To automate your sales pipeline with AI, you need more than a tool that writes emails or scores leads. Sales work rarely lives in one place: the account is in the CRM, the relationship is in email and meetings, the useful signal is on LinkedIn or a company website, and the next action is often still in someone’s head.

Strawberry is a browser with built-in AI agents. It gives you a familiar workspace where an agent can learn from your tabs, work across your apps, draw on meetings and files, research the public web, and remember useful context over time. Strawberry calls the agent working alongside you your companion.

That is different from both a chatbot that stops at an answer and a sales tool that automates one narrow slice of the job. Strawberry can connect the work across prospecting, account research, meetings, outreach, follow-up, CRM updates, and pipeline decisions. The context can carry from one step into the next instead of disappearing between tools.

Some of the most ambitious teams are already using Strawberry to automate time-consuming sales work and take on workflows that would previously have been too manual to attempt. You do not need to build that whole system on day one. This guide will show you what is possible, help you find a useful place to begin, and explain how a first win can grow into a much more powerful way to run sales.

Where to start with AI sales automation

Strawberry can help across the whole sales journey, but the best first task depends on what is happening now. You might find promising new customers, research an account and find a way in, get ready for tomorrow’s meeting, or send personalized outreach that sounds like you. If you already have something else in mind, start there.

Not sure which would be most useful? Ask Strawberry to recommend a first win from your open tabs, connected apps, and the context it already has. If there is not much to work from yet, it can suggest a useful app to connect—such as your email, calendar, or CRM—and explain what that would unlock. You do not need to connect everything or complete a setup project before doing useful work.

The focused Strawberry skills linked throughout this guide give your agent useful starting points for each job. They are not rigid sales methodologies: your companion can adapt them to the result you need and the way your team already works.

Work with Strawberry in small loops

You do not need a perfect prompt. Give Strawberry a real goal, let it take a first pass, and respond as you would to a capable teammate: say what works, what feels off, and what you want to try next. For bigger or more subjective work, begin with a small result—one account brief, a handful of prospects, or a few outreach messages—then build from there.

Your feedback becomes useful context. Over time, your companion can get to know your qualification bar, preferred sources, writing style, level of detail, CRM fields, and the way you like to review work. Once the two of you have shaped a repeatable approach, save what works as a skill so the next run starts further ahead.

Find the right customers and reach out like yourself

When you need new pipeline, Strawberry can work from a description of the customer, an ICP document, or examples of accounts you already win. It can search sources such as LinkedIn and company websites, explain why each prospect fits, and bring the result back to a sheet, CRM, dashboard, or chat.

Valuable sales signals often live outside standard databases. Strawberry can investigate niche directories, public websites, and older databases to uncover relevant evidence, while keeping the source behind each finding.

Try a small, varied set first and give your agent feedback before going wider. If you already have the companies, enrich the list instead. When you are ready, it can prepare and send personalized outreach in your voice, starting with representative messages and adding a sequence when follow-ups would help.

Research accounts and prepare for real conversations

A powerful first loop is research, preparation, conversation, and follow-through. Tell Strawberry which account or conversation matters. It can pull together the account record, recent thread, attendees, company pages, relevant LinkedIn context, prior notes, and fresh signals, then turn them into a concise brief with sources and useful questions.

A typical AI research flow searches the public web and summarizes the pages it can retrieve. Because Strawberry is an agentic browser, it can investigate more like you would: open the real source, follow links, search within sites, move across tabs, and keep digging when the first answer is incomplete. It can also work with relevant pages and apps you already use instead of relying only on whatever you paste into a chat.

Carry the context from the meeting into follow-up

The same context that helps before a meeting becomes even more useful afterwards. Strawberry can transcribe the conversation without adding a meeting bot. Once the transcript or notes are ready, your companion can prepare a sales debrief, follow-up, next actions, and suggested CRM updates—all grounded in what was actually said.

If the next step needs more than an email, your companion can also turn the account and meeting context into a customer-facing one-pager, deck, tailored page, or interactive asset built around the next decision.

You stay in control of what happens next. Reviewing a recap does not mean sending the email, sharing an asset, or changing the CRM. Once you approve the work, Strawberry can finish the follow-through without another round of copying.

Turn your pipeline into the next best move

Connect the CRM your team already uses and Strawberry can work with the records there instead of relying on an export or another round of copy-paste. It can combine deal stages with recent meetings, email activity, commitments, missing information, and fresh account signals.

That turns a pipeline review into more than a CRM summary. Strawberry can spot stalled deals, neglected follow-ups, and the actions most likely to move something forward, then propose updates for you to review.

Start with a review you can talk through together. Once the team trusts the signals and recommendations, your companion can prepare it regularly as a Routine.

Turn what works into a team advantage

A shared Sales companion can get to know the whole team: the customers you care about, the way you qualify accounts, the tone you use, the meetings you have had, and the skills that make the work run smoothly.

When someone improves a useful skill, teammates can benefit without rebuilding it from scratch. The companion becomes a living home for the team’s sales context and ways of working—something people can actually use, enjoy, and improve together, not another playbook sitting unread.

Grow one useful win into a sales operating system

Getting started is not a one-time checklist. After a useful account brief, connect preparation to follow-up. After a helpful pipeline review, improve how the CRM stays current. After a strong prospect sample, save what your companion learned about a good fit. Each useful loop can make the next one better.

As the pieces begin to work, connect them. A trusted workflow can become a skill, a shared team capability, or a Routine that runs at the right moment. The goal is not automation for its own sake. It is a sales operating system that gives the team more time, carries context forward, expands what the team can take on, and still leaves people in control.

If you are new to Strawberry, open the Getting Started with Sales skill and choose one useful piece of work. You can begin with the context already in front of you and build from there.

Official Strawberry skill

Get Started with Sales in Strawberry

Help the user see how their companion can support the whole sales journey, not just one isolated task. Briefly paint the bigger picture, explain what makes Strawberry different, and then help them choose one useful place to begin.

Paint the picture

Explain that Strawberry gives the user more than a chat box. Because Strawberry is a browser, the user and their companion can work together in the websites where sales happens. You can learn from their tabs, work inside websites, and bring together context from email, calendar, CRM, meetings, files, connected apps, and the public web. Your memory helps you get to know their customers, preferences, writing style, and way of working over time.

Show how the pieces can connect: finding customers can lead into account research, meeting prep, follow-up, CRM updates, and pipeline decisions. This can grow into a powerful way to run sales, but the best place to start is usually one real piece of work.

Keep this introduction brief, friendly, and inspiring. Do not turn it into a feature tour.

Find a good first win

Start with what the user wants to accomplish. If they already have a real task in mind, follow their lead. If they are unsure, offer a small set of useful starting points:

  • find promising new customers;
  • research an account and find a way in;
  • get ready for an upcoming meeting;
  • send personalized outreach that sounds like them;
  • recommend a first win from their open tabs, connected apps, and other available context;
  • tell you what they already have in mind.

Adapt the choices and wording to the conversation rather than treating this list as a script or a closed menu. If the user asks for a recommendation, look through their available context (e.g. apps, tabs, memory) and suggest no more than three strong options. If there is little useful context, briefly suggest a relevant app they could connect and explain what it would unlock. Do not turn this into a setup checklist or make connecting apps a requirement before helping.

Use a focused skill when it fits

First understand the result the user wants. If a focused Official Skill fits, read it and adapt it to their situation. If another collection owns the workflow, use that skill instead. If no existing skill fits, help using your normal capabilities rather than forcing the task into the closest workflow. A useful process that repeats can later become a custom skill.

  • Research an account and find a way in: read strawberry/sales/research-an-account. Use it whether or not a meeting is already scheduled. Start from the decision or conversation the research should help with.
  • Prepare for an upcoming sales conversation: read strawberry/sales/prepare-for-sales-meetings.
  • Turn a conversation into next steps: when a transcript or notes exist, read strawberry/sales/debrief-a-sales-meeting. If the meeting still needs to be captured, use strawberry/operations/transcribe-meetings-follow-up first.
  • Review the pipeline and choose what to do next: read strawberry/sales/review-sales-pipeline. Begin with a review the user can react to before suggesting a recurring version.
  • Find and qualify new customers: read strawberry/sales/find-new-customers. Start with a small, varied sample, invite feedback, and explain the tradeoff before widening the search.
  • Enrich a list the user already has: read strawberry/sales/enrich-lead-list.
  • Send personalized outreach that sounds like the user: read strawberry/sales/send-personalized-outreach. Agree on the channel and destination, show a few representative messages, and get approval before sending or enrollment.
  • Create a customer-facing sales asset: read strawberry/sales/create-a-sales-asset. Use the account and meeting context already available to recommend and create the simplest useful format for the next customer decision.
  • Clean or update CRM records: read strawberry/sales/keep-crm-updated.

When several options fit, recommend the best first one and no more than two alternatives. Explain why they fit the user's situation instead of listing every skill.

Teach through the work

Help the user become better at working with their companion while you work together:

  • Context and memory: explain that good context leads to better work. Use their feedback to learn what matters, and let them know that memory helps carry useful context into future chats.
  • Small loops: for subjective, expensive, or high-volume work, begin with a small result. Ask what they like and what they would change before doing more.
  • Skills: after feedback reveals a repeatable way of working, offer to save the useful criteria, tone, sources, fields, steps, and review points as a custom skill. Do not save the first result automatically.
  • Team collaboration: once a personal workflow works and would help others, explain how a shared Sales companion can carry the team's context, memory, and skills across the immediate team.
  • Routines: after the workflow is trusted and has a real trigger or cadence, offer to make it a Routine.

Connect the journey over time

After a useful result, suggest one natural next step when it would genuinely help. Examples:

  • account research -> meeting preparation -> follow-up -> approved CRM update;
  • small prospect sample -> feedback -> wider search -> enrichment;
  • account research and writing examples -> outreach sample -> approved sequence;
  • account research or meeting context -> customer-facing asset -> approved delivery;
  • pipeline review -> prioritized next actions -> later weekly Routine;
  • a workflow shaped through feedback -> custom skill -> shared Sales companion when useful.

When the user returns, build on what they already do with their companion. Help them improve a workflow, connect the next part of sales, or try something more ambitious instead of restarting onboarding.

Keep the user in control

Research, drafting, sending, and changing records are different steps. Never treat approval of one as approval of the next. Before sending messages, enrolling contacts, changing CRM records, or sharing private context, confirm the people, records, destination, and scope. Keep important sources and uncertainty visible when they affect a sales decision.

You have done the job when the user understands what is possible, completes a useful first workflow, and knows the most relevant way their companion can help next.