Goodwill tripled its personalised outreach and freed up 32 hours a month.

Case studies

Goodwill AI Sprint Four weeks

The firm

Goodwill is a UK social media agency helping hotels and hospitality brands improve their content and online presence. Its work spans finding new clients, researching competitors, creating social content and managing paid and organic marketing.

The problem

Much of that work was manual and fragmented.

Finding a new prospect meant moving between Instagram, Google and spreadsheets. Competitor research was time-consuming enough to sometimes be skipped. Creating content meant repeatedly turning photographs and ideas into posts by hand.

Planning, publishing and reviewing marketing involved more tools and more duplicated effort.

AI could help with isolated parts, but it was used in the browser, a chat at a time, and was not connected to Goodwill’s processes, information or creative standards.

What we did

We began by mapping how Goodwill worked and measuring the time spent across its recurring activities. That identified five workflows where AI could create the most useful capacity.

We created a local second brain — a folder of plain files holding Goodwill’s processes, preferences, examples and tone of voice, which Goodwill can read and edit itself. We then connected the Claude Cowork desktop app to the tools and information the business already used, so it could work in them directly.

Over the four-week Sprint, Will tested each workflow himself using Goodwill’s real work, with guidance on choosing the right approach, reviewing the output and improving the process.

01 Client acquisition

Starts with one example hotel, finds similar businesses, qualifies and researches each prospect, updates the pipeline and drafts personalised email and Instagram outreach. Goodwill reviews the prospects and every message before anything is sent.

02 Competitor analysis

Collects recent activity from selected competitor accounts and turns it into a structured report covering content patterns, engagement and opportunities.

03 Social content creation

Turns a short brief and photographs from a shoot into a story arc and branded Canva content, ready for final creative review.

04 Paid marketing

Develops campaign options, copy and creative, then reviews performance and recommends what to pause, change or test next.

05 Organic marketing

Creates a weekly content plan, prepares and schedules approved posts, then uses the results to improve the next plan.

What changed

20 → 60

Personalised pitches prepared each week

32 hours

Saved each month across the five workflows

5

Workflows redesigned and tested

End-of-sprint testing showed the five workflows releasing 32 hours each month, the equivalent of four working days.

The client-acquisition workflow took Goodwill from 20 to 60 personalised pitches each week, without replacing the judgement that made the outreach relevant.

Goodwill also left with the context, connected tools and working methods needed to adjust the workflows and build new ones rather than depending on an external developer.

In Will’s words

“AI had always been useful for the odd task. After the sprint, it was connected to how Goodwill actually works. We could find three times as many prospects, turn shoots into finished content much faster, and still keep control of what went out.”

Will Thompson

Founder, Goodwill

What could your most important process look like four weeks from now?

Book a free AI review
Book a free AI review