# Background Agents: Setup Guide > Learn how to set up Spiich Background Agents - autonomous AI agents that handle lead generation, contact enrichment, deal qualification, and pipeline reviews on autopilot. - Source: https://spiich.ai/articles/background-agents-setup-guide - Author: Filip Hessle - Category: Getting Started - Published: 2026-03-23 - Read time: 10 min read ## Introduction Spiich Background Agents are autonomous AI agents - running on a daily schedule or changes in your CRM. They handle repetitive sales tasks like lead generation, contact enrichment, deal qualification, and pipeline reviews. You can set them up in two ways: by telling Spiich what you need in plain language, or by configuring them manually in [**Settings → Background Agents**](https://app.spiich.ai/settings/background-agents). Read more about Background Agents [here](https://spiich.ai/product/background-agents). ## What Background Agents Can Do Spiich Background Agents have access to the same tools as your sales reps. They can: - **Read and write to your CRM** — organizations, people, deals, tasks, and notes - **Research the web** — company websites, news, job postings, and public data - **Find new leads** from prospecting databases and web sources - **Enrich contacts** with emails, phone numbers, and LinkedIn profiles - **Qualify and score** companies against your ICP using real signals - **Send you email reports** with findings and recommended actions ## Prerequisites To create a Background Agent, ensure that: - You are a Spiich admin user - You have [sufficient credits](https://app.spiich.ai/settings/background-agents) to cover background agent consumption ## Trigger Types Background Agents need a trigger - a condition that tells them when to run. There are two types. ### Event-Driven Event-driven agents fire instantly on changes in your CRM, such as a record creation or update. Use event-driven triggers when you want background agents to automatically run on an action, like qualifying new leads or enriching a new contact. ### Scheduled Scheduled agents run on a set interval - every day, every other day, or once a week. Use them for tasks that occur on a regular cadence, like lead generation, pipeline reviews, or recurring reports. ## Credits Background Agents consume credits when they run. Credit usage depends on the actions the agent takes — such as enriching contacts, updating CRM records, or generating reports. Credits are allocated monthly. You can view your credit balance, usage breakdown, and per-action costs in [Settings → Billing](https://app.spiich.ai/settings/billing). ## Creating Background Agents To create a background agent, just tell Spiich what you need. ### Chat Setup _Recommended_ Select a use case below to see the full setup sequence example. #### Lead Generation & Buyer Discovery _Daily prospecting with buyer discovery_ ##### Prerequisites Before you begin, make sure the following CRM attributes are in place: - **Source** on the deal object, set as a multi-select field with Spiich Background Agent as an option - **ICP fit score** on the company object, set as a 1-5 rating field. If a similar field already exists, you may use it instead, just ensure to update the example prompts below accordingly. Note: Due to CRM limitations, it may take up to 1 hour for the agent to see newly added fields. Set the attributes first and wait approximately 1 hour before following the message sequence below. ##### Step 1: Understand the ICP Send the message below to Spiich. It will analyze your ICP and break it down into core segments based on region, size, industry, and company archetype. **Step 1 - ICP analysis** ``` Please deeply assess my ICP based on what you know about me and what you can see on our website. Break this down into core segments based on region, size, industry, and company archetype. Develop a persona we should approach to maximise the probability of selling our solution; include role, experience, focus, and other attributes. Feel free to search my CRM for won and lost deals to understand our customer base better. ``` Reiterate until the agent is fully aligned with your ICP and target segment. ##### Step 2: Lead sourcing strategy Next, use the message below to draft a lead sourcing strategy for the agent to follow. **Step 2 - Lead sourcing strategy** ``` Based on our ICP and buyer persona breakdown, please create a comprehensive strategy for how new leads could be identified through web research and structured prospecting databases. This should include signals to track. Ensure it's possible to find through web research or prospecting databases. It's okay to rely on proxies. ``` Continue once you are satisfied with the lead sourcing strategy. You may incorporate your existing lead sourcing strategies and relevant websites to scrape. ##### Step 3: Create the lead generation agent Based on ICP and lead sourcing strategy, the agent will create a lead gen agent. Adapt the prompt based on desired number of leads and CRM output preferences. **Step 3 - Create lead gen agent** ``` Based on the ICP and buyer persona breakdown, and the strategy for lead generation, please create a background agent to generate me 10 leads per day and add them to my CRM. You must implement an intelligent rotation to avoid finding the same companies over and over again, for this you can use modulus based on the date (ex. search for companies in different regions, of different sizes, or of different archetypes based on date). For each identified company, you must search the CRM to avoid duplicates, add it if it doesn't already exist, and fully enrich the record. Make sure to qualify the company and set the "ICP fit score". For every new company added, also create a deal in lead stage with "Source" set to "Spiich Background Agent". ``` **Example prompt output:** New deal and enriched organisation objects in your CRM - daily. ##### Step 4: Create the buyer discovery agent Finally, we will create a buyer discovery agent to identify the most relevant decision-makers for each added deal. Adapt the prompt based on desired number of decision-makers and CRM output preferences. **Step 4 - Create buyer discovery agent** ``` Please create a second agent that triggers on new deals being added with "Source" = "Spiich Background Agent" to identify who I should approach at the company. For each deal, add a note to the deal object with the 5 most relevant people to approach at the company. Each person should have a name, LinkedIn URL, title, and 1 sentence why they make sense to approach. Ensure you follow the priorly identified buyer persona. ``` **Example prompt output:** A note on each new deal with the 5 most relevant decision-makers to approach. #### Deal Revival _Identify deals losing momentum and drive re-engagement_ ##### Step 1: Define what "stalling" means for your pipeline Send the following message to Spiich: **Step 1 - Stalling criteria analysis** ``` Please analyze my current pipeline to understand what a "stalling deal" looks like for us. Look at my deals across all stages and identify: - The average time deals spend in each stage - Which stages deals tend to get stuck in - Common patterns in deals that eventually went to "Lost" vs "Won" - What signals indicate a deal is losing momentum (e.g. no recent emails, no meetings scheduled, no stage change, no notes added) Based on this, propose a clear definition of "stalling" for each deal stage, including specific time thresholds and inactivity signals. ``` Review the agent's analysis and adjust the thresholds to match your sales cycle. For instance, you may decide that 14 days without activity in "In Progress" is stalling, but 7 days in "Negotiation" is a red flag. ##### Step 2: Define the re-engagement playbook Based on the stalling criteria, the agent will create a re-engagement playbook. Adapt the prompt based on desired re-engagement actions and CRM output preferences. **Step 2 - Re-engagement recommendations** ``` Based on the stalling criteria we just defined, please create a re-engagement playbook. For each deal stage, define: - What the recommended next action is when a deal stalls (e.g. send a follow-up, loop in a champion, offer a trial extension, escalate internally) - How urgent the action is (critical, high, medium) - What context the agent should surface to help me act quickly (last interaction, deal value, key contact, blockers mentioned in notes) Keep it practical — I want actionable recommendations, not generic advice. ``` Iterate until the playbook matches your desired re-engagement actions. ##### Step 3: Create the deal revival agent Finally, create a deal revival to identify stalling deals in your pipeline and create a daily email report with the stalling deals, context, and recommended actions. **Step 3 - Create the agent** ``` Please create a background agent that runs daily to identify stalling deals in my pipeline. It should: 1. Check all deals in active stages for signs of stalling based on the criteria we defined 2. For each potentially stalling deal, check the last email interaction, last calendar event, last note added, and last meeting recording 3. Classify each deal as "Critical" (severely stalled), "At Risk" (approaching stall), or "Healthy" 4. For Critical and At Risk deals, include the recommended next action from our re-engagement playbook 5. Send me an email report every morning with the stalling deals, sorted by urgency Skip deals in Lead, Won, Lost, or On Hold stages. Only flag deals that meet the stalling thresholds we agreed on. ``` You may iterate in the chat until you are fully satsified with your agent setup. **Example prompt output:** A daily email report with your stalling deals, context, and recommended actions. #### Web Scraping → CRM _Scrape a web source for leads and add them to your CRM_ ##### Prerequisites Before you begin, make sure the following CRM attributes are in place: - **Source** on the deal object, set as a multi-select field with Spiich Background Agent as an option - **ICP fit score** on the company object, set as a 1-5 rating field. If a similar field already exists, you may use it instead, just ensure to update the example prompts below accordingly. Note: Due to CRM limitations, it may take up to 1 hour for the agent to see newly added fields. Set the attributes first and wait approximately 1 hour before following the message sequence below. ##### Step 1: Define source and what to extract Ask the agent to review the source and propose a data extraction plan. Adapt the prompt based on the source and your needs. **Step 1 - Source & data mapping** ``` I want to create a background agent that scrapes [SOURCE URLS] on a daily basis and adds relevant companies and contacts to my CRM. Please visit the page, understand its structure, and propose: - What data points to extract (company name, domain, contact person, email, phone, etc.) - How to handle page pagination (if applicable) - What filters to apply to only capture relevant entries (e.g. date, sector, region) - How to identify which entries are new since the last run ``` Review the result and reiterate until you are satisfied with the data extraction plan. ##### Step 2: Define qualification criteria Next, set up qualification criteria for the scraped data. Adapt the prompt based on your ICP and the source. **Step 2 - Qualification & filtering rules** ``` Before adding scraped companies to my CRM, I want to filter them. Please propose qualification criteria based on my ICP and the data available on this source. Consider: - Company size or revenue thresholds (and how to verify - Industry or sector filters - Geography filters - Potential pain points - Company archetype - Any disqualifying signals (e.g. company already in CRM, too small, wrong sector) For each criterion, specify: where to find the data, what the threshold is, and what to do if the data is unavailable (skip or estimate). ``` ##### Step 3: Define CRM output structure Define CRM output structure for the scraped data. **Step 3 - CRM mapping & deduplication** ``` For each qualified company from the source, define exactly what should be created in my CRM: 1. **Organization record** — what fields to populate (name, domain, description, location, etc.) 2. **Person record** — what contact fields to populate (name, email, phone, linked company) 3. **Deal record** — should a deal be created? If so, in what stage and with what source tag? 4. **List entry** — should the person or company be added to a specific list? If so, which fields to set? 5. **Notes** — should a note be attached with context from the source? What should it contain? Also define rules to avoid duplicates. ``` Iterate until it matches your CRM setup and preferences. ##### Step 4: Create the scraping agent Finally, create the scraping agent based on the previously defined plan. **Step 4 — Create the agent** ``` Based on everything we've defined — the source, extraction plan, qualification criteria, and CRM output structure — please create a background agent that runs daily. The agent must: 1. Scrape [SOURCE URLS], handling pagination as needed, only processing entries from yesterday 2. Extract the data points we defined and filter for previously defined sector, region, and qualification criteria 3. Deduplicate within the run (same company appearing in multiple entries) 4. For each company: check CRM for duplicates before creating 5. Qualify each company against our criteria (verify via web search) 6. For qualifying companies: create the organization, person, deal, list entry, and notes as we defined 7. Skip companies that don't meet criteria or already exist with contacts ``` Review the generated agent prompt carefully to ensure it matches your needs. **Example prompt output:** New organization, person, deal, and list records in your CRM - daily. ### Settings Page _Advanced_ To set up a Background Agent manually, go to [**Settings → Background Agents**] and click **Create New Agent**. Each agent has five fields: 1. **Trigger** — Event-based (fires on CRM changes) or Scheduled (daily/weekly) 2. **Name** — A descriptive label (e.g., "Weekly Follow-Up Tracker") 3. **Objective** — What the agent should accomplish in 1-2 sentences 4. **Process** — Step-by-step instructions. Use numbered lists and markdown formatting for clarity. 5. **Output** — What the agent produces: CRM updates, notes, tasks, email reports, or a combination For most users, the chat setup is faster and handles these fields automatically. The settings page is available for full manual control over each field, and adjustments of existing Background Agents. ## Editing and Managing Agents Background Agents can be edited directly through the chat or by going to [**Settings → Background Agents**](https://app.spiich.ai/settings/background-agents) and selecting the edit button on the agent you wish to edit. On this page, you can also pause, manually trigger, or delete agents along with seeing the number of used and remaining credits, and recent agent runs. Pausing, triggering, and deleting agents cannot be done through the chat. ## FAQ ### Why is my Background Agent not running? Verify that you have an adequate number of credits remaining. You may troubleshoot your agent setup by sending your trigger, objective, process, and output field values in the Spiich chat. If the issue remains, reach out and we will help you resolve your issue. See [how to troubleshoot a background agent](https://spiich.ai/help/troubleshooting/agent-problems). ### How many credits does a Background Agent use? Every agent run and action consumes credits in accordance with the [Credit Table](https://app.spiich.ai/settings/billing). You can view your remaining credits at [Settings → Background Agents](https://app.spiich.ai/settings/billing). See [how credits work](https://spiich.ai/help/workspace/credits-and-usage). ### Can I run multiple Background Agents at the same time? Yes, you can create and run as many agents as you need wihin reasonable limits. See [how background agents work](https://spiich.ai/help/background-agents/background-agents-overview). ### When do scheduled agents run? Scheduled agents run at 05:00 CET (Stockholm time), ensuring that their tasks are completed before your workday starts. See [how triggers and schedules work](https://spiich.ai/help/background-agents/agent-triggers). ### Can I test an agent before it goes live? Yes. From [Settings → Background Agents](https://app.spiich.ai/settings/background-agents), you can manually trigger any agent using the trigger button. This lets you verify the agent's behaviour before it runs on its regular schedule or event trigger. See [how to test an agent before it writes anything](https://spiich.ai/help/background-agents/agent-permissions). ### Why is my agent not seeing newly added CRM fields? Due to CRM synchronisation limitations, it can take up to 1 hour for new or modified fields to become visible to Background Agents. If you recently added a field, wait approximately 1 hour and try again. See [how to fix CRM sync problems](https://spiich.ai/help/troubleshooting/crm-sync-problems). ### Can Background Agents send emails on my behalf? No. Background Agents cannot send emails as you. They can send email reports and summaries to your inbox, but they cannot compose or send outbound emails to prospects or customers. See [what Spiich will and will not do on its own](https://spiich.ai/help/chat-and-voice/what-spiich-does-on-its-own). ### How do I stop or pause an agent? Go to [**Settings → Background Agents**](https://app.spiich.ai/settings/background-agents), find the agent you want to pause, and click the pause button. Paused agents will not run until you resume them. You can also delete an agent entirely from the same page. Pausing, triggering, and deleting cannot be done through the chat. See [how to manage agent runs](https://spiich.ai/help/background-agents/monitor-agent-runs). ## Stop Typing. Start Closing. - [Book a demo](https://spiich.ai/contact) - [Estimate the hours](https://spiich.ai/value-calculator)