agreed in writing after we have read your brief and spoken to you.
In one paragraph
For an AI, machine learning or analytics company, employer branding means showing what your models do, on what data, and whether any of it runs in production. GullySales starts with an hour beside the engineer who shipped the model, then checks the publication rule with whoever signs your customer contracts and the take-home with the hiring manager before anything goes live.
Also written for
AI companies, ML companies, Data analytics companies.
What we do not do
We do not recruit, supply candidates or run payroll. We help candidates see the employer you already are.
Where this sits
For AI, ML and data analytics companies
Where it usually goes wrong, and what we would do.
“The word AI has stopped carrying information”
A dashboard company, a data labelling operation and a team training models all use it. Anybody with real modelling experience has learned to discount the word and hunt for something concrete.
“Whether a model ever shipped is the question”
Research that never left a notebook is a familiar story and candidates have heard it twice. What went into production, what it decides, and what happens when it gets something wrong.
“The post asks for a doctorate and offers reporting work”
Plenty of good analytics jobs exist. Advertised as research, they attract people who resign in six months, while the people who would have been happy in them never applied.
“Data access and compute go unmentioned”
An engineer wants to know whose data they get to touch, how dirty it is, and who signs off a GPU bill. Silence suggests the answer is a shared laptop and a pending approval.
What we do
What we deliver for AI, ML and data analytics companies.
Every deliverable, what it covers for you, and the result it is there to produce. Nothing here is an extra.
A careers page about the models, not the category
What your systems predict or generate, what data sits behind them, where they run and who uses the output. Drawn from an hour with the model builders.
Result: Engineers can judge the work instead of guessing at it. For example, a page for a data team might open: "Our churn model scores 40,000 accounts every night and the retention team calls the top 200. When it is wrong, this is who finds out." The numbers in that sentence are invented; yours come from the team.
Honest splits between research, engineering and data work
Separate posts for modelling, for the platform around it, and for annotation and pipeline work, each saying how much of the job is cleaning data.
Result: The post says up front how much of the job is cleaning data.
Publication and open-source policy
Whether people can publish, speak at a meetup or release code, who approves it and what your customer contracts allow. Confirmed with whoever handles those agreements.
Result: Researchers know what they can put their name to before they sign.
The take-home, described before it is sent
Whether there is one, how long it should take, who reviews it and whether you pay for the time. Agreed with your hiring managers so every candidate gets the same version.
Result: A candidate knows the take-home before they are asked to do it.
Engineer accounts of a model that failed
One short piece on something that did not work and what you changed afterwards, reviewed by the engineer who lived through it.
Result: Credibility with the exact people who can tell a real team from a slide.
How the result is measured
- Applications received per open role
- Careers page engagement and time on page
- Candidate drop-off during the interview process
- Referral hires as a share of total hires
Recorded as a baseline before work starts, so every later report has an honest comparison.
How it works
How the work runs for AI, ML and data analytics companies
The work itself, in order. How an engagement starts, from your brief to a written scope, is set out on our method page.
- 1
Talk to current employees
We interview a cross-section of staff honestly, including some who might raise uncomfortable points.
- 2
Find the genuine reasons to join
We identify what is actually true and attractive about working there, rather than generic claims that could apply anywhere.
- 3
Build the careers content
The careers page, job post templates and employee stories are written from what staff actually said.
- 4
Review the candidate experience
We check what happens between application and offer, since branding cannot fix a process that loses candidates through delay.
- 5
Hand over and refresh periodically
Content is handed over for ongoing use and revisited periodically as the team and roles change.
Before you hire anyone
The first thing to do yourself.
Read your company the way a good candidate would, before you change anything.
- Search for your company name with the word jobs, and read the first page of results as a candidate would.
- Read your last ten reviews and list the points that appear more than once.
- Ask your three newest joiners why they said yes, and what almost stopped them.
Judge any provider
What to ask any agency or ghost-writer.
Useful whoever you choose, including us.
- Will you work from what our people actually say, or from a template?
- Do you do recruiting or supply candidates, or only the employer-brand work?
- Who approves anything that speaks for our staff?
- Which of our existing job pages and profiles will you change, and which will you leave alone?
- What do you count as a result, and what will you not promise?
Where this fits
Three pieces of sales work, in this order.
GullySales builds every engagement from three pieces of sales work. This page is not one of them, but the same three steps apply to the business behind it.
Step 1
Sales audit
Find where deals are being lost and what to fix first.
Start here if you cannot say where orders are lost.
Step 2
Sales process design
One written way to sell, from first enquiry to repeat order.
Start here if every salesperson sells differently.
Step 3
Sales management
A target, a review routine and someone who owns the number.
Start here if the team has targets and no routine.
FAQ
Questions owners ask before they call.
Not here? More answers, or ask on WhatsApp.
We mostly build on other people's models. Will engineers write us off?
Can you write our job posts and also find candidates?
We renamed our analytics team. Is that a problem?
How much does it cost?
How long is the contract?
How soon will we see results?
Who will actually do the work?
What do you need from us?
What if we are not happy with the work?
Your next practical step
Tell us what you are trying to grow, and we will read it before we call.
Eight short steps, mostly taps. Skip anything you would rather answer on the call.