SPREAD AI
Go-to-Market
I turned go-to-market from a stack of tools into a system that works for the team.
Role: Marketing Ops, Sales Ops, Rev Ops, then GTM Engineer
Type: Internal, end to end (strategy, build, rollout)
Year: 〔2022–2026 / your timeframe〕
Company: SPREAD 〔one line on what SPREAD does〕
The one line
Over my time at SPREAD I moved from running the website and marketing operations to engineering the go-to-market system itself. Along the way I shipped two internal products: a GTM command center that connects the whole revenue motion to Salesforce, and a brand hub that brings marketing asset management into agentic GTM workflows through MCP. My job was to make the team’s daily work faster, clearer, and increasingly run by agents.
Headline results
Across a 30-person GTM org (19 AE/Sr AE, 3 BDR, 3 heads-of, 5 marketing):
- ~50 hours/week given back to the team, roughly 2,500 hours/year, by replacing manual reporting and prep with live, connected tooling
- ~1.8 hours/week per rep moved out of admin and back into selling
- 2 internal products shipped, plus 30+ workflows, in daily use by the whole GTM team
- 30 of 30 GTM members use the system weekly
- Marketing asset requests cut by ~30% through self-serve and agentic, MCP-based creation
- Salesforce as one source of truth instead of hundreds of disconnected spreadsheets and reports
These are internal operating numbers, modeled conservatively from team size and the friction the tooling removed. They show the GTM system is used daily and is changing how the team works, not just sitting in a tab.
The challenge
Go-to-market at a scaling company is rarely one problem. It is many small frictions compounding: data lives in different tools, reporting is manual, sales reps prep meetings from scratch, marketing assets are scattered, and nobody has a single view of what is actually working.
I owned this surface as it grew from marketing operations into full revenue operations, and the brief kept sharpening:
- Make the revenue motion legible. One place to see pipeline, performance and next actions instead of a dozen reports.
- Give reps leverage, not more dashboards to read. Turn Salesforce data into recommendations and prep, automatically.
- Make marketing assets usable inside modern, AI-driven workflows, not locked in a folder somewhere.
- Do all of this as a system that scales with the team, not a set of one-off fixes.
Approach
1. Operations first, engineering second.
Before building anything I ran the work by hand: marketing ops, then sales ops, then rev ops. That is what made the GTM Engineer transition real. I was not automating a process I read about, I was automating the exact frictions I had lived through. The tooling solves problems I felt personally.
2. Connect to the source of truth, then build on top.
Everything routes through Salesforce. Rather than create another data island, I made the GTM products read from and write to the system the company already trusts, so the dashboards, recommendations and task management all reflect one reality.
3. Agentic by design, not bolted on.
The newer work assumes agents are part of the workflow. The brand hub exposes its tools via MCP so an assistant can find, use and create marketing assets directly. GTM stops being a set of screens a human clicks through and becomes a system an agent can operate.
Product 1 — GTM command center
A full go-to-market dashboard and task manager, connected to Salesforce, built to run the daily revenue motion in one place.
- GTM dashboard and task manager. Pipeline, performance and the work that follows from it, in a single view.
- Interactive meeting dashboards. Reps and leadership walk into meetings with the live picture already assembled, no manual prep.
- Personal sales performance accelerator. Interactive and recommendation-driven, it reads each rep’s Salesforce data and surfaces what to do next, not just what happened.
- One source of truth. Because it is wired to Salesforce, the numbers in the dashboard are the numbers the company runs on.
Everyone in GTM opens the same tool and lands somewhere different. Reports are grouped by topic, the order of those topics follows your role, and pinning plus a recents row means the two or three screens you actually live in are never more than one click away.
This one replaced a slide deck. Opportunities that share a deal are collapsed into one row, and the status narrative, delivery team and leadership sponsor are edited inline during the meeting instead of being retyped into Salesforce afterwards. The columns are draggable and the arrangement is saved per person, because leadership, AEs and delivery each read the same table differently.
An account is not a logo, it is a few hundred people with different jobs. The map lays them out by department in product-lifecycle order and by seniority within each one, then colours every card by when we last spoke. The gaps are the point: a lane with no VP, a department nobody has called in a year, a champion sitting next to five untouched decision makers.
Signed Revenue was a spreadsheet before it was a screen — scenario weightings, sanity checks and the special-approval rules all came from that model, so the report reproduces them rather than inventing its own. Keeping the export means nobody had to give up the workflow they trusted in order to move.
Key decisions.
- Salesforce stays the system of record. The command center reads from and writes to it rather than keeping its own copy, so a status edited in the weekly meeting is a status changed in the CRM. It cost more to build and it is the only reason people trust the numbers.
- Layout is per person, not per role. Column order, pinned reports and topic order are saved for the individual, because leadership, AEs and delivery genuinely read the same table differently. One "correct" dashboard would have been wrong for everyone.
- Keep the spreadsheet's rules — and its export. Signed Revenue reproduces the Excel model's scenario weightings, sanity checks and approval logic instead of inventing better ones, and still exports .xlsx. Nobody had to give up a workflow they trusted in order to move.
Product 2 — Brand hub
A brand hub that brings marketing asset management and creation into agentic GTM workflows, with its tools exposed via MCP.
- Asset management as a system. A single home for marketing assets, structured so they can actually be found and reused.
- Creation, not just storage. The hub helps produce on-brand assets, not only file them.
- Available via MCP. The tools are exposed so AI assistants can search, use and create assets directly inside GTM workflows, a new agentic level for marketing operations.
The generator is the hub's answer to the request queue. Someone running a booth picks the event, and the artwork assembles itself from the campaign record — then they edit anything they like and export a PNG. Seven formats cover what marketing was previously making by hand.
The deck stopped being a file people copy and started being a kit people draw from. Pick the audience and the customer, take the slides that story needs, download a .pptx. The ✦ slides are the honest part: layouts built from copy assemble themselves, and the ones needing a bespoke diagram or real headshots hand off to "Ask Claude" rather than pretending.
Key decisions.
- Generators, not templates. A template is a file someone copies and then drifts from; a generator produces the asset from data every time. Brand changes land in one place instead of in a folder of stale copies.
- The CRM is the input. Assets are built from the campaign or account record, so the event dates, the attending team and the customer name are never rekeyed and never wrong.
- MCP first, screen second. Every generator is exposed as a tool an assistant can call, and this UI is just one client of it. That is what makes the hub usable inside an agentic workflow rather than only by a person with a browser open.
Website and marketing operations
Before and alongside the products, I owned the front door and the engine room.
- Website. 〔Designed / built / ran the SPREAD site, key changes and outcomes.〕
- Marketing operations. 〔Campaign ops, attribution, the stack you ran, what you fixed.〕
- This is the groundwork that made the GTM engineering credible: I understood the funnel because I had been running it.
Results and impact
- From operator to system-builder. Hands-on across marketing, sales and rev ops, then formally a GTM Engineer, so the tooling is grounded in real operational pain.
- Two internal products shipped, plus several workflows, all in daily use by the 30-person GTM team.
- ~50 hours/week given back (about 2,500 hours/year), with ~1.8 hours/week per rep moving from admin into selling, and full weekly adoption across all 30 GTM members.
- One connected revenue motion. Salesforce as the source of truth, surfaced through dashboards and recommendations the team actually opens, replacing manual reporting at every level up to heads-of.
- A head start on agentic GTM. With the brand hub on MCP, SPREAD’s GTM workflows can be driven by agents, not just people, cutting marketing asset requests by ~30% through self-serve and agentic creation.
What I’d carry forward
The best internal tools come from people who lived the problem first. Running marketing, sales and rev ops by hand was not a detour on the way to engineering, it was the spec. Every workflow I automated, I had felt as friction.
And it reframed what GTM engineering is. It is not building dashboards. It is deciding what the team should see, what the rep should do next, what an agent should be allowed to touch. The connection to Salesforce is a decision. The choice to expose tools over MCP is a decision. Each one quietly sets how the whole company works.
That is the kind of work I care about: taking the messy, human reality of go-to-market and turning it into a system that is clear, connected, and increasingly run by agents, where the team spends its energy on judgment and the tooling handles the rest.