DS Web: What DS Web Solutions Does and How AI Agents Now Run the SEO Grunt Work

If you searched "ds web" and landed here, you're probably looking for DS Web Solutions (DSWS), a Canadian web development, software and SEO shop. Short version: we build and maintain websites and apps for local businesses. More and more of the repetitive work behind those sites, like reading search data, spotting ranking gaps and drafting content, is now done by AI agents we build ourselves. This article explains what we do and how that automation works, including the parts that still need a human.

What DS Web Solutions actually does

Most clients come to us for one of a handful of jobs. We don't pretend to do everything, and we'll tell you when a job isn't a fit.

• Web development: fast, maintainable business sites that are built to be found and to turn visitors into enquiries.
• Local SEO: Google Business Profile, local landing pages and content for service-area businesses.
• AI automation: agents and workflows that take over repetitive back-office and marketing tasks.
• Analytics and optimization: tracking that answers "what is working?" instead of producing dashboards nobody opens.
• Website malware removal: cleanup and hardening when a WordPress site has been compromised.

If you're hunting for a specific service, those pages are the fastest route. If you want to know how we think about the work, keep reading.

Why we automate SEO with agents

Local SEO is a lot of small, repeatable decisions. Which queries are showing impressions but earning no clicks? Which pages are stuck at position 7 to 12? Which topics have we already covered? A person can answer those questions, but doing it every week for many sites is tedious, and tedious work gets skipped.

Our own data shows the kind of gap an agent catches. The query "ds web" has 198 impressions, 1 click and an average position of 7.1 in Search Console. That's a page-one result that's barely being clicked. The usual causes are a vague title, a weak meta description, or a page that doesn't match what the searcher wanted. An agent can flag that pattern across every query on every site in minutes. Deciding what to do about it is the harder part.

Position 7 with almost no clicks isn't a ranking problem. It's a message problem, and that's the cheapest kind to fix.

How the agent pipeline works

The system is a Python 3.12 service built on FastAPI. It runs on a 72-hour schedule (APScheduler) and goes through the same steps every cycle:

1. Pull query data for each site from the Google Search Console API, authenticated with a service account.
2. Rank the opportunities: queries with real impressions, weak click-through, and positions where a better page could move the needle.
3. Put triage events on a Redis queue so each site and query is handled independently and a failure in one doesn't block the rest.
4. Hand each event to an agent built with AWS's open-source Strands Agents framework, running Claude on Amazon Bedrock. It checks existing posts to avoid overlap, then drafts the article against a strict rule set.
5. Generate a hero image with Google Gemini, store it in S3 and serve it through CloudFront.
6. Write the finished article to our multi-tenant content backend over AppSync GraphQL, and store run history in SQLite.

Every run is traced with OpenTelemetry into Langfuse. When an article comes out wrong, we can open the trace and see which prompt, tool call or data point caused it. Without that visibility, agent systems are guesswork. The code is checked with pytest and Ruff, and the whole thing runs in Docker Compose, so a new site is mostly configuration rather than new code.

What the agents are not allowed to do

Automation without rules produces junk at scale. The agent writes under constraints: no invented statistics, no fake testimonials, no pricing claims beyond hedged market ranges, citations only from stable authoritative sources, and one topic per article. It also has to link to real service pages and avoid repeating a topic we've already published.

That's why you'll see posts like SEO Company Vaughan covering a specific local angle instead of generic advice. The same discipline applies to security coverage, such as our September 2026 WordPress vulnerability recap. Agents can gather and draft, but the rules keep the output honest.

Should you build your own agent pipeline?

Probably not, unless you run many sites or locations. For a single-location business, a well-built site, a tuned Google Business Profile and a steady publishing habit will beat a custom agent stack on cost. The pipeline pays off when the same analysis has to happen repeatedly across many properties, or when a process involves data from several systems that nobody has time to reconcile.

If that sounds like you, we're happy to look at the workflow and tell you honestly whether automation fits. We can also show you what's working on your site now with a free website trial.

The takeaway

"DS Web" means a team that builds websites and software for local businesses and uses AI agents to handle the repetitive search-data work. Agents find the gaps and draft the fixes, and clear rules keep the results trustworthy. If you want that kind of system for your business, start with our AI automation page.