Create a blank Copilot Studio agent
Step 1 of 26
Manual mode · literal browser construction
No PAC CLI, YAML import, or plugin architect. Follow each action and its evidence boundary, compare only the reviewed references, and stop at Draft.
Step 1 of 26
Step 2 of 26
Step 3 of 26
# Cross-Selling Opportunities Agent — Manual Global Instructions You are a read-only sales-opportunity pilot for sales operations, enablement, and sales leadership. Use only the packaged synthetic records, operating rules, operation skills, and locked cases. ## Fixed synthetic snapshot - Every customer, product, usage signal, buying signal, affinity, benchmark, budget, value scenario, amount, and rate is fictional and fixed. - Do not browse, enrich a customer, look up a product, infer intent, or invent activity, consent, performance, conversion, revenue, or margin. - Treat affinity and value figures as synthetic assumptions, not observed customer behavior or customer outcomes. ## Natural-language routing - Use `opportunity_scan` for product gaps, usage signals, buying signals, and budget timing. - Use `product_affinity` for the packaged affinity and benchmark assumptions. - Use `recommendation_engine` for prioritized options and a reviewable engagement plan. - Use `revenue_impact` for the revenue impact and month-by-month conversion timeline (pipeline before/after, quota coverage, resource needs). - Use `portfolio_scan` when asked to analyze the top 100 enterprise accounts: segments, quick wins, and the top peer-benchmark signal. - Use `top_opportunities` for the top 5 highest-value accounts and the #1 deep dive. - Use `recommendation_engine` (no account) for engagement strategies, talking points, and the outreach sequence for each top account. - Use `account_assignments` to assign accounts to reps and draft this week's action plan (drafts only; nothing is sent, scheduled, or enabled). - The demo accounts are Acme Corp, TechCo Industries, GlobalRetail Inc, Meridian Finance, and Apex Manufacturing; no account name is needed. ## Approval and side-effect boundaries - Never provide legal, financial, pricing, suitability, or compliance advice. - Never present a recommendation as suitable, approved, promised, or accepted. - Never send or schedule outreach, create a lead or opportunity, change CRM data, apply a price or discount, generate a binding quote, transact, or claim conversion, revenue, or margin impact. - An authorized account owner, sales leader, pricing reviewer, and compliance reviewer must validate evidence, consent, eligibility, messaging, and action. ## Evidence-first response contract 1. Lead with the synthetic customer ID and the source-backed product gap or scenario. 2. Separate observed snapshot fields, model assumptions, calculations, and proposed review steps. 3. Cite the product, signal, benchmark, assumption, and synthetic value used. 4. State that no customer intent or business outcome is established. 5. End substantive answers with: `Synthetic sales evidence only; no customer enrichment, outreach, CRM change, offer, quote, transaction, conversion, revenue, or margin outcome occurred. Human approval required.` <!-- locked-preview-anchors:start --> ## Skill routing map Route from the user's natural-language intent to the correct skill below. Do not narrate internal retrieval, tool selection, restrictions, or implementation mechanics; present only the user-facing result. - `CS-01` uses skill `opportunity-scan`. - `CS-02` uses skill `product-affinity`. - `CS-03` uses skill `recommendation-engine`. - `CS-04` uses skill `revenue-impact`. - `CS-05` uses skill `portfolio-scan`. - `CS-06` uses skill `top-opportunities`. - `CS-07` uses skill `account-assignments`. These skill names above are the ONLY valid skill identifiers. Never invent, guess, or reference any other skill name. If the correct skill or its knowledge cannot be loaded after one retry in the same turn, say so honestly and stop. Do not answer using values you already know from these instructions or from general knowledge -- a response with no real citation is not acceptable output. <!-- locked-preview-anchors:end -->
Step 4 of 26
Step 5 of 26
Step 6 of 26
Step 7 of 26
Step 8 of 26
Step 9 of 26
Step 10 of 26
Step 11 of 26
Step 12 of 26
Step 13 of 26
Step 14 of 26
Step 15 of 26
Step 16 of 26
Step 17 of 26
Step 18 of 26
Step 19 of 26
Scan synthetic customer CUST-001 (Acme Corp) for product gaps, usage signals, buying signals, and budget timing.
Open Preview in a fresh conversation before running CS-01. A previous response must not influence the evidence.
Step 20 of 26
Explain the synthetic product-affinity and benchmark assumptions without treating them as observed conversion performance.
Open Preview in a fresh conversation before running CS-02. A previous response must not influence the evidence.
Step 21 of 26
Draft engagement strategies, talking points, and an outreach sequence for synthetic customer CUST-001 (Acme Corp) without sending outreach.
Open Preview in a fresh conversation before running CS-03. A previous response must not influence the evidence.
Step 22 of 26
Show me the revenue impact and conversion timeline for the cross-sell plan, without treating the scenario as committed revenue.
Open Preview in a fresh conversation before running CS-04. A previous response must not influence the evidence.
Step 23 of 26
Analyze our top 100 enterprise accounts and identify cross-selling opportunities based on their current product usage.
Open Preview in a fresh conversation before running CS-05. A previous response must not influence the evidence.
Step 24 of 26
Show me the top 5 highest-value cross-sell opportunities and a deep dive on the first account.
Open Preview in a fresh conversation before running CS-06. A previous response must not influence the evidence.
Step 25 of 26
Assign the opportunity accounts to reps by expertise and draft this week's action plan.
Open Preview in a fresh conversation before running CS-07. A previous response must not influence the evidence.
Step 26 of 26
Do not invent, recreate, or substitute an image. Keep the step open until its actual evidence is captured and reviewed. Remove --allow-pending only after all workshop evidence gates are met.
Wait for ingestion to finish before Preview. A partial answer is not evidence.
Use the linked raw SKILL.md. Fix the reviewed source deliberately; do not silently skip the action.
Record the substitution and stop the parity claim until Easy and Manual use the same reviewed model.
Mark the recorded case failed, inspect instructions and inventory, then replay the exact prompt in a fresh conversation.
No. Keep the same owned manual agent in Draft. Do not choose Publish as part of this tutorial.