PE / Portfolio Leadership
Portfolio Commander + Weekly Flash; oversight skills for partner packs, risk, and investment decisions.
How I staff Agentic AI Bots across the firm: a lean set of manager bots, AI Bot Jobs as AI Specialist skills, bridge first for continuity and readiness-gated connections, then migrate preferred platforms when margin or control earns the move.
How I use Agentic AI Bots across the firm: from PE Leadership down to marketing services and verified results.
How I staff Agentic AI Bots across the PE firm, ready for partner use.
Portfolio Commander + Weekly Flash; oversight skills for partner packs, risk, and investment decisions.
Deal Factory lead + diligence skills; Value Creation PMO + diagnostic/redesign skills.
Finance Lead over 4.x skills; COE Lead over shared connectors, vault, measurement.
Hub Manager, Client Success, Client Strategy: protect renewals and earnouts.
Department manager Agentic AI Bots (brand, media, creative, copy, web…) + AI Specialist skills.
Traffic, measurement, Delivery QA gate; agency BD and people admin.
Portfolio Commander, Weekly Flash (weekly exception pulse), Agency Hub, Client Success, and Delivery QA sit near the top of the operating path.
Deal Factory, Finance, the Center of Excellence, and marketing department managers fill out the rest of the firm.
About 8–15 manager Agentic AI Bots are the lean always-on runtime; this page shows 140 of 147 catalog AI Bot Jobs as skills under them (Fund/LP omitted).
At close, I preserve operating continuity and begin connecting priority agency systems where access, security, and data readiness permit. Validated improvements land in stages. Preferred platforms come later, only when margin or control earns the move.
Agency tools stay live. I begin wiring priority systems only where access, security, and data readiness permit. Readiness gates and fallbacks keep the promise honest: no “every connector on day one” claim. Early gains come from the workflows that clear the gate, not from a blanket cutover or a blanket integration pledge.
Bridge first keeps value flowing until a preferred platform clearly wins on margin, control, or speed. Platform cutover is covered under Acquisitions and Shared Standards (see callout below). Client-owned systems the agency operates stay on the client’s standards unless the client chooses otherwise. My portfolio platform standard does not force a client CRM, ads setup, or infrastructure change.
Migrate a category when the preferred platform clearly wins. Bridge first keeps continuity until then.
Same bridge logic can apply to email, project tools, accounting, ads platforms, analytics, file stores, chat, and other agency SaaS, each behind its own readiness check. Story for CRM only: I buy an agency on HubSpot. They keep HubSpot. When access and data readiness clear, Agentic AI Bots connect and tighten lead management, time-to-respond, communication, and the sales process in stages. Later, peer agencies on other CRMs can land on the same HubSpot standard if the move earns itself. If the agency operates a client’s CRM or ads account, that stays on the client’s stack unless the client chooses a change.
What each capability is, who decides, where records live, how access is separated, and how mature the material pieces are.
Always-on owners for a functional area or hub seat (portfolio, hub, department, finance). Rule: a manager owns ongoing priorities and exceptions; a routine only performs a recurring task. Studio leadership talks to a few managers, not to every specialist.
Reusable recipes (diligence, Google Ads, design, research) started under a manager when scoped work exists. A skill is an implementation component. It is not proof of specialist-level performance until that workflow clears maturity and evidence.
Standing jobs on a clock: flash pack generation, covenant watch, renewal calendar, seat inventory, tracking QA. Routines write into the same systems managers already read. Weekly Flash (1.2) is a manager because it owns the weekly exception queue; the pack file is produced by its routine.
The org chart shows accountability and escalation. It does not mean every task climbs every management layer. Routine work takes the shortest reliable path: skill or routine → owning manager → system of record. Managers prioritize, delegate, review exceptions, resolve conflicts, and escalate. They do not rubber-stamp every token.
Prioritize the queue · start or stop AI Specialist skills · review exceptions · escalate consequential items to the named human · report status into the shared pack. Not: silent campaign edits, wires, employment decisions, or contract commits.
AI can recommend, prepare, and execute routine actions inside approved rails. Humans stay accountable for consequential decisions.
Changing campaign budgets above agreed bands · issuing credits · employment decisions · accepting contractual commitments · capital allocation, investment decisions, and wires · releasing client work that failed Delivery QA · legal interpretation of client or contractual obligations.
Draft packs and variance notes · queue triage · scheduled health checks · pull reports through approved connectors · propose next actions into a manager queue for human or manager review.
Paid Media Manager AI Bot + Google Ads AI Specialist skill on client Acme.
| Action | AI may? | Approval | Accountable human |
|---|---|---|---|
| Draft a bid/budget change inside a pre-approved band | Yes · prepare | Manager queue review | Paid Media lead (agency) |
| Push that change live in the ads account | Yes · only if band + client rails already approved | Logged execute under standing approval | Paid Media lead |
| Raise budget above the agreed band | No · propose only | Human required | Paid Media lead + client owner |
| Publish a landing page after Delivery QA passes | Yes · release | Delivery QA gate green | Hub / Delivery lead |
| Publish if Delivery QA failed or was skipped | No | Blocked | Hub / Delivery lead |
A thread is a durable work lane with a standing brief, the right connectors, and a clear owner (for example Deal: Agency Name, Hub: Dallas, or Client: Acme retainers). It is not a one-off chat ask that disappears.
Diligence packs land in the deal thread; Integration Planner and Hub pick up the same artifacts. No re-brief from scratch.
Each manager owns one exception queue. Skills post results there. Portfolio watches the book; hubs run the shop.
Open for a deal, hub, or client program; archive when the scorecard or engagement closes. Memory stays in the system of record.
Back-office threads never talk to clients. Delivery threads never release payroll or wires.
For each major capability I name: what information it needs, where the authoritative copy lives, what it produces, and who acts next. Bots read and write through connectors into those systems. Chat is a workbench, not the filing cabinet. Two bots must not invent conflicting records of the same fact.
Finance books, project/delivery board, CRM, email/chat, knowledge base, brand vault, ads platforms, analytics, HR/payroll, time and billing, password vault. Continuity first at close; connect what clears readiness.
API connectors when available. Export / pull paths when that is all you get. Attended computer use on a secured VM when neither works well yet, with approvals on money, clients, or access. Credentials live in the vault.
Output that passed the named check or Delivery QA gate for that workflow. It is not automatically a demonstrated business outcome (retention, EBITDA, or ROI) until that outcome is measured in the system of record.
Separate bot names or separate chat threads are not proof of access isolation. Portfolio companies, clients, and financial data sit behind scoped connectors, vault credentials, and least-privilege roles. What is enforced in a given deployment depends on connector readiness and security approval. What remains unvalidated is labeled as such in maturity below, not assumed away.
Agency A books and clients do not bleed into Agency B without an explicit shared-services path.
Client-operated systems stay on the client’s standards. Agency-owned tools are the Bridge default.
Payroll and wires stay on restricted rails; delivery bots do not hold those rights.
I label managers and flagship workflows, not every AI Bot Job. These labels are an operating view / build path, not audited claims that every workflow is fully proven in production. Where I say “early live use,” I mean bounded real-world trials on pieces of the workflow; full hub-scale evidence is still being gathered. Platform native ≠ PE-workflow proven.
| Flagship | Maturity | What was / was not tested |
|---|---|---|
| Portfolio Commander (1.1) | Early live use | Book-view and exception surfacing in bounded trials; not yet a full multi-hub production evidence pack |
| Weekly Flash (1.2) | Early live use | Pulse pack + exception follow-up pattern in trials; not audited financials; cadence still tightening |
| Agency / Delivery Hub (6.1) | Early live use | Queue and promise tracking on limited hub scope; connector sets still vary by agency |
| Client Success (6.2) | Configured, unvalidated | Renewal / health rails designed; retention outcomes still human-owned and unproven as a bot system |
| Delivery QA (8.x) | Configured, unvalidated | Gate design exists; “verified” = checklist passed, not business ROI |
| Deal Factory Lead (2.x) | Requires development | Platform pieces exist; end-to-end diligence packs still need build + pilot evidence |
| Finance Lead (4.1) | Requires development | Entity close design; not a claim of controller replacement |
| COE / Standards (5.x) | Configured, unvalidated | Shared rails design; enforcement depends on connector and security approval |
Catalog vs deployment: Full catalog = 147 AI Bot Jobs. This Studio page shows 140 (Fund / LP section omitted). A given AI Bot PE Studio run activates about 8–15 manager AI Bots plus the skills live deals and named clients require. Incomplete or undecided implementations stay labeled; they do not read as verified deployments.
Which manager AI Bots are shared, what stays client-scoped, which skills fire, who approves, and where it is recorded.
| Step | Who | Where recorded |
|---|---|---|
| 1. Strategy asks for a bid tweak inside the approved band | 6.3 Client Strategy manager AI Bot | Thread: Client Acme |
| 2. Google Ads AI Specialist skill drafts the change | Skill under 7.2 Paid Media | Same thread + ads draft note |
| 3. Push live inside band (standing approval) | 7.2 Paid Media manager AI Bot | Ads platform (system of record) + thread log |
| 4. New landing module needs publish | Web skill → 8.4 Delivery QA | QA checklist; publish only if green |
| 5. Hub sees capacity / exception if QA fails | 6.1 Agency Hub | Hub exception queue (not “healthy”) |
| 6. Weekly pulse if margin or delivery heat moves | 1.2 Weekly Flash routine + manager | Flash pack in portfolio SoR |
Accountable humans: Paid Media lead (in-band ads execute), Hub / Delivery lead (publish after QA), Studio leadership (capital and out-of-band exceptions). Agency B’s Beta work never appears in Acme’s skill context.
Organize by AI Bot Job type and org level: from the AI Bot PE Studio down to marketing production inside each agency.
AI Bot PE Studio view of cash, retention, delivery heat, deals, and board-ready packs across owned agencies.
One clear view across owned agencies: cash, client retention, delivery trouble, and deal progress. Surfaces the decisions that matter when money or clients are at risk.
Owns the weekly operating pulse: the exception queue, follow-ups on red items, and the scheduled flash-pack routine before monthly books close. The pack itself is a routine; this manager AI Bot owns priorities and exceptions between flashes. Management view, not audited financials.
Keeps value-creation projects moving: raise margins after close from idea to proven result. Diagnostic and redesign work runs as AI Specialist skills under this manager.
Started when needed: covenant watch, concentration checks, board or partner drafts, investment decision packs, exit readiness. Weekly Flash is a manager AI Bot that owns the weekly exception pulse; its flash pack runs as a routine on the same command surface.
Help the business see a cash shortage before payments become due.
Find out which clients and delivery teams actually make money, and whether an apparent improvement survives a full cost check.
Reveal where the portfolio depends too heavily on one source of revenue or one vulnerable resource, even when the exposure is spread across several agencies.
Make delivery trouble visible across teams before missed deadlines and repeated errors become client losses.
Match available people and skills to work already promised, new sales and acquisition transitions, without booking the same person's time twice.
Check whether an acquisition is genuinely becoming stable, not simply whether its integration checklist looks complete.
Help one delivery hub benefit from relevant experience or client opportunities elsewhere in the group, while keeping confidential information restricted.
Turn approved financial and operating information into a clear package for directors and company investors to review and make decisions.
Help the partners choose between competing uses of money and understand the funding and ownership consequences.
Challenge a proposed acquisition before the firm commits money.
Keep the buyer's contractual purchase obligations accurate after closing, including payments that depend on later results.
Keep business risks tied to responsible people and deadlines, rather than scattered across contracts, emails and insurance documents.
Give leadership a portfolio-wide security picture without exposing every agency's systems and client records to everyone in the group.
Make the business easier for a future buyer to investigate by keeping evidence of clients, profit, rights and improvements organized throughout ownership.
Deal Factory for buys; Value Creation PMO for margin projects after close.
Runs the buy side: find targets, diligence (deep review before a buy), and the integration plan. Financial, ops, and client-risk specialists come in as AI Specialist skills on a live deal.
Diligence & integration skills (on a live deal):
Find businesses worth evaluating and distinguish a foundation agency from a smaller client-book addition that can fit into an existing operation.
Test how much of the seller's reported profit a buyer could reasonably retain after paying the real costs of ownership and transition.
Understand how the agency actually serves clients, which people and steps are essential, and whether the receiving team can take over safely.
Establish exactly which client relationships are being bought and what might prevent their revenue from continuing after the sale.
Find out whether the buyer can legally and practically use or connect the agency's systems before relying on automation or migration savings.
Turn diligence findings into a workable handover plan that protects clients, assigns a receiving team and separates business integration from software replacement.
Help the deal team present understandable purchase terms and track the seller's agreed responsibilities through closing and handover.
Keeps value-creation projects moving: raise margins after close from idea to proven result. Diagnostic and redesign work runs as AI Specialist skills under this manager.
Improvement skills under the PMO:
Find evidence-backed improvement opportunities in an acquired agency, including problems in billing, handoffs, staffing, tools and service delivery.
Turn a validated inefficiency into a simpler, costed way of working, rather than automating a broken process unchanged.
Build and introduce an approved operational improvement, including the training and testing needed for people to use it successfully.
Diligence artifacts feed the first 90-day plan. The same evidence that priced the deal becomes the improvement backlog, so post-close work starts from facts, not a new study.
Entity control plus one Center of Excellence stack every hub shares.
Runs the books for each legal entity: AP/AR, billing leaks, unused SaaS seats, and media cash recon. Skills do the catalog AI Bot Jobs; this lead keeps one close calendar and exception list.
Back-office skills:
Organize money owed by customers and money owed to suppliers.
Check that the agency bills for the work it promised and understands how much staff effort the work consumes.
Prepare accurate payroll inputs without giving a bot authority to change pay or release wages.
Stop paying for software that nobody needs while avoiding cancellations that break client work or destroy required records.
Make sure people and bots have only the access their work requires, and remove access when that work or employment ends.
Make signed promises usable in daily operations.
Keep client advertising money separate from money the agency can spend on itself, and explain differences between funding, ad spend and invoices.
Owns one shared Center of Excellence (COE) stack: connectors, password vault, measurement rules, playbooks, and access. Every hub uses the same rails.
Shared-stack skills:
Make financial reporting consistent across the group and avoid mistaking internal charges for new profit.
Give advertising bots a common, approved set of rules for budgets, allowed changes and launch checks.
Keep the correct, approved brand files and their usage rights easy to find, so teams do not publish outdated or unauthorized material.
Turn approved experience into repeatable instructions that other bot teams can follow without copying confidential client information.
Define what the agency sells and what it costs to deliver, so sales proposals and production expectations are consistent.
Decide which supplier and software arrangements are worth keeping, combining or replacing, using the full cost and operational risk rather than license price alone.
Make the same metric mean the same thing across reports, clients and companies, so comparisons support sound decisions.
Test whether bots and their system connections behave safely and reliably before and after release.
Move approved data and workflows to a new system without losing records, permissions or live work.
Hub operations, client success, strategy, research, and reporting for each delivery unit.
Runs day-to-day delivery for one hub (the team that owns a set of clients). Keeps promises, capacity, and exceptions in one place.
Protects the client relationship after close: renewals, health scores, handoffs, early warning if a client may leave. Critical when an earnout (seller payout tied to post-close results) depends on retention.
Turns client goals into one plan and clear assignments for department managers. Research and program work run as AI Specialist skills under that plan.
Program coordinators and research specialists start under the Client Strategy plan, scoped to a named client when the work exists.
Coordinate a business-to-business demand-generation program across departments, from attracting suitable prospects to handing qualified opportunities to sales.
Coordinate account-based marketing, which focuses marketing and sales effort on an approved list of target companies rather than a broad audience.
Make the many parts of a product or service launch ready together, rather than releasing marketing before the rest of the experience works.
Give strategy and production teams real customer and market evidence, rather than filling knowledge gaps with plausible-sounding assumptions.
Organize real customer and market evidence to answer a defined business question, separating findings from guesses.
Explain what competitors publicly say and show, helping the client understand relevant patterns without seeking private information.
Turn research into testable ideas about the audience, offer and message, without presenting an assumption as a confirmed customer preference.
Explain marketing results in a form the client can use, using validated evidence rather than simply copying charts or inventing reasons for changes.
Brand identity managers and paid-media AI Specialist skills.
Positioning, naming, messaging, and visual-identity direction so production teams share one brand brief.
Runs paid advertising so Google, Meta, and other channel specialists follow one budget and approval path. Channel work runs as AI Specialist skills scoped to a client.
Explain what the brand should stand for and how its products or sub-brands relate.
Develop names and a consistent messaging structure that express the approved positioning and can be reviewed for legal and commercial suitability.
Translate brand direction into practical visual rules that designers can apply consistently, without becoming a second graphic-production team.
Perform the Google Ads work assigned by the paid-media manager, from account review and campaign preparation to checks on approved changes.
Handle Facebook and Instagram advertising tasks while keeping campaign settings, creative tests and spending aligned with the approved client plan.
Identify product-data problems that prevent shopping ads from running correctly and get corrections made by the authorized product owner.
Translate the approved client strategy into practical channel budgets and tests, so campaigns answer useful business questions rather than generate uncoordinated experiments.
Prepare and review Microsoft advertising work without assuming a campaign copied from another platform is automatically ready to run.
Prepare LinkedIn advertising and lead forms for approved business audiences, including a usable handoff into the client's sales records.
Prepare and check TikTok advertising built around approved video assets, with attention to launch requirements and what the tests actually show.
Provide a video-advertising specialist within the Google account workflow, coordinating shared settings rather than competing with the Google Ads Bot.
Manage assigned advertising work on Amazon and retail-media platforms while leaving the client's store, inventory and fulfillment with their owners.
Prepare advertising bought through automated buying platforms and connected-TV systems, with clear checks on where ads run and what results can be supported.
Creative production and written-work skills under department managers.
Manages visual, video, and audio production so assets hit format, brief, and brand. Designers and editors are on-demand AI Specialist skills.
Runs written work (ads, pages, email, sales materials) so the message stays consistent. Writers are AI Specialist skills under this manager.
Turn an approved message and brand direction into usable graphics for ads and other marketing placements, not just a written suggestion for a designer.
Turn approved footage into usable marketing videos, including shorter versions and revisions required by the campaign.
Create animated graphic treatments that explain or emphasize the approved message, while keeping animation consistent with the brand and brief.
Organize the information and logistics needed for a photography shoot, while qualified people handle the camera, location and physical production.
Prepare and edit approved audio or podcast material into usable marketing assets while keeping voice, music and other usage rights visible.
Keep creative execution consistent across formats so individual designs and edits still express the same approved campaign idea.
Write the words that connect an advertisement to its destination page, including the headline, offer and action the visitor is asked to take.
Produce useful search-focused articles and page copy from the SEO team's brief, without taking over technical website work or promising rankings.
Write follow-up messages and sales materials matched to a customer's stage, while the CRM team controls recipients and sending.
Catch unclear language, contradictions and unsupported claims before written marketing reaches a client or the public.
Adapt approved content for a different language and market, preserving the intended meaning rather than relying on a literal word-for-word translation.
Organic search, social, web/CRO, and CRM/lifecycle skills.
Technical and content search work so organic growth follows one plan and measurement standard.
Social publishing and community response under brand and client rules.
Site builds, maintenance, and conversion-rate tests so pages work and changes are measured before they go live.
Email, lifecycle journeys, and revenue-ops handoffs so nurture and sales follow one system of record.
Find technical obstacles that stop search engines from accessing or understanding useful website content, then verify the website team's approved fixes.
Improve how individual pages answer relevant search questions, including their topics, headings and links to related content.
Keep a client's location information accurate and consistent in the business listings it is authorized to manage.
Help the client request and respond to real customer reviews, while identifying service problems that require a human response.
Observe how the client appears in permitted AI-search results and identify gaps in credible public information, without promising placement or citations.
Check the quality of completed search work and available link or mention evidence, rather than treating every new link as valuable.
Plan what the brand should publish on its own social channels and turn those plans into clear production briefs.
Move approved social content from draft to scheduled or published posts using tested account access and the correct channel settings.
Sort comments and messages so routine responses are prepared promptly and sensitive issues reach the right person.
Monitor permitted public discussion for relevant themes and reputational concerns, without treating every mention as a crisis or verified fact.
Build the approved website or landing-page change in a test environment, including the actual page structure and functions rather than only a design concept.
Turn a proposed website improvement into a controlled test of a desired customer action, with measurement agreed before results are interpreted.
Identify where people may struggle to understand a page or complete a task.
Find accessibility problems within the agreed testing scope and coordinate the checks that require knowledgeable human review.
Keep website updates controlled, tested and recoverable so routine maintenance does not unexpectedly break client-facing work.
Set up and maintain the client's CRM so records, pipeline stages, and ownership stay consistent and usable for sales follow-up.
Design the follow-up sequence after a prospect or customer takes an action, including when messages should start, stop or go to a person.
Make sure messages go to the right permitted audience and investigate problems reaching recipients.
Help qualified inquiries reach the correct salesperson with the information and materials needed for follow-up.
Support repeat purchases and voluntary customer advocacy through approved customer journeys, without manufacturing testimonials or consent.
Prepare a website-chat or messaging flow that asks permitted qualification questions and hands the conversation to a person when needed.
PR, influencer, affiliate, commerce, and events skills when a client program needs them.
Coordinate public relations: helping a client communicate credible news and ideas to journalists and other public audiences.
Find relevant journalists and publications and prepare suitable outreach, keeping relationship ownership with the human communications lead.
Prepare factual press and executive content so public communications are useful, consistent and supportable.
Help spokespeople prepare for interviews or sensitive public issues using verified facts and a clear escalation route.
Organize work with outside creators so audience fit, briefs, agreements, content rights and delivery are managed together.
Identify creators whose audience and work fit the client's objectives, while checking evidence and possible conflicts before human selection.
Make the creator's brief, content-use rights and disclosure obligations clear before work is produced or reused.
Track whether creators deliver the agreed work and how that work performs, without confusing organic results with paid promotion.
Run the coordination for a client's affiliate or customer-referral program, where approved participants introduce buyers under defined reward rules.
Prepare the recruitment and setup process for people or organizations joining a client's affiliate program, using approved eligibility and reward rules.
Calculate proposed affiliate or referral commissions from eligible transactions, taking account of agreed rules and reversals such as returns.
Investigate program-quality concerns and commission disputes fairly, using evidence rather than automatically penalizing participants.
Improve how a client's products are presented and kept consistent across its store and marketplaces, without taking over fulfillment or unauthorized pricing decisions.
Make product pages and storefront content accurate and complete using the client's official product facts.
Prepare product-presentation and promotion ideas within the client's commercial rules, without independently changing prices or inventory.
Keep product information consistent across connected systems through one approved source, avoiding competing updates by commerce and advertising bots.
Coordinate event and webinar preparation across marketing, registration, speakers and vendors, with qualified people responsible for physical and live operations.
Create the practical event plan: what should happen, when, by whom and with what approved resources.
Connect webinar registration, speaker readiness and follow-up so the event produces a usable attendee experience and permitted marketing handoff.
Coordinate the suppliers and dependencies behind an in-person experience, while qualified people own contracts, safety and on-site decisions.
Traffic, measurement, independent release QA, then agency BD and people admin.
Independent check before work goes to a client: brief met, brand and legal clear, tracking intact. Keeps drafting separate from the release decision.
Turns sold scope into schedules and assignments; reserves capacity so hubs do not overpromise.
Analytics, attribution, and dashboard QA so client results are comparable and trustworthy.
Make sure each incoming client request reaches one responsible team and is not duplicated across email notifications and work systems.
Check that important customer actions are recorded correctly, so campaign decisions are not based on missing or duplicate signals.
Connect permitted marketing records to CRM outcomes so the team can see which inquiries became qualified opportunities or sales.
Analyze whether an activity produced results beyond what would otherwise have happened.
Prepare statistical analysis of how marketing activity relates to business results across channels, with assumptions and uncertainty made visible.
Check that dashboards display the right numbers, periods and definitions before those reports guide decisions.
Firm-level BD, partner referrals, and people checklists: skills under the hub path, not client media production.
Help the agency win work it can profitably deliver, rather than promising custom services the team cannot support.
Manage referral relationships that bring customers to our agency.
Keep routine employee setup, training and administration organized without turning the bot into the employer or a personnel decision-maker.