PROPOSED Sequel to the May 7 workshop · Prepared for Preston, Alan & the BU presidents

Nine AI plays that move the P&L in 90–120 days.

Pulled from the workshop output, the leadership team's stated priorities, and the candid catch-up with Alan — sequenced so the first wins ship inside Q3, the foundational lifts land before year-end, and every play has a defensible business case the board can underwrite.

For Preston Bowen, Alan Rupp, MBP Group leadership
From Daniel Townsend, Copilot Innovations
Horizon 90–120 days, P&L-impact bias
Source Workshop · Alan catch-up · BU intro call
Combined P&L Lens · Year-1 Realized
$4.5–8.5M
Estimated annualized impact once the nine plays are sequenced and at run-rate. Conservative–optimistic bracket, discounted for ramp.
Revenue lift
$2.4–5.5M
Margin / Cost-out
$1.9–2.8M
CX / Retention
material
The premise

Practical, P&L-first — not platform-first.

Preston's direction to me was unambiguous: look for things that can make a P&L impact within three to five months. Alan's catch-up sharpened it. Marwin is in trouble and needs revenue, full stop. SSI is data-rich and has the leader to run fast. Millwork 360 is bleeding the easy, profitable orders to competitors with online quoting. Summit is growing and pushing into Texas. HB&G can wait.

So this roadmap deliberately doesn't lead with the platform-of-platforms ambition. It leads with eight plays targeted at the BUs that need it most, in the order they'll show up in the P&L. The bigger cross-brand plays from the opportunity report — CPQ across all five, Customer 360, "Ask MBP" — are real and worth doing. They follow these.

"Let's look for things that can make a P&L impact within three to five months… anywhere we could grow top-line revenue with Marwin — that's number one on the list. SSI is the data-rich environment with the right leader. Millwork is the other big area where we have to get that one figured out." Alan Rupp · CFO · catch-up call, May 2026
At a glance

Nine plays. Five BUs. Three P&L levers.

Each tile is one play. Color of the dot = which P&L lever it pulls — revenue, cost / margin, or customer experience. Click through to the detail below; or read the sequenced timeline to see when each one starts paying.

Play 01
Marwin
Anonymous Traffic Recovery
Sprint 1Revenue
Play 02
SSI
Lead Scoring & Ad-Spend Tuner
Sprint 1Rev + Cost
Play 03
SSI
Stalled-Lead Re-engagement Agent
Sprint 1Revenue
Play 04
Marwin
Texas Rep Briefing Engine
Sprint 2Revenue
Play 05
Millwork 360
Interior Door Express Quote
Sprint 2–3Rev + Margin
Play 06
Summit
Builder Conversion Playbook (TX)
Sprint 2–3Revenue
Play 07
Platform
Spend Map · Phase 1
Sprint 3–4Cost
Play 08
SSI
Real-Time Call Coaching
Sprint 3–4Rev + CX
Play 09
All 5 BUs
Universal AI Chat (Omnichannel)
Sprint 1–2Rev + CX
Play
Sprint 1
Sprint 2
Sprint 3
Sprint 4
01 · Traffic Recovery Marwin · Revenue
02 · Lead Scoring + Ads SSI · Rev + Cost
03 · Re-engagement Agent SSI · Revenue
04 · TX Briefing Engine Marwin · Revenue
05 · Express Quote Millwork 360 · Rev + Margin
06 · Builder Conversion Summit · Revenue
07 · Spend Map · Ph 1 Platform · Cost
08 · Real-Time Coaching SSI · Rev + CX
09 · Universal AI Chat All 5 BUs · Rev + CX
Revenue Cost / Margin Customer Experience Mixed lift ● = first measurable P&L signal
The nine plays

Each play, with the business case in one panel.

The story is the same shape every time: what's broken today, what the AI does about it, why MBP can ship it in 90 days, and the dollar / KPI lens to manage it against. The mini-chart shows the P&L pattern in the first 12 months.

PLAY · 01 Marwin Revenue lift

Anonymous Traffic Recovery — turn the 95% of unconverted visitors into a pipeline.

Marwin.com gets meaningful B2B traffic. Today, only the 2–5% who fill out a form become known leads. Everyone else — distributors, builders, contractors evaluating an attic stair or pocket door — leaves no trace. An IP-reverse-lookup tool plus an AI enrichment + outreach pipeline turns that anonymous traffic into named accounts the sales team can actually work.

Marwin is the platform's biggest revenue problem. Demand is down, margins are compressed by Polish and Malaysian competition, and the team is going through a reduction in force. Any incremental top-line lift without adding sales headcount is critical.

(1) Deploy a reverse-IP visitor identification tool on Marwin.com (RB2B, Clearbit Reveal, or similar — <$500/mo). (2) An AI enrichment pipeline appends company size, role-likely contacts, and recent project signals. (3) An outbound agent drafts a personalized first-touch email per identified account; sales rep approves and sends. Plugs straight into the lightweight Salesforce instance Marwin has today and into the NetSuite CRM coming in February.

Pure B2B + low form-fill conversion is exactly the pattern this tooling was built for. Marwin pays for the traffic already. The cost-per-lead math collapses if even a small fraction of identified accounts become outreach-worthy. And the RIF makes it more important, not less — the smaller sales team needs a fatter qualified pipeline coming in the door.

PLAY · 02 SS Industries Revenue + Cost

Lead Scoring & Ad-Spend Tuner — stop paying full price for the leads that don't close.

SSI runs a full Salesforce instance, has 21 reps closing $10K stairs over the phone, and — per Alan — spends "incredibly expensive" Google PPC dollars to fuel it. An AI lead-scoring layer + an ad-spend feedback loop does two things at once: routes the top-decile leads to the top-decile reps, and re-allocates Google budget away from the keyword × geo combinations that statistically don't convert.

"Our pay-per-click advertising is incredibly expensive in that business. We ought to be able to refine how we're looking for leads and figure out a way to increase our lead conversion rates so that the quality of incoming leads is better and more likely to convert."

Two coordinated layers on top of Salesforce + the Google Ads account. Lead scoring: a model trained on historical SSI lead → close patterns scores every inbound lead in <30 seconds and routes accordingly. Ad-spend tuner: a weekly AI agent reads what actually closed, identifies the patterns (keyword, landing page, geo, device, time-of-day), and re-weights bid strategy. Joe gets a weekly "what to keep, what to kill" report.

SSI is the platform's only B2C business and — as Preston flagged — has the right leader to run fast. The data exists (Salesforce is doing everything including manufacturing scheduling, so close-loop is unusually clean for a building products business). And Joe's already running a Dialpad AI evaluation — this fits the energy.

PLAY · 03 SS Industries Revenue

Stalled-Lead Re-engagement Agent — keep talking to the customers SSI already paid to find.

For every closed deal SSI has dozens of leads that went quiet — got a quote, never replied, or said "next month" and dropped off. An AI agent watches the Salesforce pipeline, drafts the right re-engagement touch (email or SMS, personalized to the original conversation), queues it for the rep to approve and send. Same product, same reps, same lead pool — more revenue per lead.

"The feedback process in that business, where we're constantly staying in touch with those potential customers and actual customers… we're providing great interaction with them. That would be a perfect place for AI to help."

An agent that pulls every lead older than N days with no recent activity, reads the original conversation history (Dialpad transcripts already live), drafts a tailored next-touch ("you were asking about a 6′ spiral in oak — here's the lead time and pricing as of this week"), and routes it for one-click rep approval. Branches for hot ("call in next 24h"), warm ("email + 7-day follow-up"), cold ("monthly nurture").

$10K AOV means even single-digit additional closes per month moves the P&L. The Dialpad + Salesforce stack already captures everything needed. And because the rep stays in the loop with one-click send, it doesn't disrupt Joe's existing sales motion — it just multiplies it.

PLAY · 04 Marwin Revenue

Texas Rep Briefing Engine — ramp the two new reps in half the time.

View interactive prototype →

Marwin just replaced a retiring Texas rep with two new ones covering what's apparently the biggest single geography on the platform — bigger than the entire Southeast. New reps in a territory they don't yet know is an obvious AI assist: pre-call briefings on every dealer and builder, prioritized account lists by metro, tariff and pricing context built in. Productizes Alan's existing housing-starts + net-migration work for the people who need it most right now.

Preston has redirected Mark and Jason to focus Marwin's leadership bandwidth on sales. Dave Conrad is going down to Florida to take operational load off Jason. The two new Texas reps need to ramp fast — there is no budget or appetite to wait 6–9 months for natural acclimation in a market that big.

(1) Productize Alan's net-migration + housing-starts model into a metro-level demand score for Marwin's product mix. (2) An AI account-briefing agent pulls public construction data, permits, and (where available) prior order history per dealer/builder, drafts a one-page brief before every call. (3) Weekly territory dashboard surfaces "10 accounts to call this week" with the why.

The data foundation is already half built — Alan has the pilot. The rep team is small and named, so we can white-glove the rollout. And it directly answers Preston's stated priority: anything that helps Mark Sendar grow Marwin's top line is the most important thing on the platform right now.

PLAY · 05 Millwork 360 Revenue + Margin

Interior Door Express Quote — recapture the easy, profitable orders Millwork lost to faster competitors.

View interactive prototype →

Alan's diagnosis was sharp: Millwork 360 got out-paced by competitors who built online quoting. The simpler, margin-rich doors went to those competitors. What's left at Millwork is the complex, expensive-to-quote business that's eating the org. This play doesn't try to solve all CPQ at once — it carves off the "interior door" pattern and lets a customer get a quote in five minutes, online, no human touch.

"The lower-end easy business, which is fairly margin-rich, has left. It's been replaced with this complex stuff… cataclysmic event there with all that. It's not a profitable business — it's barely break-even, and we've got to get that one figured out."

A narrow CPQ slice: a customer-facing express-quote form for the door types Millwork knows it loses to competitors. AI applies the BOM, design, and pricing rules, returns a quote in under 5 minutes. Quotes that fail any guardrail (custom dimensions, non-stock species, panel complexity) fall through to the existing human flow. Sits alongside, not on top of, Synergetics' current work — narrower scope, faster ship, same eventual platform pattern.

The room voted CPQ #2 and Quoting/Take-Off #1. This is the smallest viable slice of that idea — the one that fixes the specific revenue leak Alan named, not the full agentic vision. It also creates the template that Marwin and Summit can adopt later. Build once, prove margin recovery, then expand.

PLAY · 06 Summit Stairs Revenue

Builder Conversion Playbook — find the on-site builders ready to switch to pre-fab, starting in Texas.

View interactive prototype →

Summit's growth story is converting builders from on-site stair building (two days) to pre-built Summit staircases (30-minute install). Today that's done by the sales team on intuition. An AI agent that scores every active builder in Summit's target metros on switch-probability — based on permit volume, project mix, review signals, and observable construction patterns — gives Matt Fox's team a ranked target list and a personalized opener for each.

"With Summit, the play is you've got builders who are outside building their stairs. That's where we're getting share from — we go in and say, hey, we can sell you that completed staircase. Going big into Texas right now."

Three-step agent. (1) Pull public permit and construction data for Summit's target metros (initially Texas). (2) Score every builder on switch-probability — homes/year, span types, observable bottlenecks. (3) For the top 100/week, draft a personalized outreach with the install-time and labor-cost math specific to that builder's home types. Matt's team prioritizes from the ranked list.

Summit is growing in a declining market and Texas is the strategic frontier. Sales prioritization is the single most valuable AI lever for a sales motion that's expanding faster than the team can hire. And the same scoring engine extends to the other distribution centers Summit is standing up — it's a multiplier on every future market entry.

PLAY · 07 Platform / Procurement Cost / Margin

Spend Map · Phase 1 — give John a single view of the $90M he buys across five disconnected ERPs.

View interactive prototype →

Alan's clearest operations call-out: $90M of material spend across five ERPs that don't talk to each other, with one person — John — responsible for finding consolidation savings. The AI play is straightforward and high-confidence: pull PO data from each BU, normalize vendor names and SKU descriptions with AI-assisted entity resolution, surface the overlapping vendors and unconsolidated categories. Even a 2–3% consolidation save on $90M is $1.8–$2.7M. This is the highest-confidence dollar handle on the platform.

"He's responsible for that across five ERPs that aren't connected. So it's disjointed at best. Anything we could do to simplify the data access and then the ability to analyze that data and figure out how we get commonality in the product lines we're buying… that would be phenomenal."

Phase 1 is visibility, not optimization. (1) Pull PO extracts from each of the 5 BUs (acknowledging the manual extract pain). (2) AI normalizes vendor names ("BFS Raleigh" = "Builders FirstSource - Raleigh" = "bfs raleigh"), SKU descriptions, units of measure. (3) Output: a single dashboard for John showing where the same material is being bought from different vendors at different prices across BUs. Phase 2 (negotiation playbooks, supplier risk) follows once the data foundation lands.

John was hired exactly for this mandate — the AI doesn't replace him, it gives him in 90 days what would otherwise take 18 months. The savings is the highest-confidence number in this roadmap because the math is the same math every PE-backed platform runs on consolidation. And it's the Validor-readable proof point the next acquisition narrative needs.

PLAY · 08 SS Industries Revenue + CX

Real-Time Call Coaching — lift the bottom-half SSI reps to the top-half's close rate.

View interactive prototype →

Joe told us SSI's lead conversion swings from 30% down to 10% across reps — and called it explicitly a "people and coaching issue." A real-time AI listening layer on top of Dialpad surfaces objections, pricing pushback, and missed buying signals to the rep in-call. Builds on the Dialpad sales-assist evaluation Joe already has running. Lifting the bottom-half rep population by even 4 points of close rate is seven-figure incremental revenue at high contribution margin.

21 reps, $10K AOV, 21 reps' worth of variance in close rate. Today's coaching is post-hoc and manual; the top reps stay top, the bottom reps stay bottom, and the playbook lives in two or three senior heads. Joe is actively evaluating Dialpad's native AI — augmenting that with SSI-specific win/loss patterns is the difference between a generic sales-assist tool and one that actually moves SSI's numbers.

Layer on top of the Dialpad evaluation. The AI listens live, identifies the buying-signal/objection pattern from SSI's historical wins, and surfaces a rep-side prompt ("she mentioned a step-down — confirm height before quoting") + the right next move. Post-call, the same AI auto-grades the call against the playbook and feeds the rep a 60-second debrief. The patterns get fed back into onboarding for the next hire.

Same product, same reps, same lead flow — better close rate. The cleanest revenue ROI story on the platform. And because the data is mostly already captured (Dialpad transcripts, Salesforce close patterns), the build is implementation, not invention. The PE-side narrative also writes itself: AI-coached sales force, generalizable across the platform.

PLAY · 09 All 5 BUs · Cross-channel Revenue + CX

Universal AI Chat — one smart conversation, every channel, every brand.

Customers, dealers, builders, contractors, and suppliers reach MBP brands today through one of two doors: a phone call or a contact form. Most people won't do either. They'll DM on Instagram, message on Facebook, or text — and if nobody answers there, they go to a competitor. A universal AI chat sits across the website, Instagram, Facebook, and SMS for each brand, understands intent, and either answers, captures the lead and triggers automation, books the consult, or escalates to a human. Low cost, fast to ship, and a proven pattern we've delivered before.

Across all five brands, the contactable surface is narrow and unfriendly to how buyers actually behave in 2026. Mark, Joe, Jason, Jeff, and Matt all sit on engaged social audiences and trafficked websites where the friction to "ask a question" is too high. The lost intent — the visitor who almost reached out — is invisible in every BU's metrics today.

A single conversational layer wired into each brand's web chat, Instagram DMs, Facebook Messenger, and SMS short code. The AI is trained per-brand on product catalog, lead time, freight, install notes, and FAQ — and routes intent to the right action: answer the question, book a consultation or quote appointment, capture contact info and kick off an email/SMS nurture, or escalate to the named sales rep with the conversation context attached. One platform-level build; five branded deployments.

Three reasons. First, it's the cheapest play in this roadmap to ship and the one with the broadest cross-BU surface — same engine, five front doors. Second, it's a pattern we've deployed before for comparable clients, so the implementation risk is low and the timeline is real. Third, it pairs naturally with Plays 01, 02, and 03 — the chat captures intent the visitor-recovery tool couldn't, and feeds the SSI lead-scoring + re-engagement pipelines with cleaner data.

Sequencing

What ships when, and why this order.

The first four plays ship across Sprints 1–2 and don't depend on any cross-BU data work — they get the organization to a P&L signal fast and build credibility for the bigger lifts. The Universal AI Chat (Play 09) runs alongside as the platform-level CX layer that every other revenue play benefits from. The middle plays (Express Quote, Builder Conversion) are the BU-specific structural moves. Spend Map and Call Coaching land in Sprints 3–4 because they require more setup but bring the largest dollar handles. Sprint length is intentionally flexible — leadership defines it (2 weeks, 3 weeks, monthly) based on how fast the org wants to move.

Sprint 1Quick wins ship
Sprint 2Pipeline visible
Sprint 3BU lifts land
Sprint 4P&L compounds

Stand up the no-regret plays

  • Play 01 — Marwin Traffic Recovery launched, first identified accounts in pipeline
  • Play 02 — SSI lead scoring live; first weekly ad-spend tuner report
  • Play 03 — SSI re-engagement agent in pilot with 3 reps
  • Play 09 — Universal AI Chat scoping + pilot brand selected (SSI or Marwin)
  • Spend Map data extraction begins (long-pole)

Texas + the build sprints

  • Play 04 — Texas Briefing Engine in pilot with new reps
  • Play 05 — Express Quote scoping with Millwork 360 + Synergetics
  • Play 06 — Summit builder-data ingestion + scoring model
  • Play 09 — Chat live on pilot brand (web + IG + FB + SMS); template ready for BU 2–5
  • Plays 01–03 hit first P&L signal; published to leadership weekly

BU lifts go live

  • Play 05 — Express Quote MVP live for top-N interior door SKUs
  • Play 06 — Summit ranked target list shipping weekly to TX reps
  • Play 07 — Spend Map dashboard live for John; first dup-vendor categories surfaced
  • Play 09 — Chat rolled out to remaining brands; cross-channel intent funnel reporting weekly
  • Plays 01–04 in run-rate; reporting cadence established

Coaching + compounding

  • Play 08 — SSI Real-Time Coaching in pilot with bottom-half reps
  • Play 07 — First negotiated savings under contract
  • 90-day board check-in with the four-metric scoreboard per play
  • Decision gate on extending Express Quote pattern to Marwin and Summit
Program options

Three ways to run this with us.

The roadmap scales to how much leverage MBP wants from this engagement. Every tier includes the "combined intel" enablement layer — the move from individuals quietly using Copilot to a platform where the team's AI patterns compound. The tiers differ in how many of the nine plays we run in parallel and how deeply embedded the program is.

Foundational
$5,000/mo
Retainer · 1 focus area

One focused initiative plus org-wide enablement. Designed to move MBP from individual AI usage to combined intel that compounds — without overwhelming the BUs that aren't ready to run multiple builds in parallel yet.

  • One named initiative from the nine-play roadmap, scoped, shipped, and reported against its KPI scoreboard
  • Monthly AI enablement office hours for BU presidents and the corporate team
  • "Combined intel" pattern library — surface and share what individuals are doing in Copilot so the platform learns from itself
  • Microsoft Copilot adoption coaching plus an MBP-specific prompt library
  • Quarterly board-readable readout on the initiative and AI adoption
  • Live channel for quick questions and small asks between sessions
Accelerator
$7.5–12K/mo
Most common · up to 3 initiatives

Everything in Foundational, plus up to three additional initiatives running in parallel — sized to the leadership team's appetite for change in any given quarter and the BUs ready to move first.

  • Everything included in Foundational
  • Up to three additional initiatives from the nine-play roadmap, prioritized with leadership
  • Weekly working session per active initiative
  • Dedicated solution architect across the initiatives
  • Cross-initiative scoreboard for the executive team — one view, all plays, all KPIs
  • Build-vs-buy guidance and vendor evaluation support on the tooling each play needs
Platform
$15K+/mo
Tip of the spear · 5+ initiatives

Everything in Accelerator, plus five or more initiatives in parallel and an embedded program lead. The shape of the engagement when MBP wants to be — in Preston's words — the tip of the spear in AI within the building products space.

  • Everything included in Accelerator
  • Five or more initiatives running in parallel across the BUs
  • Embedded program lead for MBP Group
  • Custom build capacity — visitor-ID pipelines, lead-scoring models, chat integrations, scoring agents
  • Quarterly board-meeting prep with Preston directly
  • AI-forward positioning collateral for Validor and the next acquisition narrative
  • First call on emerging AI capabilities applied to specialty building products
A note on sizing. Sprint length is intentionally flexible across all three tiers — leadership defines whether a sprint is two weeks, three weeks, or monthly, based on the pace the org wants. Tiers can also start at one level and step up as the team's confidence grows; the typical path is Foundational for the first sprint, Accelerator from Sprint 2 onward once leadership has named the additional initiatives.
What's next

Three decisions for Preston and Alan, this month.

This roadmap is sized to be executable inside a single quarter and a half. To unlock it, the decisions are small in number but real in commitment — and they're all things only Preston and Alan can authorize.

DECISION · 01

Pick the first four plays to commission.

The recommendation: Play 01 (Marwin Traffic), Play 02 (SSI Scoring + Ads), Play 07 (Spend Map data work), and Play 09 (Universal AI Chat — pilot brand). Each is independent. Each has a clean dollar handle. And running them in parallel gives leadership four reading points by the end of Sprint 2 with the chat acting as the cross-BU CX layer the other plays plug into.

DECISION · 02

Name a per-BU AI champion for Marwin, SSI, Millwork, and Summit.

Each play needs one named owner inside the BU — typically a Director of Ops or Commercial — empowered to make the small day-to-day decisions and report against the four-metric scoreboard. Without it the platform stays the only point of accountability.

DECISION · 03

Set the 90-day board readout in the calendar today.

Treat this roadmap as a living artifact, not a one-time deliverable. The Sprint-4 board check-in is where results get presented against the per-play KPIs, where what's working scales, and where the next quarter's plays get commissioned. Putting the date on the calendar makes the rest real.

Limited · Intro Sprint
Special offer · 90-day launch sprint

A bundled kickoff at $5,000/mo for 90 days.

Five concrete deliverables wrapped into one engagement — designed to put proof points on the board before MBP commits to the full program. Bundled at well below the value of the individual parts because the goal is momentum.

All-in
$15K · 3 months
Format
90-day sprint

What's included

  • Universal AI Chat on 1–3 MBP brand sites — web, Instagram, Facebook, and SMS, with intent routing, lead capture, and the appropriate hand-offs to the right brand's sales team.
  • Anonymous website visitor identification & automated outreach — deployed on Marwin (or whichever brand leadership picks first), with AI-drafted first touches sent for rep approval.
  • Marwin website refresh — modernized design and structure, set up to feed the chat and visitor-identification pipelines and to support AI-driven discovery going forward.
  • Fractional AI support for leadership — monthly office hours for the executive team, plus a pool of working hours each month to help BUs get more out of Copilot and the other AI tools they're already using.
  • Spend Map design & scoping — we use the 90 days to plan the bigger Phase 1 build with John, so MBP is ready to commission it on Day 91 with the data work already mapped.
Wild Card · New Concept
Wild card idea · A bold swing

MBP Pro — a mobile app for the platform.

A new concept the leadership team responded to with energy in the catch-up — they called it potentially game-changing. One MBP-branded mobile app that serves every side of the platform (reps, dealers, builders, contractors) and quietly carries every other AI play in this roadmap into the pocket of the person at the moment of need. Early stage; included here as the strategic North Star for what platform AI looks like to the outside world.

Stage
Concept · loved in the room
Audience
Reps · Dealers · Builders · Contractors

What the app does

  • Request a quote from anywhere — Builders and contractors snap a photo of plans or specs, pick the brand, and get back an Express Quote in minutes. Powered by the same engine as Play 05, now in the field.
  • Real-time order status across all five BUs — Dealers and reps see where every open order sits without calling a CSR. Push updates when ready, scheduled, or shipped.
  • The catalog & spec sheets in your pocket — All five brands, fully searchable, with lead times, freight policy, and code-compliance notes attached.
  • "Ask MBP" assistant — The conversational layer from Play 09, mobile-first. "What's the lead time on a 4-foot oak spiral right now?" answered in seconds.
  • Two-tap re-order — Dealers re-order recurring SKUs in seconds; AI suggests when to order based on past cadence and current inventory.
  • Notifications that actually matter — Order ready, price changes, new product drops, low-stock alerts on the SKUs you actually buy. No noise.
MBP Group · 90–120 Day AI Roadmap · Prepared by Daniel Townsend, Copilot Innovations · May 2026
Sourced from the May 7 workshop output, the Alan Rupp catch-up, the BU leadership intro call, and the MBP Group AI Opportunity Report.