How Adapt uses Adapt for AI workflows

Adapt improves rapidly because every team at the company, from engineering to marketing and sales, uses it every day.

In this post, we share how different teams at Adapt actually use the product, which workflows have stuck, and what’s proven valuable in real work.

Fun fact: this post itself was researched and drafted with Adapt, using actual internal data. We prompted Adapt to:

“Analyze 30 days of Slack messages and BigQuery chat data to surface concrete examples of how each function uses Adapt in practice”.

What follows is what we’ve learned from being our own most demanding customer.

Engineering: code that reviews itself

The AI workflow of Sean Smith, CEO

Engineering's use of Adapt goes beyond what a standard coding assistant is capable of. Sean uses Adapt's cross-system access into the production environment of the company’s apps.

Debugging with full stack visibility

When a Next.js Server Action failed in production, the traditional workflow would have been to manually inspect logs, dig through cloud dashboards, reproduce the issue locally, then cross-reference docs and configs across multiple tools.

With Adapt, Sean shared the error and asked,

“What’s failing here?”

Adapt then reasoned through the codebase and the surrounding infrastructure. This includes the deployment config, runtime environment, and IAP settings. Adapt was able to surface misconfigurations that were not visible in the code alone.

Linear issues from RFCs

When an engineering RFC was ready for implementation, Sean asked Adapt to create Linear issues. Adapt downloaded the RFC, checked for existing duplicates, and created tickets. It then added them to the current sprint cycle.

Code review with codebase context

For pull requests, Adapt can review changes with access to the full codebase. Adapt created a detailed review covering backward compatibility, race conditions, and missing rollback handling.

Sean rates PRs on a 1-10 scale. Adapt gave that PR a 7/10 with six concrete issues to address.

AI for marketing: research at query speed

The AI workflow of Ashley McClelland, CMO

Marketing's use of Adapt centers on research that would otherwise take hours, and uses automations that would otherwise require engineering time.

Competitive intelligence in real-time

When new competitors launch, Adapt helps Ashley create useful intelligence that can be shared across the organization.

A recent example was with a competitive solution that appeared on the market. Within Slack, Ashley prompted:

"@Adapt how does [competitor] compare to Adapt? In 100 words or less."

Adapt crawled their website and Product Hunt page, analyzed their positioning ("Figma for Thinking" focused on collaborative chat), and identified the key differentiator: "[competitor] is for talking about work; Adapt actually does the work."

From that single Slack thread, Adapt generated:

• A competitive analysis
• A role-play as a skeptical founder evaluating both tools
• A PowerPoint battlecard slide for the sales deck

What would have been a half-day of competitive research was handled by Adapt in 10 minutes.

Automated daily competitive reports

Ashley also runs a daily competitive intelligence report that searches relevant subreddits to surface relevant conversations and pulls in the latest posts from competitor blogs. Adapt then relates these posts to what it knows about its own product, surfacing opportunities for competitive positioning and campaigns daily. This report shows up in the #competitors channel and often gets cross-functional attention and comments for its usefulness in tracking market dynamics.

AI SEO monitoring

Ashley runs two automated SEO reports with Adapt:

• Sitemap Audit: Crawls adapt.com, compares live pages to sitemap.xml, identifies anything that needs to be added to sitemap
• Google Index Status Check: Verifies that expected pages are indexed to avoid any surprises on key pages being de-indexed

Both reports are posted by Adapt to the #marketing Slack channel automatically.

Press briefing preparation

For press outreach, Ashley asked Adapt to draft a full Q&A document, pulling from the website, investment memo, and codebase.

The output was a comprehensive press-ready document covering company background, product details, funding, and roadmap, and was delivered as a Google doc for human editing.

Sales: pipeline intelligence that updates itself

The AI workflows of Jack Welsh & David Maldonado, sales professionals

Sales uses Adapt for keeping systems in sync and surfacing the right information at the right time.

Automated pilot health monitoring

David runs a scheduled task that syncs pilot usage data from BigQuery to HubSpot deals. The task pulls:

• Total chats per account
• Output tokens consumed
• Active users
• Days since last activity
• Messages in the last 7 days

When pilots go quiet, Adapt flags them immediately as an “At Risk” account. Before this automation, Sales would only discover a pilot was struggling when it was too late. With Adapt they know it on the same day.

Revenue projections from raw data

Jack analyzed a customer invoice by asking Adapt to reverse-engineer the billing methodology from BigQuery. Adapt calculated:

• COGS per query
• Total queries in period
• FY26 projections under different model scenarios

The output was a chart comparing revenue if usage doubled monthly under different model cost structures.

CRM always current

Adapt performs a daily sync from BigQuery to HubSpot, updated with product usage metrics from BigQuery. Deals show real engagement data, not just what was discussed on a call.

Leadership: cross-system visibility

The AI workflow of Jim Benton, co-founder

Adapt helps our leadership maintain visibility across every system, every day.

Daily company briefing

Jim runs a scheduled task that aggregates:

• HubSpot: Meetings created, deals closed this week
• Stripe: Current MRR vs. monthly goal
• Linear: Hiring updates
• Salesforce: Sales funnel metrics

Adapt creates one daily digest so he doesn’t waste time logging into three different dashboards.

Metric selection with real data

When evaluating northstar metrics, Jim asked Adapt to pull:

• 12 weeks of website visitor data from Google Analytics 4
• Usage patterns by account

Adapt delivered a set of comparison visualizations to inform the decision.

Product: building the docs Adapt deserves

Adapt is also used in product management, our most cross functional responsibility that touches multiple teams.

When reviewing the adapt.com website against actual product capabilities, Adapt identified a gap: the marketing promised "plug-and-play connectors" when the architecture actually delivered "powerful code execution that can connect to anything."

From that analysis:

• A new Platform page was drafted matching the site's existing components
• A 10-page docs sitemap was outlined in Stripe-style concise format
• A PR was opened to the docs repo with all new content

The docs now accurately reflect the product, including what's native OAuth vs. API-powered, how the sandbox works, and how secrets are managed.

What we've learned using Adapt

Here are the AI workflow patterns that work for our team:

1. Scheduled tasks compound value. The SEO audits, pilot health checks, and daily briefings run without anyone remembering to trigger them.
2. Cross-system queries are the unlock. The full power of Adapt is experienced when team members use it to ask a question that touches multiple systems simultaneously.
3. Non-developers use it for technical tasks. The CMO runs web crawls, while Sales ops syncs BigQuery to HubSpot, and no one had to write code to enable that.
4. Rich artifacts without a designer. Adapt helps create charts, tables, and downloadable files, leaving the product designers free to do higher leverage work.

What we're still figuring out:

• The right balance between proactive suggestions and waiting to be asked.
• How to surface scheduled task failures without creating alert fatigue.
• When to use sub-agents vs. sequential tool calls for complex queries.

By the numbers (last 30 days)

In most companies, AI usage is led by engineering and the most technical employees. We built Adapt as a cross-functional tool that can help the leadership and business functions in your company gain valuable knowledge from shared company context.

We’re seeing those patterns appear within our own organization:

The heaviest users aren't developers using AI in a silo to code. We use Adapt as a team to pull context from multiple systems to make collective decisions.

See these workflows in Adapt

If any of these workflows resonate, we'd love to show you how they work in practice. Get a demo.