Introduction: Women in AI and the Power of Allyship

I'm actually just going to cede one minute of my time

just because I want to share

just some information about women in AI

just because I think it's great to see so many women here.

It's just fantastic.

But we're a massively underrepresented community in AI.

Just 1 % of venture capital funding went to women in 2024

and that's actually down from 2 % the year before.

The Funding Gap and Capital Efficiency

So interestingly though,

women -founded startups are far more capital efficient.

We tend to do better.

We don't hold as many roles.

We're definitely not a senior.

And the decision makers are not female either.

Call to Action for Inclusive Networking

So just, I think Josh and Jo from Mindstone, they're allies.

Mindstone is a nice, diverse and inclusive community.

But just when we're networking later on,

just make an effort to speak to the women

and make space for women in AI

because we've got a lot to add, a lot of value to bring.

Closing the Opening Remarks

Thank you.

Speaker Background and Journey

This is me.

Please connect on LinkedIn.

I studied business management at King's.

I then went to Deloitte, where I was a management consultant for NHR Technology.

I then went to BlackRock, where I helped run the fixed income business and then was in corporate strategy.

I left the capitalist model for a while.

I went and learned about community healing and trying to learn from indigenous cultures.

Watchers.

I set up an incubator in India, co -founded a charity to support Palestinians,

and I campaigned with Extinction Rebellion.

From Corporate to Community and Activism

When I decided to sort of come back into founding

things at the advent of AI, I could see the power and the value that AI agents can bring.

Re-entering Tech: Why AI Agents

So I'm going to talk to you about how I'm using AI agents and hopefully encourage a

lot of you guys to use them yourself.

And if you have any questions about the work I've

done um you know message me on linkedin and um i'm happy to talk and share everything that i've

Introducing ABLE: An AI Multi‑Agent Platform for Gig Work

learned so this is able i say it's the world first gig work ai multi -agent there are a lot

of gig work platforms out there but none are like able um and we are we're launching with

Initial Markets and Expansion Vision

hospitality and events because that makes sense that's the third largest employment sector in the

uk but we've got i've got eyes on healthcare which is the largest retail as well as the second

and largest but for any kind of shift work able can um is going to to change the landscape we

Reorganizing Work with AI

think about order we think about automation when we think about this these big changes that happen

we think things are going to be automated but actually it's more disrupted than that things

are being reorganized and ai is a platform to really reorganize systems and that's what um

able is coming in to help change the way that shift work is allocated because at the moment

The Broken On‑Demand Hiring Process

if you want to hire somebody

you have to

for an on -demand shift

you have to follow the traditional recruitment process

you post a job description

you wait for your applications

you review the applications

it's just nonsensical

so

ABLE’s Personalized Service Experience

Abel's come in to provide a

I say a personalised service experience

outside of

what currently happens

so let's talk about

The Gig Economy Landscape

the gig economy very quickly

politically it's it's massive we're looking at almost two trillion dollars by 2032 50 of the

u .s workforce are freelance or will be freelance very soon in the uk we're more like 20 25 but

obviously um there's there's a lot of precarious there's a precarious state in the labor market at

the moment and and until we know exactly how ai is going to impact us all that's probably going

to get worse so platforms that offer resilient options to help people earn money that's definitely

Ethical AI as Opportunity and Responsibility

what able is and i'm going to talk a bit about ethical ai because i think well a ethical ai is

the biggest opportunity of our lifetime and i think it's really important that we try and

Challenges Gig Workers Face

build ai systems that are ethically minded gig workers can spend up to 10 hours a week searching

for work applying applying applying applying it's tedious and it's unfair they often also get billed

so they don't even earn 100 of their wages in london as an example for hospitality events

Market Size and Ambition

There are 30 ,000 temp shifts a day.

So the market's just massive.

So what we're building, intelligent hiring, empowered talent,

and ultimately the hiring infrastructure for 3 .6 billion deskless workers.

How ABLE Works

For Freelancers: Profiles, Availability, and Feedback

So for freelancers, an AI agent helps you create

one of these really funky shareable profiles.

You film a bio, and a generative AI agent

it sort of interviews you

then helps you create a script

to help sell yourself

you say when you're

available in a calendar, that means that

when somebody's looking for a chef, we know

that Sarah's available and can be put forward for the job

we gather

feedback, there are no star ratings

in ABLE, the academic perspective

on star ratings, they are not helpful

it leads to

the productification of people

you shouldn't be rating people out, giving people

stars, so instead

Instead, we take feedback at the end of the shift and an AI agent then transposes that feedback into development points, trying to support and empower and improve our users.

For Managers: Value, Matching, and Bias Reduction

For managers, we are by far and away the best value.

Often these agents charge 25 task rabbit charges, almost 50 % fees.

We're charging eight.

That includes public liability insurance and Stripe payments.

So, Able itself is operating about 5%.

The buyer views video bios and the gig folios of up to five candidates matching.

We don't say, this is the candidate for the job.

The buyer, and I'll show you a quick demo of what it looks like when the buyer goes on to search.

The buyer asks, describes what they're looking for, and then the AI matches.

matches.

When I talk a bit more about ethical IR, I mention it, but our matching agent does

not have access to demographic data.

The age, gender, nationality, race, none of that is

known by the matching agent.

The matching agent matches purely based on skills, location,

availability, and then other components like responsiveness inside the app, things like

that.

We're really trying to remove the bias.

Tech Stack and Agent Ecosystem

A little bit about the tech stack.

We're front

Frontend, Next .js, React 19, a whole load of APIs, obviously, Stripe and what have you.

Current Stack and Launch Setup

At the moment, we're launching with six agents, LLMs based on...

We're using Gemini for our launch, Postgres data, and we're using Amazon at the moment.

We're going to be switching to Google because they've given me a load of credits.

Thank you, Google.

Future Agent Ecosystem and Launch Agents

got um so this is the the future agent ecosystem kind of this will be version post mvp but you

know the intention for able is that it's going to be sort of like a one -stop shop to help freelancers

manage their money manage their savings investments that kind of thing that's later down the line

what we're launching with is this pure hr model um and this whole um like gambit of um agents

but yeah we're launching with six of them gig folio coach that helps create these profiles

an agent that checks that your your right to work an agent that helps you improve your profile

an agent that creates these trainings based on the feedback you're receiving

an agent that makes sure that conversations that are happening are following our able values be

Building Ethical AI at ABLE

good be fair be kind so i'm just going to talk quickly about ethical ai um i'm learning a lot

Learning and Community Influence

There's a guy here, Joe Fennell, I don't know if he's here,

but I met him through the MindStone community.

He's definitely helped educate me a bit.

I do a lot of self -learning, but I'm not an expert by any means on ethical AI,

but I want to share what I'm doing to try and build an ethical product

because I think we should all be thinking about how can we build an ethical product

because obviously the risks around AI are massive

and ethical AI is something that really protects us against those risks.

Values, Mission, and Governance

We're a values -based business, be good, be fair, be kind,

and those values all the ai agents are instructed to adhere to those values at all times

we're a mission -based business we're low cost we deliver this training we're all about social

mobility um i learned about peace tech through mindstone actually so we have a bit of a peace

tech pledge around um the long -term aim of only working with other technology companies that are

have a peacefully minded we want to be a trusted provider so we follow um the ethical ai charters

that are coming out from the OECD,

the EU, AI Code of Conduct.

We use all of those.

I give them to all the agents.

I instruct the agents

that these are the measures

by which we'll be measuring ourselves.

We're GDPR and SOC 2 compliant,

or we have a compliance roadmap for SOC 2

for hopefully when we launch into the US.

And then we're trusted in the sense

that we have these right -to -work checks

and what have you.

Practices: Human-in-the-Loop, Bias Removal, Transparency

In terms of pure ethical AI,

we have humans in the loop,

So we're monitoring what the AI is delivering to make sure that it makes sense and is fair.

There are ways to dispute any result.

As I've mentioned, we've removed the bias from the matching, so we don't have those demographics.

We monitor any hallucinations.

We're instructing our agents not to guess.

and we're going to be publishing our algorithm weights

which I think is a balance between

what's your IP versus wanting to be really open

and I think the best way that we can be ethical at this stage

is trying to be as open as possible

Traction and Programs

so I'm going to give you a bit of a demo in a second

Onboarding, Timeline, and Partnerships

we've got over a thousand freelancers

onboarding at the moment

the demo should work, it will work a bit

but we should be live by before Christmas

that's definitely the target we've got maybe 40 or 50 venues signed up if you know any venues

if you're interested please come and talk to me afterwards we're obviously keen to get people

involved we're part working with the department of trade and business part of the uk innovative

program we're part of the london mayor's grow london program which is fantastic if anybody

hears i would recommend looking into both of those programs and i've got right and bandstead my home

hometown council has sponsored me as well

University Engagement and Talent Pipeline

when I'm taught I presented

to the ASAC

community a few weeks ago

which is a student body

international student body but when I

contact universities

I'm now in with a whole load of London universities

and that actually includes Imperial, Kings

and UCL

so in terms of the cadre

of students that will be on the app

we've got a great

selection of students that you can be hiring

Product Demo: Hiring a Bartender

but also we've got some excellent professionals as well okay time for the demo okay so here's

Dashboard and Search Flow

the dashboard um for this is the you can switch between your buyer and your freelancer your

manager and your giggy view but here's i'm going to just try and show you what it would be like

if you want to hire a bartender so you chat with the agent says what do you what do you what are

you looking for oh i did not spell that very well this is part of ethical ai as well having having

the agent repeat back to you what you're saying just to make sure that everyone's on the same

page so you make sure there's no assumptions so here it's asking for any specific skills or

experience okay maybe i'm not giving it enough detail oh yeah we've coded it so you have to pay

more than the london minimum wage of the london london waiting wage um and ultimately we're going

Toward Richer Matching with Semantic Search

going to be using vector and semantic search so the the as we get really rich profiles the ability

to match according to you know what you really need cocktail making japanese knife skills whatever

it might be is going to be um it's going to mean that the um the candidates that you get suggested

to at the end are going to be quite perfect so it creates this a summary and then confirm oh yeah

Okay, sorry.

Candidate Results and Closing the Demo

At this point, we get a full list of candidates

that you can then go and you click

and you can view their profiles.

So that's unfortunately half the demo.

But ultimately...

Okay, thank you.