Independent AI Consultancy
Throughline
Better decisions, earlier.
I build AI solutions that work with your organization, not the other way around. I work with leaders to understand the tools you already have, focus the effort, and leave your people owning it.
Not more drafts. Clearer decisions.
The hard problem isn’t generation. It’s preserving intent.
Making things is getting easy. But making was only ever part of the job. The rest is managing it, and for a lot of teams, planning hasn’t changed. A plan is briefed in a Word doc, presented in a PowerPoint, rebuilt in a spreadsheet, and still held in individuals’ heads and chats, without a system that knows the reasoning.
We are comfortable letting AI speed up our own work. Working with it together, across an organization, is still new. Faster drafting alone does not fix the plan. It produces more work that is still missing the why. And when the why is not written down, quality comes down to taste, and taste loses to a deadline. The real opportunity is AI that carries out a strategy, not AI that generates marketing: capture what the organization knows, settle the decisions early, and let systems carry them consistently.
-
AutomationOrchestration
Automation makes one step faster. Orchestration keeps every team, and every step, connected to the decision that started it.
-
GenerationDecisioning
Generation fills the page. Decisioning settles what belongs on it before anything is made: who it is for, what it is allowed to say, and why.
-
Always producingKnowing when to stop
A system that always returns something plausible is not mature. A good one flags, holds or rejects work that does not meet the rules.
Most work starts one of three ways.
Every team is somewhere different, so the shape follows the problem. A first conversation usually tells us both where to start, and whether larger work is worth doing.
Assess what’s needed
A working session on how work moves through your team today, which tools and pilots are earning their place, and where the effort should go first. I provide an assessment of where AI can accelerate decisions or where it’s slowing you down. This process helps leaders before they buy a platform, renew one, or start a pilot.
- A working session with your team or leadership
- A short written read-out of what is working and what is not
- Where to focus first, and what to stop
Rethink one process
We pick one real piece of your work, identifying a core challenge or roadblock. I map how your work moves today, rebuild the process inside the tools you already have, and test it with a working prototype.
- A map of how the work moves now
- A working prototype on a real problem
- The decisions it surfaced, made or queued
- A specification to build against, or a clear reason to stop
Guide the team
A senior seat beside your team through a launch, a platform choice, an agency change, or a new AI lead finding their feet. I help make the calls that matter, early, and build the habit in your people.
- Regular working sessions with your leads
- An outside read on vendor and agency direction
- Prototypes as new questions come up
- Your team owning the work when I step back
Throughline is Kris LoCascio.
I’ve worked in marketing and advertising since 2003. In 2008 I joined Heartbeat, part of Publicis Health. I am a creative director who never stopped building. I came up through design and interaction, most of it in healthcare, where every claim needs a source and an approval. I led a creative department of more than 60 people, and I saw the potential of AI for our business earlier than most. Because of this, I became our agency’s first-ever EVP, AI Transformation, leading AI adoption there and supporting it across Publicis Health.
Leading AI Transformation meant working directly with strategy, media, analytics and operations. That is where I learned to read a process and find where it breaks.
I build to find out. A real-life, tangible prototype teaches a team faster than a slide, and it shows where an idea breaks before anyone commits budget. You work with me directly, inside the tools you already have, not with a team assembled after the pitch.
- 23
- years in marketing and advertising
- 161
- AI tools built or tested
- 470+
- people in AI adoption I led or supported
Two examples, each in a short film.
-
Teaching taste to a machine
Every creative organization runs on a few people whose judgement everyone trusts and nobody can describe. This is an attempt to write that judgement down accurately enough to be useful, and honestly enough that the person it belongs to recognizes it.
The reviewer in the film is invented, the session is synthetic and the moodboards are generated. The film argues for where this goes; the system has been built and run one reviewer at a time.
Full case study -
Keeping a brand intact from hero to halo
The hero asset gets everyone’s attention. The halo is the hundred adaptations behind it, and it is where brand quality quietly goes. This turns the brand system, the approved claims and the business rules into something a production process can carry all the way down.
An explainer of the approach. No client work is shown.
Full case study
What people ask first.
-
We already bought a platform. Is that a problem?
No. The tools you have can often do more than they are being asked to. I start inside them, find what holds up, and only suggest something new when the work clearly needs it. I do not resell software.
-
Do you work with agencies or in-house teams?
Both. Agency leaders usually need to deliver what clients now expect from AI. In-house leaders usually need to make it work once the vendor has moved on. The work is the same: find what holds up, focus the effort, and leave the team owning it.
-
Do you only work with healthcare teams?
No. Most of my work has been in healthcare and pharmaceutical advertising, where every claim needs a source and medical, legal and regulatory review. That sets a high bar for quality, and it carries to any team whose work has to be checked before it ships.
-
Will we end up depending on you?
No, and that is the point. Each engagement ends with your people owning the work: the process, the prototype and the reasons behind them.
-
Do you build the software?
I build the process, not the software. I make prototypes and early versions to learn and to show. When something should be built for production, I write the specification and work with whoever builds it.
-
What does it cost?
It depends on the work, and I will not quote a number before we have spoken. There are usually three costs: setting up the foundation, keeping it current, and the work you run through it. I show each one separately so you can judge whether it paid off.
-
What happens in a first conversation?
About 30 minutes. You describe where work gets stuck. I ask questions. We decide together whether it is worth going further. There is no pitch deck.
What should AI solutions look like in your organization, with your tools and your people?
If your team keeps settling the same questions again, or you are looking at AI tools and unsure what they would change, that is a good conversation to have.
Or write to hello@throughlineai.co